Introduction

What is statistics?
The Word statistics has been derived from Latin word “Status” or the Italian word “Statista”, the
meaning of these words is “Political State” or a Government. Early applications of statistical thinking
revolved around the needs of states to base policy on demographic and economic data.

Definition
Statistics: a branch of science that deals with collection presentation, analysis, and interpretation of
data. The definition points out 4 key aspects of statistics namely
(i) Data collection (iii) Data analysis, and
(ii) Data presentation, (iv) Data interpretation

Statistics is divided into 2 broad categories namely descriptive and inferential statistics.
Descriptive Statistics: summary values and presentations which gives some information about the
data Eg the mean height of a 1st year student in JKUAT is170cm. 170cm is a statistics which describes
the central point of the heights data.
Inferential Statistics: summary values calculated from the sample in order to make conclusions about
the target population.

Types of Variables
Qualitative Variables: Variables whose values fall into groups or categories. They are called
categorical variables and are further divided into 2 classes namely nominal and ordinal variables
a) Nominal variables: variables whose categories are just names with no natural ordering. Eg gender
marital status, skin colour, district of birth etc
b) Ordinal variables: variables whose categories have a natural ordering. Eg education level,
performance category, degree classifications etc

Quantitative Variables: these are numeric variables and are further divided into 2 classes namely
discrete and continuous variables
a) Discrete variables: can only assume certain values and there are gaps between them. Eg the
number of calls one makes in a day, the number of vehicles passing through a certain point etc
b) Continuous variables: can assume any value in a specified range. Eg length of a telephone call,
height of a 1st year student in JKUAT etc

1. Data Collection:
1.1 Sources of Data
There are 2 sources for data collection namely Primary, and Secondary data
Primary data:- freshly collected ie for the first time. They are original in character ie they are the
first hand information collected, compiled and published for some purpose. They haven’t undergone
any statistical treatment
Secondary Data:- 2nd hand information mainly obtained from published sources such as statistical
abstracts books encyclopaedias periodicals, media reports eg census report CD-roms and other

1
 electronic devices, internet. They are not original in character and have undergone some statistical
treatment at least once.

1.2 Data Collection Methods

The 1st step in any investigation (inquiry) is data collection. Information can either be collected
directly or indirectly from the entire population or a sample.
There are many methods of collecting data which includes the ones illustrated in the flow chart below
Methods of data collection

Experimental or laboratory methods Non experimental methods

Simulation Lab methods Field expt Field methods Library methods

Sample case Field Census

Surveys study study

Experimental methods are so called because in them the investigator in a laboratory tests the
hypothesis about the cause and effect relationship by manipulating the independent variables under
controlled conditions.
Non-Experimental methods are so called because in them the investigator does not control or change
any aspect of the situation under study but simply describes what naturally occurs at a certain point or
period of time.
Non-Experimental methods are widely used in social sciences. Some of the Non-Experimental
methods used for data collection are outlined below.

a) Field study:- aims at testing hypothesis in natural life situations. It differs from field experiment in
that the researcher does not control or manipulate the independent variables but both of them are
carried out in natural conditions
Merits:
(i) The method is realistic as it is carried out in natural conditions
(ii) It’s easy to obtain data with large number of variables.

Demerits
(iii) Independent variables are not manipulated.
(iv) Co-operation of the organization is often difficult to obtain.
(v) Data is likely to contain unknown sampling biasness.
(vi) The dross rate (proportion of irrelevant data) may be high in such studies.
(vii) Measurement is not precise as in laboratory because of influence of confounding
variables.
b) Census. A census is a study that obtains data from every member of a population (totality of
individuals /items pertaining to certain characteristics). In most studies, a census is not practical,
because of the cost and/or time required.

2
c) Sample survey. A sample survey is a study that obtains data from a subset of a population, in order to
estimate population attributes/ characteristics. Surveys of human populations and institutions are
common in government, health, social science and marketing research.
d) Case study –It’s a method of intensively exploring and analyzing the life of a single social unit be it a
family, person, an institution, cultural group or even an entire community. In this method no attempt
is made to exercise experimental or statistical control and phenomena related to the unit are studied in
natural. The researcher has several discretion in gathering information from a variety of sources such
as diaries, letters, autobiographies, records in office, files or personal interviews.
Merits:
(i) The method is less expensive than other methods.
(ii) Very intensive in nature –aims at studying a few units rather than several
(iii) Data collection is flexible since the researcher is free to approach the problem from any angle.
(iv) Data is collected from natural settings.

Demerits
(i) It lacks internal validity which is basic to scientific evidence.
(ii) Only one unit of the defined population is studied. Hence the findings of case study cannot be
used as abase for generalization about a large population. They lack external validity.
(iii) Case studies are more time consuming than other methods.

e) Experiment. An experiment is a controlled study in which the researcher attempts to understand

cause-and-effect relationships. In experiments actual experiment is carried out on certain individuals /
units about whom information is drawn. The study is "controlled" in the sense that the researcher
controls how subjects are assigned to groups and which treatments each group receives.
f) Observational study. Like experiments, observational studies attempt to understand cause-and-effect
relationships. However, unlike experiments, the researcher is not able to control how subjects are
assigned to groups and/or which treatments each group receives. Under this method information, is
sought by direct observation by the investigator.

1.3 Population and Sample

Population: The entire set of individuals about which findings of a survey refer to.
Sample: A subset of population selected for a study.
Sample Design: The scheme by which items are chosen for the sample.
Sample unit: The element of the sample selected from the population.
Unit of analysis: Unit at which analysis will be done for inferring about the population. Consider that
you want to examine the effect of health care facilities in a community on prenatal care. What is the
unit of analysis: health facility or the individual woman?.

Sampling Frames
For probability sampling, we must have a list of all the individuals (units) in the population. This list
or sampling frame is the basis for the selection process of the sample. “A [sampling] frame is a clear
and concise description of the population under study, by virtue of which the population units can be
identified unambiguously and contacted, if desired, for the purpose of the survey” - Hedayet and
Sinha, 1991
Based on the sampling frame, the sampling design could also be classified as:
Individual Surveys if List of individuals is available or when the size of population is small
Special population
Household Surveys; If it’s Based on the census of the households and if the individual level
information is unlikely to be available In practice, it’s limited to small geographical areas and know
as “area sampling frame” Example: Demographic and Health Surveys (DHS)
Institutional Surveys If it’s Based on the census of say Hospital/clinic lists eg

3
i) 1990 National Hospital Discharge Survey
ii) National Ambulatory Medical Care Survey

Problems of Sampling Frame

(i) Missing elements (v) Undercoverage
(ii) Noncoverage (vi) May not be readily available
(iii) Incomplete frame (vii) Expensive to gather
(iv) Old list

1.4 Sampling
Sampling is a statistical process of selecting a representative sample. We have probability sampling
and non-probability sampling Probability Samples involves a mathematical chance of selecting the
respondent. Every unit in the population has a chance, greater than zero, of being selected in the
sample. Thus producing unbiased estimates. They include;
(i) Simple random sampling (iv) Cluster sampling
(ii) Systematic sampling (v) multi-stage sampling
(iii) Stratified sampling

Non-probability sampling is any sampling method where some elements of the population have no
chance of selection (also referred to as “out of coverage”/”undercovered”), or where the probability of
selection can't be accurately determined. It yields a non-random sample therefore making it difficult to
extrapolate from the sample to the population. They include; Judgement sample, purposive sample,
convenience sample: subjective Snow-ball sampling: rare group/disease study

1.4.1 Sampling Procedure

Sampling involves two tasks
 How to select the elements?
 How to estimate the population characteristics – from the sampling units?

4
We employ some randomization process for sample selection so that there is no preferential treatment
in selection which may introduce selectivity bias

1.4.2 Reasons Behind sampling

(i) Cost; the sample can furnish data 0f sufficient accuracy at much lower cost.
(ii) Time; the sample provides information faster than census thus ensuring timely decision making.
(iii) Accuracy; it is easier to control data collection errors in a sample survey as opposed to census.
(iv) Risky or destructive test call for sample survey not census eg testing a new drug.

1.4.3 Probability Sampling Techniques

a)...Simple Random Sampling (SRS)
In this design, each element has an equal probability of being selected from a list of all population units
(sample of n from N population). Though it’s attractive for its simplicity, the design is not usually used
in the sample survey in practice for several reasons:
(i) Lack of listing frame: the method requires that a list of population elements be available, which is
not the case for many populations.
(ii) Problem of small area estimation or domain analysis: For a small sample from a large population,
all the areas may not have enough sample size for making small area estimation or for domain
analysis by variables of interest.
(iii) Not cost effective: SRS requires covering of whole population which may reside in a large
geographic area; interviewing few samples spread sparsely over a large area would be very costly.

Implementation of SRS sampling:

(i) Listing (sampling) Frame
(ii) Random number table (from published table or computer generated)
(iii) Selection of sample

b)..Systematic Sampling
Systematic sampling, either by itself or in combination with some other method, may be the most
widely used method of sampling.” In systematic sampling we select samples “evenly” from the list
(sampling frame): First, let us consider that we are dividing the list evenly into some “blocks”. Then, we
select a sample element from each block.

In systematic sampling, only the first unit is selected at random, the rest being selected according to a
predetermined pattern. To select a systematic sample of n units, the first unit is selected with a random
start r from 1 to k sample, where k=N/n sample intervals, and after the selection of first sample, every
kth unit is included where 1≤ r ≤ k.

5
 An example:
Let N=100, n=10, then k=100/10. Then the random start r is selected between 1 and 10 (say,
r=7). So, the sample will be selected from the population with serial indexes of: 7, 17, 27, . . . .
. .,97. i.e., r, r+k, r+2k,......., r+(n-1)k

What could be done if k=N/n is not an integer?

Selection of systematic sampling when sampling interval (k) is not an integer
Consider, n=175 and N=1000. So, k=1000/175 = 5.71
One of the solution is to make k rounded to an integer, i.e., k=5 or k=6.Now, if
k=5, then n=1000/5=200; or, If k=6, then n=1000/6 = 166.67 ~ 167. Which n
should be chosen?

Solution
if k=5 is considered, stop the selection of samples when n=175 achieved.
if k=6 is considered, treat the sampling frame as a circular list and continue the
selection of samples from the beginning of the list after exhausting the list during
the first cycle.

An alternative procedure is to keep k non-integer and continue the sample

selection as follows: Let us consider, k=5.71, and r=4. So, the first sample is 4th
in the list. The second = (4+5.71) =9.71 ~9th in the list, the third =(4+2*5.71)
=15.42 ~ 15th in the list, and so on. (The last sample is: 4+5.71*(175-1) = 997.54
~ 997th in the list). Note that, k is switching between 5 and 6

Advantages:
Systematic sampling has many attractiveness:
(i) Provides a better random distribution than SRS
(ii) Simple to implement
(iii)May be started without a complete listing frame (say, interview of every 9th patient
coming to a clinic).
(iv) With ordered list, the variance may be smaller than SRS (see below for exceptions

Disadvantages:
(i) Periodicity (cyclic variation)
(ii) linear trend
i.
When to use systematic sampling?
i) Even preferred over SRS
ii) When no list of population exists
iii) When the list is roughly of random order
iv) Small area/population

c)..Stratified Sampling
In stratified sampling the population is partitioned into groups, called strata, and sampling is performed
separately within each stratum.
This sampling technique is used when;
i) Population groups may have different values for the responses of interest.
ii) we want to improve our estimation for each group separately.
iii) To ensure adequate sample size for each group.

In stratified sampling designs:

6
i) Stratum variables are mutually exclusive (no over lapping), e.g., urban/rural areas, economic
categories, geographic regions, race, sex, etc. The principal objective of stratification is to reduce
sampling errors.
ii) The population (elements) should be homogenous within-stratum, and the population
(elements) should be heterogeneous between the strata.

Advantages
(i) Provides opportunity to study the stratum; variations - estimation could be made for each stratum
(ii) Disproportionate sample may be selected from each stratum
(iii) The precision is likely to increase as variance may be smaller than simple random case with same
sample size
(iv) Field works can be organized using the strata (e.g., by geographical areas or regions)
(v) Reduce survey costs.

Disadvantages
(i) Sampling frame is needed for each stratum
(ii) Analysis method is complex
(iii) Correct variance estimation
(iv) Data analysis should take sampling “weight” into account for disproportionate sampling of strata
(v) Sample size estimation is difficult in practice

Allocation of Stratified Sampling

The major task of stratified sampling design is the appropriate allocation of samples to
different strata.
Types of allocation methods:
(i) Equal allocation
(ii) Proportional to stratum size
(iii) Cost based sample allocation

Equal Allocation
Divide the number of sample units n equally among the K strata. ie n i  n
k Example: n = 100 and k= 4
strata n i  100
4
 25 units in each stratum.
Disadvantages of equal allocation:
May need to use weighting to have unbiased estimates

Proportional allocation
Make the proportion of each stratum sampled identical to the proportion of the population. Ie
Let the sample fraction be f= n/N. So, ni  fN i  n Ni , Where NN is the stratum weight.
N i

Example: N = 1000, n = 100 f  1000

100
 0.1 now suppose N1  700 and N 2  300 then
n1  700 * 0.1  70 and n2  300 * 0.3  30
Disadvantage of proportional allocation:
Sample size in a stratum may be low thus providing unreliable stratum-specific results.

d)..Cluster Sampling
In many practical situations the population elements are grouped into a number of clusters. A list of
clusters can be constructed as the sampling frame but a complete list of elements is often unavailable, or
too expensive to construct. In this case it is necessary to use cluster sampling where a random sample of
clusters is taken and some or all elements in the selected clusters are observed. Cluster sampling is also
preferable in terms of cost, because it is much cheaper, easier and quicker to collect data from adjoining
7
 elements than elements chosen at random. On the other hand, cluster sampling is less informative and
less efficient per elements in the sample, due to similarities of elements within the same cluster. The
loss of efficiency, however, can often be compensated by increasing the overall sample size. Thus, in
terms of unit cost, the cluster sampling plan is efficient.

e)..Multi-Stage Samples
Here the respondents are chosen through a process of defined stages. Eg residents within Kibera
(Nairobi) may have been chosen for a survey through the following process:
Throughout the country (Kenya) the Nairobi may have been selected at random, ( stage 1), within
Nairobi, Langata (constituency) is selected again at random (stage 2), Kibera is then selected within
Langata (stage 3), then polling stations from Kibera (stage 4) and then individuals from the electoral
voters’ register (stage 5)! As demonstrated five stages were gone through before the final selection of
respondents were selected from the electoral voters’ register.

Advantages of probability sample

(i) Provides a quantitative measure of the extent of variation due to random effects
(ii) Provides data of known quality
(iii)Provides data in timely fashion
(iv) Provides acceptable data at minimum cost
(v) Better control over nonsampling sources of errors
(vi) Mathematical statistics and probability can be applied to analyze and interpret the data

1.4.4 Non-probability Sampling

Social research is often conducted in situations where a researcher cannot select the kinds of probability
samples used in large-scale social surveys. For example, say you wanted to study homelessness - there
is no list of homeless individuals nor are you likely to create such a list. However, you need to get some
kind of a sample of respondents in order to conduct your research. To gather such a sample, you would
likely use some form of non-probability sampling.
There are four primary types of non-probability sampling methods:

a)..Convinience Sampling
It’s a method of choosing subjects who are available or easy to find. This method is also sometimes
referred to as haphazard, accidental, or availability sampling. The primary advantage of the method is
that it is very easy to carry out, relative to other methods.
Demerit
 One can never be certain what population the participants in the study represent. The population is
unknown.
 The method is haphazard, and the cases studied probably don't represent any population you could come
up with. However, it’s very useful for pilot studies
Advantages of convenience sample
(i) It’s very easy to carry out with few rules governing how the sample should be collected.
(ii) The relative cost and time required to carry out a convenience sample are small in comparison to
probability sampling techniques. This enables you to achieve the sample size you want in a
relatively fast and inexpensive way.

8
 (iii) The convenience sample may help you gather useful data and information that would not have been
possible using probability sampling techniques, which require more formal access to lists of
populations [see, for example, the article on simple random sampling].

For example, imagine you were interested in understanding more about employee satisfaction in a
single, large organisation in the United States. You intended to collect your data using a
questionnaire. The manager who has kindly given you access to conduct your research is unable to
get permission to get a list of all employees in the organisation, which you would need to use a
probability sampling technique such as simple random sampling or systematic random sampling.
However, the manager has managed to secure permission for you to spend two days in the
organisation to collect as many questionnaire responses as possible. You decide to spend the two
days at the entrance of the organisation where all employees have to pass through to get to their
desks. Whilst a probability sampling technique would have been preferred, the convenience sample
was the only sampling technique that you could use to collect data. Irrespective of the
disadvantages of convenience sampling, discussed below, without the use of this sampling
technique, you may not have been able to get access to any data on employee satisfaction in the
organisation.

Disadvantages of convenience sampling

 The convenience sample often suffers from a number of biases. This can be seen in both of our
examples, whether the 10,000 students we were studying, or the employees at the large organisation. In
both cases, a convenience sample can lead to the under-representation or over-representation of
particular groups within the sample. If we take the large organisation:

It may be that the organisation has multiple sites, with employee satisfaction varying considerably
between these sites. By conducting the survey at the headquarters of the organisation, we may have
missed the differences in employee satisfaction amongst those at different sites, including non-office
workers. We also do not know why some employees agreed to take part in the survey, whilst others did
not. Was it because some employees were simply too busy? Did they not trust the intentions of the
survey? Did others take part out of kindness or because they had a particular grievance with the
organisation? These types of biases are quite typical in convenience sampling.
 Since the sampling frame is not know, and the sample is not chosen at random, the inherent bias in
convenience sampling means that the sample is unlikely to be representative of the population being
studied. This undermines your ability to make generalisations from your sample to the population you
are studying.
If you are an undergraduate or master’s level dissertation student considering using convenience
sampling, you may also want to read more about how to put together your sampling strategy [see the
section: Sampling Strategy

b)..Quota Sampling
Quota sampling is designed to overcome the most obvious flaw of availability sampling. Rather than
taking just anyone, you set quotas to ensure that the sample you get represents certain characteristics in
proportion to their prevalence in the population. Note that for this method, you have to know something
about the characteristics of the population ahead of time. Say you want to make sure you have a sample
proportional to the population in terms of gender - you have to know what percentage of the population
is male and female, then collect sample until yours matches. Marketing studies are particularly fond of
this form of research design.
The primary problem with this form of sampling is that even when we know that a quota sample is
representative of the particular characteristics for which quotas have been set, we have no way of
knowing if sample is representative in terms of any other characteristics. If we set quotas for gender and

9
age, we are likely to attain a sample with good representativeness on age and gender, but one that may
not be very representative in terms of income and education or other factors.
Moreover, because researchers can set quotas for only a small fraction of the characteristics relevant to
a study quota sampling is really not much better than availability sampling. To reiterate, you must know
the characteristics of the entire population to set quotas; otherwise there's not much point to setting up
quotas. Finally, interviewers often introduce bias when allowed to self-select respondents, which is
usually the case in this form of research. In choosing males 18-25, interviewers are more likely to
choose those that are better-dressed, seem more approachable or less threatening. That may be
understandable from a practical point of view, but it introduces bias into research findings.
Imagine that a researcher wants to understand more about the career goals of students at a single
university. Let’s say that the university has roughly 10,000 students. suppose we were interested in
comparing the differences in career goals between male and female students at the single university. If
this was the case, we would want to ensure that the sample we selected had a proportional number of
male and female students relative to the population.
To create a quota sample, there are three steps:
Choose the relevant grouping chsr and divide the population accordingly gender
Calculate a quota (number of units that should be included in each for group
Continue to invite units until the quota for each group is met

Advantages of quota sampling

i) It particularly useful when you are unable to obtain a probability sample, but you are still trying to
create a sample that is as representative as possible of the population being studied. In this respect,
it is the non-probability based equivalent of the stratified random sample.
ii) Unlike probability sampling techniques, especially stratified random sampling, quota sampling is
much quicker and easier to carry out because it does not require a sampling frame and the strict use
of random sampling techniques.
iii) The quota sample improves the representation of particular strata (groups) within the population,
as well as ensuring that these strata are not over-represented. For example, it would ensure that we
have sufficient male students taking part in the research (60% of our sample size of 100; hence, 60
male students). It would also make sure we did not have more than 60 male students, which would
result in an over-representation of male students in our research.
iv) It allows comparison of groups.

Disadvantages of quota sampling

i) In quota sampling, the sample has not been chosen using random selection, which makes it
impossible to determine the possible sampling error.
ii) this sampling bias. Thus nostatistical inferences from the sample to the population. This can lead to
problems of external validity.
iii) Also, with quota sampling it must be possible to clearly divide the population into strata; that is,
each unit from the population must only belong to one stratum. In our example, this would be fairly
simple, since our strata are male and female students. Clearly, a student could only be classified as
either male or female. No student could fit into both categories (ignoring transgender issues).

c)..Purposive Sampling
Purposive sampling is a sampling method in which elements are chosen based on purpose of the study.
Purposive sampling may involve studying the entire population of some limited group or a subset of a
population. As with other non-probability sampling methods, purposive sampling does not produce a
sample that is representative of a larger population, but it can be exactly what is needed in some cases -
study of organization, community, or some other clearly defined and relatively limited group.
10
Advantages of purposive sampling
i) There are a wide range of qualitative research designs that researchers can draw on. Achieving the
goals of such qualitative research designs requires different types of sampling strategy and
sampling technique. One of the major benefits of purposive sampling is the wide range of sampling
techniques that can be used across such qualitative research designs; purposive sampling techniques
that range from homogeneous sampling through to critical case sampling, expert sampling, and
more.
ii) Whilst the various purposive sampling techniques each have different goals, they can provide
researchers with the justification to make generalisations from the sample that is being studied,
whether such generalisations are theoretical, analytic and/or logical in nature. However, since each
of these types of purposive sampling differs in terms of the nature and ability to make
generalisations, you should read the articles on each of these purposive sampling techniques to
understand their relative advantages.
iii) Qualitative research designs can involve multiple phases, with each phase building on the previous
one. In such instances, different types of sampling technique may be required at each phase.
Purposive sampling is useful in these instances because it provides a wide range of non-probability
sampling techniques for the researcher to draw on. For example, critical case sampling may be used
to investigate whether a phenomenon is worth investigating further, before adopting an expert
sampling approach to examine specific issues further.

Disadvantages of purposive sampling

i) Purposive samples, irrespective of the type of purposive sampling used, can be highly prone to
researcher bias. The idea that a purposive sample has been created based on the judgement of the
researcher is not a good defence when it comes to alleviating possible researcher biases,
ii) specially when compared with probability sampling techniques that are designed to reduce such
biases. However, this judgemental, subjective component of purpose sampling is only a major
disadvantage when such judgements are ill-conceived or poorly considered; that is, where
judgements have not been based on clear criteria, whether a theoretical framework, expert
elicitation, or some other accepted criteria.
iii) The subjectivity and non-probability based nature of unit selection (i.e. selecting people,
cases/organisations, etc.) in purposive sampling means that it can be difficult to defend the
representativeness of the sample. In other words, it can be difficult to convince the reader that the
judgement you used to select units to study was appropriate. For this reason, it can also be difficult
to convince the reader that research using purposive sampling achieved theoretical/analytic/logical
generalisation. After all, if different units had been selected, would the results and any
generalisations have been the same?

d)..Snowball Sampling
Snowball sampling is a method in which a researcher identifies one member of some population of
interest, speaks to him/her, and then asks that person to identify others in the population that the
researcher might speak to. This person is then asked to refer the researcher to yet another person, and so
on.
Snowball sampling is very good for cases where members of a special population are difficult to locate.
For example,.populations that are subject to social stigma and marginalisation, such as suffers of
AIDS/HIV, as well as individuals engaged in illicit or illegal activities, including prostitution and drug
use. Snowball sampling is useful in such scenarios because:
The method creates a sample with questionable representativeness. A researcher is not sure who is in
the sample. In effect snowball sampling often leads the researcher into a realm he/she knows little
about. It can be difficult to determine how a sample compares to a larger population. Also, there's an
issue of who respondents refer you to - friends refer to friends, less likely to refer to ones they don't like,
fear, etc.
Snowball sampling is a useful choice of sampling strategy when the population you are interested in
studying is hidden or hard-to-reach.

11
Advantages of Snowball Sampling
(i) The chain referral process allows the researcher to reach populations that are difficult to sample
when using other sampling methods.
(ii) The process is cheap, simple and cost-efficient.
(iii) This sampling technique needs little planning and fewer workforce compared to other sampling
techniques.

Disadvantages of Snowball Sampling

(i) The researcher has little control over the sampling method. The subjects that the researcher can
obtain rely mainly on the previous subjects that were observed.
(ii) Representativeness of the sample is not guaranteed. The researcher has no idea of the true
distribution of the population and of the sample.
(iii) Sampling bias is also a fear of researchers when using this sampling technique. Initial subjects tend
to nominate people that they know well. Because of this, it is highly possible that the subjects share
the same traits and characteristics, thus, it is possible that the sample that the researcher will obtain
is only a small subgroup of the entire populatio

1.4.5 Limitations of Sampling

a) Sampling frame: may need complete enumeration
b) Errors of sampling may be high in small areas
c) May not be appropriate for the study objectives/questions
d) Representativeness may be vague, controversial

1.4.6 Characteristics of Good sampling

A good sample should;
a) Meet the requirements of the study objectives
b) Provides reliable results
c) Clearly understandable
d) Manageable/realistic: could be implemented
e) Time consideration: reasonable and timely
f) Cost consideration: economical
g) Interpretation: accurate, representative
h) Acceptability

1.5 Survey Administration

1.5.1 Steps in Survey
1. Setting the study objectives; What are the objectives of the study? Is survey the best procedure to
collect data? Why other study design (experimental, quasi-experimental, community randomized trials,
epidemiologic designs,,e.g., case-control study) is not appropriate for the study? What information/data
need to be collected?

2. Defining the study population; Representativeness Sampling frame

3. Decide sample design: alternative considerations
4. Questionnaire design; Appropriateness, acceptability, culturally appropriate, understandable Pre-
test: Appropriate, acceptable, culturally appropriate, will answer
5. Fieldwork; Training/Supervision Quality monitoring Timing: seasonality
6. Quality assurance Every steps Minimizing errors/bias/cheating
7. Data entry/compilation Validation Feedback
8. Analysis: Design consideration
9. Dissemination
10. Plans for next survey: what did you learn, what did you miss?
1.5.2 Modes of Survey Administration
12
 a) Self-Administered Surveys
b) Personal interview
c) Telephone
d) Mail
e) Computer assisted self-interviewing(CASI)Variants: CAPI (personal interview); CATI
(telephone interview) – Replaces the papers
f) Combination of methods
a)..Self-Administered Surveys
Self-administered surveys have special strengths and weaknesses.
They are useful in describing the characteristics of a large population and make large samples feasible.
Advantages:
i) Low cost. Extensive training is not required to administer the survey. Processing and analysis
are usually simpler and cheaper than for other methods.
ii) Reduction in biasing error. The questionnaire reduces the bias that might result from personal
characteristics of interviewers and/or their interviewing skills.
iii) Greater anonymity. Absence of an interviewer provides greater anonymity for the respondent.
This is especially helpful when the survey deals with sensitive issues.
iv) Convenience to the respondents (may complete any time at his/her own convenient time)
v) Accessibility (greater coverage, even in the remote areas)
vi) May provide more reliable information (e.g., may consult with others or check records to avoid
recall bias)

Disadvantages:
i) Requires simple questions. The questions must be straightforward enough to be comprehended
solely on the basis of printed instructions and definitions.
ii) No opportunity for probing. The answers must be accepted as final. Researchers have no
opportunity to clarify ambiguous answers.
iii) Low response rate; respondents may not respond to all questions and/or may not return
questionnaire
iv) The respondent must be literate to read and understand the questionnaire
v) Introduce self selection bias
vi) Not suitable for complex questionnaire

b). . .Interview Surveys

Unlike questionnaires interviewers ask questions orally and record respondents’ answers. This type of
survey generally decreases the number of ―”do not know” and ―”no answer” responses, compared
with self-administered surveys. Interviewers also provide a guard against confusing items. If a
respondent has misunderstood a question, the interviewer can clarify, thereby obtaining relevant
responses.
Interviewer selection: background characteristics (race, sex, education, culture) listening skill
recording skill experience unbiased observation/recording
Interviewer training: be familiar with the study objectives and significance thorough familiarity with
the questionnaire contextual and cultural issues privacy and confidentiality informed consent and ethical
issues unbiased view mock interview session
Supervision of the interviewer: Spot check Questionnaire check Reinterview (reliability check)
Advantages
i) Flexibility. Allows flexibility in the questioning process and allows the interviewer to clarify
terms that are unclear.
ii) Control of the interview situation. Can ensure that the interview is conducted in private, and
respondents do not have the opportunity to consult one another before giving their answers.
iii) High response rate. Respondents who would not normally respond to a mail questionnaire will
often respond to a request for a personal interview.

13
 iv) May record non-verbal behaviour, activities, facilities, contexts
v) Complex questionnaire may be used
vi) Illiterate respondents may participate

Disadvantages
i) Higher cost. Costs are involved in selecting, training, and supervising interviewers; perhaps in
paying them; and in the travel and time required to conduct interviews.
ii) Interviewer bias. The advantage of flexibility leaves room for the interviewer’s personal
influence and bias, making an interview subject to interviewer bias.
iii) Lack of anonymity. Often the interviewer knows all or many of the respondents. Respondents
may feel threatened or intimidated by the interviewer, especially if a respondent is sensitive to
the topic or to some of the questions.
iv) Less accessibility
v) Inconvenience
vi) Often no opportunity to consult records, families, relatives

c).. Telephone Interview

Advantages:
(i) Less expensive
(ii) Shorter data collection period than personal interviews
(iii) Better response than mail surveys

Disadvantages:
(i) Biased against households without telephone, unlisted number
(ii) Nonresponse
(iii) Difficult for sensitive issues or complex topics
(iv) Limited to verbal responses

d)… Focus Groups

Focus groups are useful in obtaining a particular kind of information that would be difficult to obtain
using other methodologies. A focus group typically can be defined as a group of people who possess
certain characteristics and provide information of a qualitative nature in a focused discussion
Focus groups generally are composed of six to twelve people. Size is conditioned by two factors: the
group must be small enough for everyone to participate, yet large enough to provide diversity. This
group is special in terms of purpose, size, composition, and procedures. Participants are selected
because they have certain characteristics in common that relate to the topic at hand, such as parents of
gang members, and, generally, the participants are unfamiliar with each other. Typically, more than one
focus group should be convened, since a group of seven to twelve people could be too atypical to offer
any general insights on the gang problem.
A trained moderator probes for different perceptions and points of view, without pressure to reach
consensus. Focus groups have been found helpful in assessing needs, developing plans, testing new
ideas, or improving existing programs
Advantages:
i) Flexibility allows the moderator to probe for more in-depth analysis and ask participants to
elaborate on their responses.
ii) Outcomes are quickly known.
iii) They may cost less in terms of planning and conducting than large surveys and personal
interviews.
Limitations
i) A skilled moderator is essential.
ii) Differences between groups can be troublesome to analyze because of the qualitative nature of
the data.
14
 iii) Groups are difficult to assemble. People must take the time to come to a designated place at a
particular time.
iv) Participants may be less candid in their responses in front of peers.

2. Data Presentation
2.1 Frequency Distributions Tables
Definitions
Raw Data: unprocessed data ie data in its original form.
Frequency Distribution: The organization of raw data in table form with classes and
frequencies. Rather it’s a list of values and the number of times they appear in the data set.
We have grouped and ungrouped frequency distribution tables for large and small data sets
respectively.

2.1.1 Construction of Ungrouped Frequency Distributions

 Note the largest and smallest observations in the data
 Starting with the smallest value, tally the observations of each quantity.
 Count the number of tallies for each quantity and record it as frequency.

Example: Construct an ungrouped frequency table for the data below

16 14 15 13 12 14 16 15 15 14 17 16 13 16 15 14 18 13 15 17

Solution
Value 12 13 14 15 16 17 18
Tally / /// //// //// //// // /
Frequency 1 3 4 5 4 2 1
Note ; when tallying //// is used for 5 counts and not /////

2.1.2 Construction of Grouped Frequency Distributions

When the number of observations is too large and/or when the variable of interest is
continuous, it’s cumbersome to consider the repetition of each observation. A quick and more
convenient way is to group the range of values into a number of exclusive groups or classes
and count the class frequency. The resulting table is called a grouped frequency distribution
table.
A grouped frequency distribution consists of classes and their corresponding frequencies. Each raw data
value is placed into a quantitative or qualitative category called a class. The frequency of a class then is the
number of data values contained in a specific class.

Steps in construction
 Select the number of classes k. Choose the smallest integer k such that
log n log 30
2k  n  k  Eg if n = 30 k   k 5
log 2 log 2
 Identify the largest and smallest observation and compute the ranrge R=largest –smallest
value.
 Identify the smallest unit of measurement (u) used in the data collection (ie the accuracy of
the measurement.)
Eg for the data 10 30 20 50 30 60 u=10 For the data 12 15 11 17 13 u=1
For the data 1.6 3.2 2.8 5.6 3.5 1.6 u=0.1
15
 Estimate the class width/interval. i = round up  Rk  to the nearest u
Eg round up 3.13 t0 the nearest 0.1 is 3.2
 Pick the starting value (lower class limit of the 1st class (LCL1)) as the smallest value used
in the computation of R above. Successive LCL s are got by adding I to the previous LCL
 Find the upper class limit of the 1st class (UCL1)) by subtracting u from LCL2. Successive
UCLs are got by adding I to the previous UCL
 If necessary find the class boundaries as follows LCB  LCL  12 u and UCB  UCL  12 u
 Tally the number of observations falling in each class and record the frequency.
NB a value x fall into a class LCL UCL if LCB  x  UCB
Example 1
Organize the data below into a grouped frequency table.
15.0 17.4 10.3 9.2 20.7 18.9 16.6 22.4 23.7 18.6 26.1 16.5 19.7 12.9 15.7
30.8 15.4 20.3 24.0 29.6 18.3 23.7 17.8 24.6 23.0 21.4 32.8 12.5 17.5 18.3
23.2 21.6 20.8 29.8 24.5 28.4 13.5 17.1 27.1 27.9
Solution
u  0.1 n  40  k  log 40
log 2
 k 6 Range = 32.8 - 9.2  23.6
i = round up  236.6  to the nearest 0.1  4.0
Now LCL 1 = 9.2  LCL 2  13.2 LCL 3  17.2 etc
 UCL 1  LCL 2  u  13.2  0.1  13.1, UCL 2  17.1 UCL 3  21.1 etc
The frequency table is as shown belo

Class Boundaries tally Freq C.F

9.2 - 13-1 9.15 - 13-15 //// 4 4
13.2 - 17.1 13.15 - 17.15 //// // 7 11
17.2 - 21-1 17.15 - 21-15 //// //// / 11 22
21.2 - 25.1 21.15 - 25.15 //// //// 10 32
25.2 - 29.1 25.15 - 29.15 //// 4 36
29.2 - 33- 29.15 - 33-15 //// 4 40
Sometimes due to convenience, it may be necessary to slightly lower the starting value. Eg in
the above case we may use 9.0 in place of 9.2.

Example 2
These data represent the record high temperatures in degrees Fahrenheit (F) for each of the 50
states. Construct a grouped frequency distribution for the data.
112 100 127 120 134 118 105 110 109 112 110 118 117 116 118 122 114 114 105
109 107 112 114 115 118 117 118 122 106 110 116 108 110 121 113 120 119 111
104 111 120 113 120 117 105 110 118 112 114 114
Solution
u  1 n  50  k  log 50
log 2
 k 6 Range = 134 - 100  34
i = round up 346  to the nearest 1  6
Now LCL 1 = 100  LCL 2  106 LCL 3  112 etc
 UCL1  LCL 2  u  106  1  105, UCL 2  111 UCL 3  117 etc
The frequency table is as shown below
Class Boundaries tally Freq C.F
100 - 105 99.5 – 105.5 //// 5 5
16
 106 - 111 105.5 - 111.5 //// //// /// 13 18
112 - 117 111.5 – 117.5 //// //// //// // 16 34
118 - 123 117.5 -123.5 //// //// //// 14 48
124 - 129 124.5 - 129.5 / 1 49
130 - 135 129.5 - 135-5 / 1 50

Exercise
1) The data shown here represent the number of miles per gallon (mpg) that 30 selected four-
wheel-drive sports utility vehicles obtained in city driving. Construct a frequency
distribution, and analyze the distribution. 12 17 12 14 16 18 16 18 12 16 17 15 15
16 12 15 16 16 12 14 15 12 15 15 19 13 16 18 16 14
2) Suppose a researcher wished to do a study on the ages of the top 50 wealthiest people in the
world. The researcher first would have to get the data on the ages of the people. In this
case, these ages are listed in Forbes Magazine. 49 57 38 73 81 74 59 76 65 69 54 56
69 68 78 65 85 49 69 61 48 81 68 37 43 78 82 43 64 67 52 56 81 77 79 85 40
85 59 80 60 71 57 61 69 61 83 90 87 74 Organize the data into a grouped frequency
table
3) The data represent the ages of our Presidents at the time they were first inaugurated. 57
61 57 57 58 57 61 54 68 51 49 64 50 48 65 52 56 46 54 49 50 47 55 55 54
42 51 56 55 54 51 60 62 43 55 56 61 52 69 64 46 54
a) Were the data obtained from a population or a sample? Explain your answer.
b) What was the age of the oldest and youngest President?
c) Construct a frequency distribution for the data.
d) Are there any peaks in the distribution?
e) identify any possible outliers.
4) The state gas tax in cents per gallon for 25 states is given below. Construct a grouped
frequency distribution for the data. 7.5 16 23.5 17 22 21.5 19 20 27.1 20 22 20.7 17
28 20 23 18.5 25.3 24 31 14.5 25.9 18 30 31.5
5) Listed are the weights of the NBA’s top 50 players. Construct a grouped frequency
distribution and analyze the results in terms of peaks, extreme values, etc.
240 210 220 260 250 195 230 270 325 225 165 295 205 230 250 210 220 210
230 202 250 265 230 210 240 245 225 180 175 215 215 235 245 250 215 210
195 240 240 225 260 210 190 260 230 190 210 230 185 260
6) The number of stories in each of the world’s 30 tallest buildings is listed below. Construct a
grouped frequency distribution and analyze the results in terms of peaks, extreme values,
etc. 88 88 110 88 80 69 102 78 70 55 79 85 80 100 60 90 77 55 75 55 54 60
75 64 105 56 71 70 65 72
7) The average quantitative GRE scores for the top 30 graduate schools of engineering are
listed. Construct a grouped frequency distribution and analyze the results in terms of peaks,
extreme values, etc. 767 770 761 760 771 768 776 771 756 770 763 760 747 766
754 771 771 778 766 762 780 750 746 764 769 759 757 753 758 746
8) The number of passengers (in thousands) for the leading U.S. passenger airlines in 2004 is
indicated below. Use the data to construct a grouped frequency distribution with a
reasonable number of classes and comment on the shape of the distribution.
9) 91.570 86. 755 81.066 70.786 55.373 42.400 40.551 21.119 16.280 14.869
13.659 13.417 3.170 12.632 11.731 10.420 10.024 9.122 7.041 6.954 6.406
6.362 5. 930 5.585
10) The heights (in feet above sea level) of the major active volcanoes in Alaska are given here.
Construct a grouped frequency distribution for the data. 4,265 3,545 4,025 7,050
11,413 3,490 5,370 4,885 5,030 6,830 4,450 5,775 3,945 7,545 8,450 3,995
17
 10,140 6,050 10,265 6,965 150 8,185 7,295 2,015 5,055 5,315 2,945 6,720
3,465 1,980 2,560 4,450 2,759 9,430 7,985 7,540 3,540 11,070 5,710 885
8,960 7,015

2.2 Graphical Displays

After you have organized the data into a frequency distribution, you can present them in
graphical form. The purpose of graphs in statistics is to convey the data to the viewers in
pictorial form. It is easier for most people to comprehend the meaning of data presented
graphically than data presented numerically in tables or frequency distributions. This is
especially true if the users have little or no statistical knowledge.
Statistical graphs can be used to describe the data set or to analyze it. Graphs are also useful in
getting the audience’s attention in a publication or a speaking presentation. They can be used to
discuss an issue, reinforce a critical point, or summarize a data set. They can also be used to
discover a trend or pattern in a situation over a period of time.
The commonly used graphs in research are; the pie chart, bar chart, histogram, frequency
polygon and the cumulative frequency curve (Ogive).

2.2.1 Pie Chart

It’s a circular graph having radii divide a circle into sectors proportional in angle to the relative
size of the quantities in the category being represented.
How to Draw
(i) Add up the given quantities and let s be the sum of the values
(ii) For each quantity x, calculate the representative angle and percentage as xs (3600 ) and
x
s (100%) respectively
(iii)Draw a circle and divide it into sectors using the angles calculated in step ii above
(iv) Label the sector by the group represented and indicate the corresponding percentage.
Example
This frequency distribution shows the number of pounds of each snack food eaten during the
Super Bowl. Construct a pie graph for the data.
Snack Potato chips Tortilla chips Pretzels Popcorn Snack nuts
Pounds (in millions) 11.2 8.2 4.3 3.8 2.5
Solution
Snack Potato chips Tortilla chips Pretzels Popcorn Snack nuts Total
Pounds (in millions) 11.2 8.2 4.3 3.8 2.5 30.0
Representative Angle 134 98 52 46 30 360
Representative %age 37.3 27.3 14.3 12.7 8.3 99.9

18
 .
2.2.2 Bar chart
A bar chart consists of a set of equal spaced rectangles whose heights are proportional to the
frequency of the category /item being considered. The X axis in a bar chart can represent the
number of categories.
Note: Bars are of uniform width and there is equal spacing between the bars.
Example
A sample of 250 students was asked to indicate their favourite TV channels and their
responses were as follows.
TV station KBC NTV CITIZEN KTN FAMILY
No. of viewers 28 52 92 63 15
Draw a bar chart to represent this information.
Solution see the bar chart above and on the right.

2.23 Pareto Charts

It consist of a set of continuous rectangles where the variable displayed on the horizontal axis
is qualitative or categorical and the frequencies are displayed by the heights of vertical bars,
which are arranged in order from highest to lowest. A Pareto chart is used to represent a
frequency distribution for a categorical variable,
Points to note when drawing a Pareto Chart
i) Make the bars the same width.
ii) Arrange the data from largest to smallest according to frequency.
iii) Make the units that are used for the frequency equal in size.
When you analyze a Pareto chart, make comparisons by looking at the heights of the bars.

Example
The table shown here is the average cost per mile for passenger vehicles on state turnpikes.
Construct and analyze a Pareto chart for the data.
State Indiana Oklahoma Florida Maine Pennsylvania
Number 2.9 4.3 6.0 3.8 5.8

Solution
Arrange the data from the largest to smallest according to frequency.

19
 State Florida Pennsylvania Oklahoma Maine Indiana
Number 6.0 5.8 4.3 3.8 2.9

Draw and label the x and y axes and then the bars corresponding to the frequencies. The Pareto chart
shows that Florida has the highest cost per mile. The cost is more than twice as high as the cost for Indiana.

2.2.4 Histogram
It consists of a set of continuous rectangles such that the areas of the rectangles are
proportional to the frequency. For ungrouped data, the heights of each bar is proportional to
frequency. For grouped data, the height of each rectangle is the relative frequency h and is
frequency
given by h = . The width of the bars is determined by the class boundaries.
IClass Interval
Example
Construct a histogram to represent the data shown below
Class 100-104 105-109 110 -114 115-119 120 - 24 125-129 130 -134
Freq 2 8 18 13 7 1 1

Solution
99.5- 104.5- 109.5 - 114.5- 119.5 – 124.5- 129.5 -
Boundaries 104.5 109.5 114.5 119.5 124.5 129.5 134.5
Heights 2 8 18 13 7 1 1

20
2.2.5 Frequency polygon
It’s a plot of frequency against mid points joined with straight line segments between
consecutive points. It can also be obtained by joining the mid point of the tops of the bars in a
histogram. The gaps at both ends are filled by extending to the next lower and upper
imaginary classes assuming frequency zero.

Example: Consider the following frequency distribution.

Class 5-9 10-14 15-19 20-24 25-29 30-34 35-39
freq 5 12 32 40 16 9 6
Draw a histogram to represent this information and fit a frequency polygon on it.
Solution
Boundaries 4.5-9.5 9.5-14.5 14.5-19.5 19.5-24.5 24.5-29.5 29.5-34.5 34.5-39.5
heights 1 2.4 6.4 8 3.2 1.8 1.2
The corresponding histogram is as shown below.
9
A HISTOGRAM & A FREQUENCY POLYGON FITTED ON IT
8
Relative Frequency

7
6
5
4
3
2
1
0

Class Boundaries

2.2.6 Cumulative frequency curve (ogive)

It is a plot of cumulative frequency against upper boundaries joined with a smooth curve. The
gap on the lower end is filled by extending to the next lower imaginary class assuming
frequency zero. This graph is useful in estimating median and other measures of location.

Example:
21
Construct an ogive to represent the data shown below
Class 100-104 105-109 110 -114 115-119 120 - 24 125-129 130 -134
Freq 2 8 18 13 7 1 1

Solution
Upper Boundaries 99.5 104.5 109.5 114.5 119.5 124.5 129.5 134.5
CF 0 2 10 28 41 48 49 50

Exercise
1) Construct a pie chart and a bar graph showing the blood types of the army inductees
described in the frequency distribution is repeated here.
Blood group A B AB O
Frequency 5 7 4 9
2) The table below shows the average money spent by first-year college students. Draw a pie
chart and a bar graph for the data.
Nature of Expense Electronics Dorm decor Clothing Shoes
Amount(in $) 728 344 141 72
3) The table shown here is the average cost per mile for passenger vehicles on state
turnpikes. Draw a pie chart and a bar graph for the data.
State Indiana Oklahoma Florida Maine Pennsylvania
Number 2.9 4.3 6.0 3.8 5.8
4) The following data are based on a survey from American Travel Survey on why people
travel. Construct a pie chart a bar graph and a Pareto chart for the data and analyze the
results.
Purpose Personal business Visit friends or relatives Work-related Leisure
Number 146 330 225 299
5) The following percentages indicate the source of energy used worldwide. Construct a
Pareto chart and a vertical pie chart, a bar graph and a Pareto graph for the energy used.
Energy Type Petroleum Coal Dry natural gas Hydroelectric Nuclear Others
percentage 39.8 23.2 22.4 7.0 6.4 1.2
6) The following elements comprise the earth’s crust, the outermost solid layer. Illustrate the
composition of the earth’s crust with a pie chart and a bargraph for this data.
Element Oxygen Silicon Aluminum Iron Calcium Others

22
 percentage 45.6 27.3 8.4 6.2 4.7 7.8
7) The sales of recorded music in 2004 by genre are listed below. Represent the data with an
appropriate graph.
Rock Country Rap/hip R&B/u Pop Religious Children’s Jazz Classical Oldie Soundtrack New Othe
-hop rban s s age rs
23.9 13.0 12.1 11.3 10.0 6.0 2.8 2.7 2.0 1.4 1.1 1.0 8.9

8) The top 10 airlines with the most aircraft are listed. Represent these data with an
appropriate graph.
American Continental United Southwest Northwest American U.S. Lufthansa
Eagle Airways (Ger.)
714 364 603 327 424 245 384 233
9) The top prize-winning countries for Nobel Prizes in Physiology or Medicine are listed
here. Represent the data with an appropriate graph.
United Denmark United Austria Germany Belgium Sweden Italy Fran Aust Switzer
States Kingdom ce ralia land
80 5 24 4 16 4 8 3 7 3 6
10) Construct a histogram, frequency polygon, and an ogive for the distribution (shown here)
of the miles that 20 randomly selected runners ran during a given week.
Class 6-10 11-15 16 -20 21-25 26 - 30 31-35 36 -40
Freq 1 2 3 5 4 3 2
11) For 108 randomly selected college applicants, the following frequency distribution for
entrance exam scores was obtained. Construct a histogram, frequency polygon, and ogive
for the data.
Class 90-98 99-107 108-116 117125 126-134
Freq 6 22 43 28 9
Applicants who score above 107 need not enrol in a summer developmental program.
In this group, how many students do not have to enroll in the developmental program?
12) Thirty automobiles were tested for fuel efficiency, in miles per gallon (mpg). The
following frequency distribution was obtained. Construct a histogram, a frequency
polygon, and an ogive for the data.
Class 8-12 13-17 18-22 23-27 28-32
Freq 3 5 15 5 2
13) The salaries (in millions of dollars) for 31 NFL teams for a specific season are given in
this frequency distribution.
Class 39.9-42.8 42.9-45.8 45.9-48.8 48.9-51.8 51.9-54.8 54.9-57.8
Freq 2 2 5 5 12 5
Construct a histogram, a frequency polygon, and an ogive for the data; and comment on
the shape of the distribution.
14) In a study of reaction times of dogs to a specific stimulus, an animal trainer obtained the
following data, given in seconds. Construct a histogram, a frequency polygon, and an
ogive for the data; analyze the results.
Class 2.3-2.9 3.0-3.6 3.7-4.3 4.4-5.0 5.1-5.7 5.8-6.4
Freq 10 12 6 8 4 2
15) The animal trainer in question above selected another group of dogs who were much older
than the first group and measured their reaction times to the same stimulus. Construct a
histogram, a frequency polygon, and an ogive for the data.
Class 2.3-2.9 3.0-3.6 3.7-4.3 4.4-5.0 5.1-5.7 5.8-6.4

23
 Freq 1 3 4 16 14 4
Analyze the results and compare the histogram for this group with the one obtained in the
above question. Are there any differences in the histograms?
16) The frequency distributions shown indicate the percentages of public school students in
fourth-grade reading and mathematics who performed at or above the required proficiency
levels for the 50 states in the United States. Draw histograms for each, and decide if there
is any difference in the performance of the students in the subjects.
Class 18-22 23-27 28-32 33-37 38-42 43-48
Reading Freq 7 6 14 19 3 1
Math Freq 5 9 11 16 8 1
Using the histogram shown here, Construct a frequency distribution; include class limits,
class frequencies, midpoints, and cumulative frequencies. Hence answer these questions.
a) How many values are in the class 27.5–30.5?
b) How many values fall between 24.5 and 36.5?
c) How many values are below 33.5?
d) How many values are above 30.5?

3 NUMERICAL SUMMARIES
A numerical summary for a set of data is referred to as a statistic if the data set is a sample and
a parameter if the data set is the entire population.
Numerical summaries are categorized as measures of location and measures of spread.
Measures of location can further be classified into measures of central tendancy and measures
of relative positioning (quantiles).
3.1 Measures of Location
Before discussing the measures of location, its important to consider summation notation and
indexing
Index (subscript) Notation: Let the symbol xi (read ‘x sub t’i) denote any of the n values
x1 , x2 ,...., xn assumed by a variable X. The letter i in xi (i=1,2,. . . ,n) is called an index or
subscript. The letters j, k , p, q or s can also be used.
n
Summation Notation: x
i 1
i  x1  x2  ......  xn
n
Example: X Y
i 1
i i  X 1Y1  X 2Y2  ......  X N YN and

24
 n n

 aX i  aX 1  aX 2  ......  aX N  a X 1  X 2  ......  X N   a  X i
i 1 i 1

3.1.1 Measures of Central Tendency (Averages)

A Measures of Central Tendency of a set of numbers is a value which best represents it.
There are three different types of Central Tendencies namely the mean, median and
mode. Each has advantages and disadvantages depending on the data and intended purpose.

Arithmetic Mean
The arithmetic mean of a set of values x1 , x2 ,...., xn , denoted x if the data set is a sample, is
found by dividing the sum of the set of numbers with the actual number of values. Ie
1 n x  x  ......  xn
x 
n i 1
xi  1 2
n

Example 1 Find the mean of 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10.

Solution
Sum of values: 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 + 10 = 55
Number of values = 10
Mean of values x  10
55
 5.5

Note: If the numbers x1 , x2 ,...., xn occur f1 , f 2 ,...., f n times respectively, (occur with
frequencies f1 , f 2 ,...., f n ), the arithmetic mean is, given by
1 n f x  f 2 x2  ......  f n xn
x 
n i 1
f i xi  1 1
f1  f 2  ......  f n
where n is the total frequency. This is the formular for the mean of a grouped data.

Example 2 The grades of a student on six examinations were 84, 91, 72, 68, 91 and 72. Find
the arithmetic mean.

1 n 1(84)  2(91)  2(72)  1(68)

The arithmetic mean x  
n i 1
fi xi 
1 2  2 1
 79.67

Example 3 If 5, 8, 6 and 2 occur with frequencies 3, 2, 4 and 1 respectively, the arithmetic

1 n 3(5)  2(8)  4(6)  1(2)
mean is x   f i xi   5 .7
n i 1 3  2  4 1
Exercise
1 Find the mean of 9, 3, 4, 2, 1, 5, 8, 4, 7, 3
2 A sample of 5 executives received the following amount of bonus last year: sh 14,000, sh
15,000, sh 17,000, sh 16,000 and sh y. Find the value of y if the average bonus for the
5executives is sh 15,400

Properties of the Arithmetic Mean

25
 (1) The algebraic sum of the deviations of a set of numbers from their arithmetic mean is

 x  x   0 .
n
zero, that is i
i 1
n
(2)  x  a 
i 1
i
2
is minimum if and only if a  x .

(3) If n1 numbers have mean x1 , n2 numbers have mean x 2 ,…, nk numbers have mean
x k , then the mean of all the numbers called the combined mean is given by
n x1  n2 x 2  ...  nk x k  ni x i
xc  1 
n11  n2  ...  nk  ni
Median
It’s the value below which and above which half of the observations fall when ranked in order
of size. The position of the median term is given by  n2 1 th Value .
NB if n is even we average the middle 2 terms
For grouped data median is estimated using the formular
  n 1   Cf a 
Median = LCB   2   i
 f 
where LCB, f and i are the lower class boundary, frequency and class interval of the median
class. Cfa is the cumuilative frequency of the class above the median class.
Remark: The disadvantage of median is that it is not sensitive against changes in the data.

Mode
It’s the value occurring most frequently in a data set. If each observation occurs the same
number of times, then there is no mode. When 2 or more observation occurs mos frequently in
a data then the data is said to be multimodal.
For ungrouped data it’s very easy to pick out the mode. However If the data is grouped, mode
is estimated using the formular
 f  fa 
Mode = LCB     i
 2 f  f a  fb 
where LCB, f and i are the lower class boundary, frequency and class interval of the modal
class. fa and fb are frequencies of the class above and below the modal class. respectively

Example 1
Find the median and mode of the following data: 19, 13, 18, 14, 12, 25, 11, 10, 17, 23, 19.
Solution
Sorted data: 10, 11, 12, 13, 14, 17, 18, 19, 19, 23, 25.
n = 11 thus Median = 1121  Value  6th Value  17
th

Mode=19 since it appears most frequently in this data set as compared to other observations.

Example 2 Find the median and mode of the data: 2, 4, 8, 7, 9, 4, 6, 10, 8, and 5.
Solution

26
Array: 2, 4, 4, 5, 6, 7, 8, 8, 9, 10
67
n = 10 thus Median = 1021  Value  5.5th Value   6.5
th

2
Mode 4 and 8 ie the data is bimodal.
Example 3
Estimate the mean, median and mode for the following frequency distribution:
Class 5-9 10-14 15-19 20-24 25-29 30-34 35-39
Freq 5 12 32 40 16 9 6

Solution
Boundaries 4.5-9.5 9.5-14.5 14.5-19.5 19.5-24.5 24.5-29.5 29.5-34.5 34.5-39.5
Mid pts (x) 7 12 17 22 27 32 37
Frequency 5 12 32 40 16 9 6
Xf 35 144 544 880 432 288 222
CF 5 17 49 89 105 114 120

Mean x 
 fx  35  144  ...  222  2545  21.2083
n 120 120

n = 120 thus Median = 60.5th Value  Median class is 19.5-24.5 thus

  n1   Cf a   60.5  49 
Median = LCB   2   i  19.5     5  20.9375
 f   40 
The modal class (class with the highest frequency) is 19.5-24.5 therefore
 f  fa  40  32 
Mode = LCB     i = 19.5  5   20.75
 2 f  f a  f b   80  32  16 
Exercise
1. Find the mean median and mode for the following data: 9, 3, 4, 2, 9, 5, 8, 4, 7, 4
2. Find the mean median and mode of 1, 2, 2, 3, 4, 4, 5, 5, 5, 5, 7, 8, 8 and 9
3. The number of goals scored in 15 hockey matches is shown in the table.
No of goals 1 2 3 4 5
No of matches 2 1 5 3 4
Calculate the mean number of goals cored

4. The table shows the heights of 30 students in a class calculate an estimate of the mean mode
and median height.
Height (cm) 140<x<144 144<x<148 148<x<152 152<x<156 156<x<160 160<x<164
No of students 4 5 8 7 5 1
5. Estimate the mean, median and mode for the following frequency distribution:
Class 1-4 5-8 9-12 13-16 17-20 21-24
frequency 10 14 20 16 12 8

27
 Class 40-59 60-79 80-99 100-119 120-139 140-159 160-179 180-199
freq 5 12 28 40 16 9 6 4

The Empirical Relation between the Mean, Median and Mode

MEAN  MODE  3  MEAN  MEDIAN 
The above relation is true for unimodal frequency curves which are asymmetrical.

3.1.2 Other Types of Means

These will include weighted, harmonic and geometric means.

The Weighted Arithmetic Mean

The weighted arithmetic mean of a set of n numbers x1 , x2 ,...., xn having corresponding weights
w1 , w2 ,...., wn is defined as
w1 x1  w2 x2  ...  wn xn  wi x i
xw  
w11  w2  ...  wn  wi
Example1 Consider the following table with marks obtained by two students James (mark x)
and John (mark y). The weights are to be used in determining who joins the engineering course
whose requirement is a weighted mean of 58% on the four subjects below;
Subject Maths English History Physics Total
Mark x 25 87 83 30 225
Mark y 70 45 35 75 225
Weight 3.6 2.3 1.5 2.6 10

Working the products of the marks and the weights we get

Subject Maths English History Physics Total
Wx 90 200.1 124.5 78 492.6
Wy 252 103.5 52.5 195 603

Now x w 
 wi xi  492.6  49.26 and y   wi y i  603  60.3
 wi 10 w
 wi 10
Clearly John qualifies but James does not.

Example 2 If a final examination is weighted 4 times as much as a quiz, a midterm examination

3 times as much as a quiz, and a student has a final examination grade of 80, a midterm
examination grade
of 95 and quiz grades of 90, 65 and 70, the mean grade is
1(90)+ 1(65)+ 1(70)+ 3(95)+ 4 (80) 830
X= = = 83 .
1+ 1+ 1+ 3 + 4 10

Question A tycoom has 3 house girls who he pays Ksh 4,000 each per month, 2 watch men
who he pays Ksh 5,000 each and some garden men who receives Ksh 7,000 each. If he pays
out an average of Ksh 5,700 per month to these people, find the number of garden men.

28
Question A student’s grades in laboratory, lecture, and recitation parts of a computer course
were 71, 78, and 89, respectively.
(a) If the weights accorded these grades are 2,4, and 5, respectively, what is an average grade?
(b) What is the average grade if equal weights are used?

The Geometric and Harmonic Means

Let x1 , x2 ,...., xn be the sample values, the geometric mean GM is given by
n
GM  n x1  x2  .... xn  n x i 1
i

and the harmonic mean is given by

n n
HM  
 
.
1
x1  1
x2  ....  x1n 1
xi

The Relation between the Arithmetic, Geometric and Harmonic Means:

HM  GM  x .

The formulas for geometric and harmonic means of a frequency distribution are
respectively given by;
n
1 n
GM  n x 1f1  x 2f 2  .... x nf n  n  x ifi  log GM  
i 1
 f i log xi
n i 1
and
HM 
n

 
n
    1 fi 1

  ....  
f1 f2 fn fi n xi
x1 x2 xn xi

where n   f i and xi are the midpoints

Example 1: Find the harmonic and the geometric mean of the numbers 2,4 and 8
Solution The geometric mean GM  3 2  4  8  3 64  4 and
3 3 34
the harmonic mean HM  1 1 1  7   3.43
2  4  8 8 7
Example 2:
Find the harmonic and geometric mean of the frequency table below
x 13 14 15 16 17
f 2 5 13 7 3
Solution
30
The harmonic mean HM  2 5 13 7 3  15. and
13  14  15  16  17

The geometric mean GM  30 132  145  1513  167  173  15.09837

Exercise:
1. Find the harmonic and the geometric mean of the numbers 10, 12, 15, 5 and 8
2. The number of goals scored in 15 hockey matches is shown in the table below .
Calculate the harmonic and geometric mean number of goals scored.
29
 No of goals 1 3 5 6 9
No of matches 2 1 5 3 4
3. Find the harmonic and geometric mean of the frequency table below
Class 0-29 30-49 50-79 80-99
Frequency 20 30 40 10

3.1.3 Measures of Relative Positioning (Quantiles)

These are values which divide a sorted data set into N equal parts. They are also known as
quantiles or N-tiles. The commonly used quantiles are; Quartiles, Deciles and Percentiles
These 3 divides a sorted data set into four, ten and hundred divisions, respectively. These
measures of position are useful for comparing scores within one set of data. You probably all
took some type of college placement exam at some point. If your composite math score was
say 28, it might have been reported that this score was in the 94th percentile. What does this
mean? This does not mean you received a 94% on the test. It does mean that of all the
students who took that exam, 94% of them scored lower than you did (and 6% higher).

Remark For a set of data you can divide the data into three quartiles ( Q1 , Q2 , Q3 ), nine deciles
( D1 , D2 ,...D9 ) and 99 percentiles ( P1 , P2 ,...., P99 ).To work with percentiles, deciles and
quartiles - you need to learn to do two different tasks. First you should learn how to find the
percentile that corresponds to a particular score and then how to find the score in a set of data
that corresponds to a given percentile.

Quartiles
They divide a sorted data set into 4 equal parts and we have lower, middle and upper quartiles
denoted Q1, Q2 and Q3 respectively. The lower quartile Q1 separates the bottom 25% from the
top 75%, Q2 is the median and Q3 separates the top 25% from the bottom 75% as illustrated
below .

The Kth quartile is given by: Qk  k4 (n  1)th value where k=1,2,3

Deciles and Percentiles

Similarly the Kth Deciles Dk and the Kth Percentiles Pk are respectively given by;
Dk  10k (n  1)th Value and Pk  100
k
(n  1)th value
NB For ungrouped data we may be forced to use linear interpolation for us to get the required
Kth quantile. However for grouped data the Kth Value is given by
 K  Cf a 
K thValue  LCB     i
 f 
where LCB, i and f are the lower class boundary. class interval and frequency of the class
containing the K th value. Cfa is the cumuilative frequency of the class above this particular
class

30
Example 1: Find the lower and upper quartiles, the 7th decile and the 85th percentile of the
following data. 3, 6, 9, 10, 7, 12, 13, 15, 6, 5, 13

Solution
Sorted data: 3, 5, 6, 6, 7, 9, 10, 12, 13, 13, 15 Here n=11
Q1  14 (11  1)th  3rd value  6 Similarly Q3  34 (11  1)th  9th value  13
D7  107 (11  1)th  7.7th value  7th value  0.7(8th value  7th value)  10  0.7(12  10)  11.4

linear interpolat ion

P85  100
85
(11  1)th  10.2th value  10th value  0.2(11th value  10th value)  13  0.2(15  13)  13.4
  
linear interpolat ion

Example 2:
Estimate the lower quartile, 4th decile and the 72nd percentile for the frequency table below
Class 1-4 5-8 9-12 13-16 17-20 21-24
frequency 10 14 20 16 12 8
Solution
Boundaries 0.5-4.5 4.5-8.5 8.5-12.5 12.5-16.5 16.5-20.5 20.5-24.5
C.F 10 24 44 60 72 80
For this data n=80
 20.25  10 
Q1  14 (80  1) th  20.25th value  4.5     4  7.428571
 14 
 32.4  24 
D4  104 (80  1) th  32.4th value  8.5     4  10.18
 20 
 58.32  44 
P72  100
72
(80  1) th  58.32 th value  12.5     4  16.08
 16 
Exercise
a) Find the lower and upper quartiles, the 7th decile and the 85th percentile of the data.
a) 9, 3, 4, 2, 9, 5, 8, 4, 7, 4 b) 1, 2, 2, 3, 4, 4, 5, 5, 5, 5, 7, 8, 8 and 9
2) The number of goals scored in 15 hockey matches is shown in the table.
No of goals 1 2 3 4 5
No of matches 2 1 5 3 4
Estimate the lower quartile, 4 decile and the 72nd percentile of the number of goals cored
th

4) The table shows the heights of 30 students in a class calculate an estimate of the upper and
lower quartile of the height.
Height (cm) 140<x<144 144<x<148 148<x<152 152<x<156 156<x<160 160<x<164
No of students 4 5 8 7 5 1
5) The grouped frequency table gives information about the distance each of 150 people travel to
work.
Height (cm) 0<d<5 5<d<10 10<d<15 15<d<20 20<d<25 25<d<30
No of students 4 5 8 7 5 1
a) Work out what percentage of the 150 (b) Calculate an estimate for the median
people travel more than 20 km to work distance travelled to work by the people?

31
Properties of measures of Location
(i) They are affected by change of origin. Adding or subtracting a constant from
each and every observation in a data set causes all the measures of location to
shift by the same magnitude. That is New measure  old measure  k
(ii) They are affected by change of scale. Multiplying each and every observation
in a data set by a constant value scales up all the measures of location by the
same magnitude.. That is New measure  K (old measure )

Example: Consider the three sets of data A, B and C below

Set A: 65, 53, 42, 52, 53 x A  53 and Median A  53
Set B: 15, 3, -8, 2, 3 x B  3 and Median B  3
Set C: 45, 9, -24, 6, 9 x C  9 and Median C  9
 Notice that set B is obtained by subtracting 50 from each and every observation in set A
and clearly x B  x A  50 and Median B  Median A  50 Therefore
New measure  old measure  k. This is referred to as change of origin.
 Effectively set C is obtained by multiplying each and every observation in set B by 3 and
clearly x C  3 x B and Median C  3Median B Thus New measure  K (old measure ) This
is referred to as change of scale.

3.2 Measures of Spread/ Dispersion

Spread is the degree of scatter or variation of the variable about the central value.
Examples of these measures includes: the range, Inter-Quartile range, Quartile
Deviation also called semi Inter-Quartile range, Mean Absolute Deviation,
Variance and standard deviation.

Inter-Quartile range and Semi Inter-Quartile Range

Inter-Quartile range (IQR) is the difference between the upper and lower quartiles. Half of
this difference is called Quartile Deviation or the semi Inter-Quartile range (SIQR) Ie
IQR Q 3 -Q1 and SIQR  12 Q 3 -Q1 

Mean Absolute Deviation (MAD)

It is the average of the absolute deviations from the mean and it’s given by

MAD =
 x-x for ungrouped data but for grouped data MAD =
f x - x
n n

Example 1:
Find the quartile deviation and the mean absolute deviation for the following data.
3, 6, 9, 10, 7, 12, 13, 15, 6, 5, 13
Solution
Sorted data: 3, 5, 6, 6, 7, 9, 10, 12, 13, 13, 15
Recall Q1  6 and Q3  13 ie from earlier calculations.
Thus SIQR  1
2
Q3 -Q1   12 (13  6)  3.5

32
 3  5  6  6  7  9  10  12  13  13  15
x 9
11

MAD =
 
 x - x  3 - 9  5 - 9  6 - 9  ...  15 - 9  6  4  3  ...  6  36  3.2727
n 11 11 11

Variance and Standard Deviation

Ignoring the negative sign in order to compute MAD is not the only option we have to deal
with deviations. We can square the deviations and then average. The average of the squared
deviations from the mean is called the variance denoted s 2 and its given by

 (x - x) x
2
s2 = 1
n
2
A little algebraic simplification of this formular gives s 2 = 1
n
2
x

 f (x - x)  fx
2
For grouped data s 2 = 1
n
2
 1
n
2
 x where n is the sum of frequencies.
To reverse the squaring on the units we find the square root of the variance.Standard Deviation
denoted s is the square root of variance.

Example 1: Find the variance and standard deviation for the data.
3, 6, 9, 10, 7, 12, 13, 15, 6, 5, 13
Solution
3  5  6  6  7  9  10  12  13  13  15
x 9
11
 x - x   3 - 9  5 - 9  6 - 9  ...  15 - 9
2
36  16  9  ...  36 143
2 2 2 2
S 2
=    13
n 11 11 11
Standard deviation s = variance   13  3.60555 .

Example 2 Find the standard deviation of the data: 2, 4, 8, 7, 9, 4, 6, 10, 8, and 5.

Solution
 x  2  4  8  ...  5  63  6.3 and
Mean x 
n 10 10
x 2
 2 2  42  82  ...  52  455

x
2
Standard deviation s = 1
n
2
 x  45.5  6.32  2.4104 .
Example 3 Estimate the mean, and standard deviation for the frequency table below:
Class 5-9 10-14 15-19 20-24 25-29 30-34 35-39
freq 5 12 32 40 16 9 6

Solution
Mid pts (x) 7 12 17 22 27 32 37 sum
Freq (f) 5 12 32 40 16 9 6 120
xf 35 144 544 880 432 288 222 2545
fx 2 245 1728 9248 19360 11664 9216 8214 59675

 fx  2545  21.2083
Mean x 
n 120
and  fx 2
 59675

33
 59675
 fx
2
Standard deviation s = 1
n
2
x   21.20832  6.8919 .
120

Exercise
1) Find the quartile deviation, the mean absolute deviation and the standard deviation of the
data: a) 9, 3, 4, 2, 9, 5, 8, 4, 7, 4 b) 1, 2, 2, 3, 4, 4, 5, 5, 5, 5, 7, 8, 8 and 9
2) 2) The number of goals scored in 20 hockey matches is shown in the table.
No of goals 1 2 3 4 5
No of matches 2 5 6 3 4
Estimate the quartile deviation, the mean absolute deviation and the standard deviation of
the number of goals cored
3) consider the frequency table below and estimate quartile deviation, the mean absolute
deviation and the standard deviation
Class 8-12 13-17 18-22 23-27 28-32 33-37
Freq 3 10 12 9 5 1
4) The table shows the heights of 30 students in a class calculate an estimate of the quartile
deviation, the mean absolute deviation and the standard deviation of the height.
Height (cm) 140<x<144 144<x<148 148<x<152 152<x<156 156<x<160 160<x<164
No of students 4 5 8 7 5 1

5) The grouped frequency table gives information about the distance each of 150 people travel
to work. Calculate an estimate for the quartile deviation and the standard deviation of the
distance travelled to work by the people
Height (cm) 0<d<5 5<d<10 10<d<15 15<d<20 20<d<25 25<d<30
# of students 4 5 8 7 5 1
.
Properties of measures of Spread
i) They are not affected by change of origin. Adding or subtracting a constant
from each and every observation in a data set does not affect any measures of
spread. That is New measure  old measure
iii) They are affected by change of scale. Multiplying each and every observation
in a data set by a constant value scales up all the measures of spread by the
same value except in the case of variance which is scaled up by a square of the
same constant.
ie New measure  K (old measure ) but New variance = k 2  old variance

Example: Consider the three sets of data A, B and C below

Set A: 65, 53, 42, 52, 53 Range=23, MAD A  4.8 and Variance A  66.5
Set B: 15, 3, -8, 2, 3 Range=23, MAD B  4.8 and Variance B  66.5
Set C: 45, 9, -24, 6, 9 Range=69, MAD C  14.4 and VarianceC  598.5
 Notice that set B is obtained by subtracting 50 from each and every observation in set A
and clearly MAD B  MAD A and Variance B  Variance A Therefore there is no effect on
the change of origin ie New measure  old measure . .

34
 Effectively set C is obtained by multiplying each and every observation in set B by 3 and
clearly MAD C  3  MAD B and VarianceC  32  VarianceB Thus
New measure  K (old measure ) and New VarianceC  k 2  old VarianceB

Mean and Standard Deviation Using a Calculator

 When on, press mode key to get;
COMP SD REG
1 2 3
 Press 2 to select SD for statistical data.
 Enter data one by one pressing m+ after every value entered. The screen will
be showing the number of observations that are fully entered.
 Pressing shift then 1 gives  Pressing shift then 2 gives
x 2
x n x x n x n 1
1 2 3 1 2 3
Typing 1 then = gives the value of x 2 Which are the mean uncorrected
standard deviation and the corrected
Similarly typing 2 then = gives the standard deviation
value of  x

Example using your calculator, obtain the mean and standard deviation of the following
data: 31, 52, 29, 60, 58
Solution
Entering data 31M+ 52 M+ 29 M+ 60 M+ 58 M+
x  46 and s  14.91643

Question Redo the above example using the data: 235, 693, 484, 118, 470

3.3 Assumed Mean and the Coding Formular

If the observations are too large such that the natural computation of totals is tedious,
we can take one of the observations as the working/assumed mean. Let A be any
guessed or assumed arithmetic mean and let d i =x i -A be the deviations of x i from A,
then
mean x  A  1n  fd  A  d
and
 fd    fd 
 1n  fd 2  d
2 2
Variance S 
2 1 2 1
n n

Respectably where A = Assumed mean which is generally taken as mid point of the
middle class or the class where frequency is large
Remark:, in most cases deviations (d) of x i from A is a multiple of the class interval ie
d x -A
di = t i  i  t i  i  i .
i i
In these cases we can use t rather than d in computation. The above formulae reduces to
x  A  ni  ft  A  it and
S 2  i2   ft
1
n
2
   ft
 1n  ft   i 2
2 1
n
2
t
2

35
respectably the latter formulae are referred to as coding formulae

Example
Using coding formulae, find the mean and standard deviation of the following data
Class 340-349 350-359 360-369 370-379 380-389
Freq 2 3 7 5 3

Solution
x - 364.5
Class Mid pts Freq t
i
ft ft 2
340-349 344.5 2 -2 -4 8
350-359 354.5 3 -1 -3 3
360-369 364.5 7 0 0 0
370-379 374.5 5 1 5 5
380-389 384.5 3 2 6 12
Total 20 4 28
x  A  ni  ft  364.5  10
20 ( 4)  366.5

 ft    ft    204   11.6619
2
S  i  10 
1 2 1 28 2
n n 20

Exercise
1) Consider the following frequency distribution.
classes 10-14 15-19 20-24 25-29 30-34
frequency 7 11 14 13 5
Estimate the mean and standard deviation using coding formula
2) Using coding formular, find the mean and standard deviation of the frequency table
below
Class 10-20- 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100 100-110
Freq 4 5 7 13 16 11 9 6 4 3
3) The table shows the speed distribution of vehicles on Thika Super high way on a typical
day.
Speed (km/hr) 60-69 70-79 80-89 90-99 100-109 110-119 120-129 130-139 140-149
No of vehicles 138 163 325 541 427 214 110 52 30
Using coding formulae, find the mean speed and the standard deviation of the speeds.
4) The following table shows a frequency distribution of the weekly wages of 65 employees
at the P&R Company.
Wages $250.00 $260.00 $270.00- $280.00- $290.00- $300.00- $310.00-
-259.99 -269.99 279.99 289.99 299.99 309.99 319.99
No. of employees 8 10 16 14 10 5 2
Find the mean wage and the standard deviation of the wages using coding formular

3.4 Measures of Relative Dispersion:

These measures are used in comparing spreads of two or more sets of observations. These
measures are independent of the units of measurement. These are a sort of ratio and are called
coefficients.

36
Suppose that the two distributions to be compared are expressed in the same units and their
means are equal or nearly equal. Then their variability can be compared directly by using
their standard deviations. However, if their means are widely different or if they are expressed
in different units of measurement, we can not use the standard deviations as such for
comparing their variability. We have to use the relative measures of dispersion in such
situations. Examples of these Measures of relative dispersion includes; Coefficient of quartile
deviation, Coefficient of mean deviation and the Coefficient of variation

3.4.1 Coefficient of Quartile Deviation and Coefficient of Mean Deviation

Q3  Q1
The Coefficient of Quartile Deviation of x CQD(x) is given by CQD(x)   100%
Q3  Q1
MAD
The Coefficient of Mean Deviation CMD(x) is given by CMD(x)   100%
Mean

3.4.2 Coefficient of Variation:

Coefficient of variation is the percentage ratio of standard deviation and the arithmetic mean.
It is usually expressed in percentage. The coefficient of variation of x denoted C.V(x) is given
by the formula
C.V(x) = Sx  100%
where x is the mean and S is the standard deviation of x.
The coefficient has no units ie it’s independent of the units of measurements. It is useful in
comparing spreads of two or more populations. The smaller the coefficient of variation, the
higher the peak and the lower the spread.
Note: Standard deviation is absolute measure of dispersion while. Coefficient of variation is
relative measure of dispersion.

Example 1: Consider the distribution of the yields (per plot) of two ground nut varieties. For
the first variety, the mean and standard deviation are 82 kg and 16 kg respectively. For the
second variety, the mean and standard deviation are 55 kg and 8 kg respectively. Then we
have, for the first variety
C.V(x) = 16
82
 100  19.5%
For the second variety
C.V(x) = 558  100  14.5%
It is apparent that the variability in second variety is less as compared to that in the first
variety. But in terms of standard deviation the interpretation could be reverse.

Example 2: Below are the scores of two cricketers in 10 innings. Find who is more
„consistent scorer‟ by Indirect method.
A 204 68 150 30 70 95 60 76 24 19
B 99 190 130 94 80 89 69 85 65 40

Solution:
From a calculator, x A  79.6 , S A  58.2 x B  94.1 and S B  41.1
Coefficient of variation for player A is C.V(x) = 58.2
79.6
 100  73.153%
Coefficient of variation for player B is C.V(x) = 41.1
94.1
 100  43.7028%

37
Coefficient of variation of A is greater than coefficient of variation of B and hence we
conclude that player B is more consistent

Exercise
1) Find the coefficient of quartile deviation, the coefficient of mean deviation and the
Coefficient of variation n of x for the following data: a) 9, 3, 4, 2, 9, 5, 8, 4, 7, 4
b) 1, 2, 2, 3, 4, 4, 5, 5, 6, 6, 7, 8, 8 and 9 c) 3, 6, 9, 10, 7, 12, 13, 15, 6, 5, 13
d) data on marks given by the table below
Marks Obtained 0-10 10-20 20-30 30-40 40-50 50-60 60-70
No. of Students 6 12 22 24 16 12 8
2) If the weights of 7 ear-heads of sorghum are 89, 94, 102, 107, 108, 115 and 126 g. Find
the arithmetic mean and standard deviation using a calculator hence determine the
coefficient of variation of the ear-heads of sorghum
3) The following are the 381soybean plant heights in Cms collected from a particular plot.
Using coding formula, Find the mean and Standard deviation of the plants hence
determine the coefficient of variation of the 1soybean plant heights:
Plant heights 6.8- 7.3- 7.8- 8.3- 8.8- 9.3- 9.8- 10.3- 10.8- 11.3- 11.8- 12.3-
(Cms) 7.2 7.7 8.2 8.7 9.2 9.7 10.2 10.7 11.2 11.7 12.2 12.7
No. of Plants 9 10 11 32 42 58 65 55 37 31 24 7

3.5 Measures of Skewness and Kurtosis

3.5.1 Skewness
Before discussing the concept of skewness, an understanding of the concept of symmetry is
essential. A plot of frequency against class mark joined with a smooth curve can help
us to visually assess the symmetry of a distribution. Usually symmetry is about the
central value. Symmetry is said to exist in a distribution if the smoothed frequency polygon of
the distribution can be divided into two identical halves wherein each half is a mirror image of the other
Skewness on the other hand means lack of symmetry and it can be positive or negative.
Basically, if the distribution has a tail on the right, (See figure below), then the
distribution is positively skewed Eg Most students having vey low marks in an
examination. However if the distribution has a tail on the left, then the distribution is
negatively skewed. (see figure below). Eg Most students having vey high marks in an
examination

Measures of Skewness

38
Generally for any set of values x1 , x2 , x3 , ..... xn , the moment coefficient of skewness  3 is

 f x - x 
3

given by  3 = where S is the standard deviation of X. It’s worth noting that if

nS 3
 3  0, the distribution is negatively skewed, if  3  0, the distribution is positively skewed
and if  3  0 the distribution is normal

Other measures of Skewness includes the Karl Pearson coefficient of Skewness SK p ,

Bowley’s coefficient of Skewness SK B and Kelley’s coefficient of Skewness SK k .
The Karl Pearson’s coefficient of Skewness is based upon the divergence of mean from
mob in a skewed distribution. Recall the empirical relation between mean, median and
mode which states that, for a moderately symmetrical distribution, we have
Mean - Mode = 3 (Mean - Median)
Hence Karl Pearson's coefficient of skewness is defined by;
Mean  Mode 3Mean  Median 
SK p   ,
Standard Deviation Standard Deviation
The Bowley’s coefficient of Skewness is based on quartiles. For a symmetrical
distribution, it is seen that Q1 and Q 3 are equidistant ftom median.
Q3 - 2Q 2  Q1
SK B  where Q k is the Kth quartile.
Q3 - Q1
The Kelly’s coefficient of Skewness is based on P90 and, P10 so that only 10% of the
observations on each extreme are ignored.. This is an improvement over the Bowley’s
coefficient which leaves 25% of the observatories on each extreme of the distribution.
P - 2P50  P10
SK k  90 where Pk is the Kth percentile.
P90 - P10
Interpreting Skewness
If the coefficient of skewness is positive, the data are positively skewed or skewed right,
meaning that the right tail of the distribution is longer than the left. If the coefficient of
skewness is negative, the data are negatively skewed or skewed left, meaning that the left tail
is longer. If the coefficient of skewness = 0, the data are perfectly symmetrical. But a
skewness of exactly zero is quite unlikely for real-world data, so how can you interpret the
skewness number? Bulmer, M. G., Principles of Statistics (Dover,1979) — a classic —
suggests this rule of thumb: If the coefficient of skewness is:-
 less than −1 or greater than +1, the distribution is highly skewed.
 between −1 and  12 . or between  12 . and +1, the distribution is moderately skewed.
 between  12 and  12 .., the distribution is approximately symmetric.

Example: The following figures relate to the size of capital of 285 companies :
Capital (in Ks lacs.) 1-5 6-10 11-15 16-20 21-25 26-30 31-35
No. of companies 20 27 29 38 48 53 70
Compute the Bowley's coefficients of skewness and interpret the results.
Solution
Boundaries 0.5-5.5 5.5-10.5 10.5-15.5 15.5-20.5 20.5-25.5 25.5-30.5 30.5-35.5
CF 20 47 76 114 162 215 285
39
  71.5 - 47 
Q1  14 (286) th value  71.5 th value  10.5     5  14.7241
 29 
 143 - 114 
Q 2  12 (286) th value  143rd value  20.5     5  23.5208
 48 
 214.5 - 162 
Q3  34 (286) th value  214.5th value  25.5     5  30.4528
 53 
Q - 2Q 2  Q1 30.4528 - 2  23.5208  14.7241
SK p  3   -0.11855.
Q 3 - Q1 30.4528 - 14.72411
This value lies between  12 and  12 , therefore the distribution is approximately symmetric.
Question: Compute the Karl Pearson's and the Kelly’s coefficient of skewness for the
above data and interpret the results.

3.5.2 Kurtosis
It measures the peakedness of a distribution. If the values of x are very close to the mean, the
peak is very high and the distribution is said to be Leptokurtic. On the other hand if the
values of x are very far away from the mean, the peak is very low and the distribution is said
to be Pletykurtic. Finally if x values are at a moderate distance from the mean then the peak
is moderate and the distribution is said to be mesokurtic..

Measures of Kurtosis
Generally for a set of values x1 , x2 , x3 , ..... xn , the moment coefficient of kurtosis  4 is given

 f x - x 
4

by  4 = where x and S are the arithmetic mean and standard deviation of X.

nS 4
Example: Calculate the coefficient of Skewness  3 and the coefficient of kurtosis  4 for the
data 5, 6, 7, 6, 9, 4, 5
Solution
x= 1
n x  42
7
 6 and Standard deviation s = 1
n  (x - x) 2
 4
7

x 5 6 7 6 9 4 5 Sum
x - x  2 1 0 1 0 9 4 1
16
x - x  3 -1 0 1 0 27 -8 -1
18

40
 x - x 
4 1 0 1 0 81 16 1 100

Coefficient of Skewness  3 =
x - x  3


18
    0.744118
7
3
3 4
nS 7
f x - x 
=
4

Coefficient of kurtosis  4
nS 4

100
7
    2.73438
4
7
4

Exercise
1. Find the moment coefficient of Skewness and kurtosis for the dat below. a) 9, 3, 4, 2, 9, 5,
8, 4, 7, 4 b) 1, 2, 2, 3, 4, 4, 5, 5, 6, 6, 7, 8, 8 and 9 c) 3, 6, 9, 10, 7, 12, 13, 15, 6, 5, 13
d) data on marks given by the table below
Marks Obtained 0-10 10-20 20-30 30-40 40-50 50-60 60-70
No. of Students 6 12 22 24 16 12 8
e) Data given by the table below
Marks Obtained 0-10 10-20 20-30 30-40
No. of Students 1 3 4 2
2. Compute the Bowley's coefficient of skewness, the Kelly’s coefficient of skewness
and the Percentile coefficient of kurtosis for the following data and interpret the results.
a) 9, 3, 4, 2, 9, 5, 8, 4, 7, 4 b) 1, 2, 2, 3, 4, 4, 5, 5, 6, 6, 7, 8, 8 and 9
c) 3, 6, 9, 10, 7, 12, 13, 15, 6, 5, 13 d) data on heights given by the table below
Heightl (in inches.) 58 59 60 61 62 63 64 65
No. of persons 10 18 30 42 35 28 16 8
e) data on daily expenditure of families given by the table below
Daily Expenditure (Rs) 0-20 20-40 40-60 60-80 80-100
No. of persons 13 25 27 19 16
f) Data on marks given by the table below
Marks Obtained 0-20 20-40 40-60 60-80 80-100
No. of Students 8 28 35 17 12
3. The following measures were computed for a frequency distribution : Mean = 50,
coefficient of Variation = 35% and Karl Pearson's Coefficient of Skewness
SK p  - 0.25 . Compute Standard Deviation, Mode and Median of the distribution.

4. Bivariate Data
4.1 Introduction
So far we have confined our discussion to the distributions involving only one
variable. Sometimes, in practical applications, we might come across certain set of data,
where each item of the set may comprise of the values of two or more variables.
A Bivariate Data is a a set of paired measurements which are of the form
(x1 , y1 ) , (x 2 , y 2 ), ....., (x n , y n )
Examples
i. Marks obtained in two subjects by 60 students in a class.
ii. The series of sales revenue and advertising expenditure of the various branches of a
company in a particular year.
iii. The series of ages of husbands and wives in a sample of selected married couples.

41
 In a bivariate data, each pair represents the values of the two variables. Our interest is to
find a relationship (if it exists) between the two variables under study.

4.2 Scatter Diagrams and Correlation

A scatter diagram is a tool for analyzing relationships between two variables. One
variable is plotted on the horizontal axis and the other is plotted on the vertical axis. The
pattern of their intersecting points can graphically show relationship patterns. Most often a
scatter diagram is used to prove or disprove cause-and-effect relationships. While the diagram
shows relationships, it does not by itself prove that one variable causes the other. In brief, the
easiest way to visualize Bivariate Data is through a Scatter Plot.
“Two variables are said to be correlated if the change in one of the variables results in a
change in the other variable”.

4.2.1: Positive and Negative Correlation

If the values of the two variables deviate in the same direction i.e. if an increase (or
decrease) in the values of one variable results, on an average, in a corresponding increase (or
decrease) in the values of the other variable the correlation is said to be positive.
Some examples of series of positive correlation are:
i. Heights and weights;
ii. Household income and expenditure;
iii. Price and supply of commodities;
iv. Amount of rainfall and yield of crops.

Correlation between two variables is said to be negative or inverse if the variables deviate
in opposite direction. That is, if the increase in the variables deviate in opposite direction.
That is, if increase (or decrease) in the values of one variable results on an average, in
corresponding decrease (or increase) in the values of other variable.
Eg Price and demand of goods.

42
4.2.2 Interpreting a Scatter Plot
Scatter diagrams will generally show one of six possible correlations between the
variables:
i. Strong Positive Correlation The value of Y clearly increases as the value of X
increases.
ii. Strong Negative Correlation The value of Y clearly decreases as the value of X
increases.
iii. Weak Positive Correlation The value of Y increases slightly as the value of X
increases.
iv. Weak Negative Correlation The value of Y decreases slightly as the value of X
increases.
v. Complex Correlation The value of Y seems to be related to the value of X, but the
relationship is not easily determined.
vi. No Correlation There is no demonstrated connection between the two variables

4,3 Correlation Coefficient

Correlation coefficient measures the degree of linear association between 2 paired
variables It takes values from + 1 to – 1.
i. If r = +1,we have perfect positive relationship
ii. If r = -1,we have perfect negative relationship
iii. If r = 0 there is no relationship ie the variables are uncorrelated.

4,3 .1 Pearson's Product Moment Correlation Coefficient

Pearson's product moment correlation coefficient, usually denoted by r, is one example of a
correlation coefficient. It is a measure of the linear association between two variables that
have been measured on interval or ratio scales, such as the relationship between height in
inches and weight in pounds. However, it can be misleadingly small when there is a
relationship between the variables but it is a non-linear one.
43
 n xy   x y
The correlation coefficient r is given by r 
[n x 2  ( x) 2 ][n y 2  ( y ) 2 ]
Example:
: A study was conducted to find whether there is any relationship between the weight and
blood pressure of an individual. The following set of data was arrived at from a clinical study.
Let us determine the coefficient of correlation for this set of data. The first column represents
the serial number and the second and third columns represent the weight and blood pressure
of each patient.
Weight 78 86 72 822 80 86 84 89 68 71
Blood Pressure 140 160 134 144 180 176 174 178 128 132

Thus
10(124206) - (796) (1546) 11444
r   0.5966
[ (10)63776 - (796) 2 (10) ][(243036) - (1546) 2 ] (1144) (40244)

It can be shown that r   ( x  x )( y  y )

 ( x  x)  ( y  y )
2 2

Example:
Obtain the correlation coefficient of the following data
Mean Temp. (x) 14.2 14.3 14.6 14.9 15.2 15.6 15.9
Pirates (y) 35000 45000 20000 15000 5000 400 17

Solution

44
 - 62583
We then have that r   0 : 93
2.5(1828695447 )

4.3 .2 Spearman rank correlation coefficient

Data which are arranged in ascending order are said to be in ranks or ranked data.. The
coefficient of correlation for such type of data is given by Spearman rank difference
correlation coefficient and is denoted by R.
6 d
2

R is given by the formula R  1 

n(n 2  1)

Example
The data given below are obtained from student records.( Grade Point Average (x) and
Graduate Record exam score (y)) Calculate the rank correlation coefficient ‘R’ for the data.
Subject 1 2 3 4 5 6 7 8 9 10
X 8.3 8.6 9.2 9.8 8.0 7.8 9.4 9.0 7.2 8.6
y 2300 2250 2380 2400 2000 2100 2360 2350 2000 2260
Solution
Note that in the x row, we have two students having a grade point average of 8.6 also in the y
row; there is a tie for 2000.
Now we arrange the data in descending order and then rank 1,2,3,. . . . .10 accordingly. In
case of a tie, the rank of each tied value is the mean of all positions they occupy. In x, for
56
instance, 8.6 occupy ranks 5 and 6. So each has a rank  5.5
2
Similarly in ‘y’ 2000 occupies ranks 9 and 10, so each has rank 9.5
6 d
2

Now we come back to our formula R  1 

n(n 2  1)
We compute d, square it and substitute its value in the formula

45
So here, n = 10 and  d  12 . So
2

6(12)
R 1  1  0.0727  0.9273
10(100  1)
Note: If we are provided with only ranks without giving the values of x and y we can still find
Spearman rank difference correlation R by taking the difference of the ranks and proceeding
in the above shown manner.

4,4 Regression
If two variables are significantly correlated, and if there is some theoretical basis for
doing so, it is possible to predict values of one variable from the other.
Regression analysis, in general sense, means the estimation or prediction of the
unknown value of one variable from the known value of the other variable. It is one of the
most important statistical tools which is extensively used in almost all sciences – Natural,
Social and Physical.
Regression analysis was explained by M. M. Blair as follows:
“Regression analysis is a mathematical measure of the average relationship between two or
more variables in terms of the original units of the data.”

3.4.1 Regression Equation

Regression analysis can be thought of as being sort of like the flip side of correlation.
It has to do with finding the equation for the kind of straight lines you were just looking at
Suppose we have a sample of size n and it has two sets of measures, denoted by x and y. We
can predict the values of y given the values of x by using the equation, y *  a  bx
Where the coefficients ‘a’ and ‘b’ are real numbers given by
n xy   x y  y  b x
b and a
n x 2  ( x) 2 n
The symbol y * refers to the predicted value of y from a given value of x from the regression
equation.

Example:
46
Scores made by students in a statistics class in the mid-term and final examination are given
here. Develop a regression equation which may be used to predict final examination scores
from the mid – term score.
Student 1 2 3 4 5 6 7 8 9 10
Mid term 98 66 100 96 88 45 76 60 74 82
Final 90 74 98 88 80 62 78 74 86 80
Solution:
We want to predict the final exam scores from the mid term scores. So let us designate ‘y’ for
the final exam scores and ‘x’ for the mid term exam scores. We open the following table for
the calculations.

10 (65,071) - 785 (810) 14,860 810 – 785( 0.5127)

b 2
  0.5127 and a   40.7531
10( 64, 521) - (785) 28,985 10
Thus, the regression equation is given by y*  40.7531  (0.5127) x

We can use this to find the projected or estimated final scores of the students.
Eg for the midterm score of 50 the projected final score is
y *  40.7531  (0.5127) 50  66.3881 , which is a quite a good estimation.
To give another example, consider the midterm score of 70. Then the projected final score is
y*  40.7531  (0.5127) 70  76.6421 , which is again a very good estimation.

Practice Problems:
1. Consider the following data and draw a scatter plot
X 1.0 1.9 2.0 2.9 3.0 3.1 4.0 4.1 5
Y 10 99 100 999 1,000 1,001 10,000 10,001 100,000
2. . Let variable X is the number of hamburgers consumed at a cook-out, and variable Y is
the number of beers consumed. Develop a regression equation to predict how many beers
a person will consume given that we know how many hamburgers that person will
consume.
Subject 1 2 3 4 5
Hamburgers 5 4 3 2 1

47
 Beers 8 10 4 6 2
3. A horse owner is investigating the relationship between weight carried and the finish
position of several horses in his stable. Calculate r and R for the data given
Weight carried 110 113 120 115 110 115 117 123 106 108 110 110
Position Finished 2 6 3 4 6 5 4 2 1 4 1 3
4. The top and bottom number which may appear on a die are as follows Calculate r and R
for these values. Are the results surprising?
Top 1 2 3 4 5 6
Bottom 5 6 4 3 1 2
5. The ranks of two sets of variables (Heights and Weights) are given below. Calculate the
Spearman rank difference correlation coefficient R.
1 2 3 4 5 6 7 8 9 10
Heights 2 6 8 4 7 4 9.5 4 1 9.5
Weights 9 1 9 4 5 9 2 7 6 3
6. Researchers interested in determining if there is a relationship between death anxiety and
religiosity conducted the following study. Subjects completed a death anxiety scale (high score =
high anxiety) and also completed a checklist designed to measure an individuals degree of
religiosity (belief in a particular religion, regular attendance at religious services, number of times
per week they regularly pray, etc.) (high score = greater religiosity . A data sample is provided
below:
X 38 42 29 31 28 15 24 17 19 11 8 19 3 14 6
y 4 3 11 5 9 6 14 9 10 15 19 17 10 14 18
a) What is your computed answer?
b) What does this statistic mean concerning the relationship between death anxiety and
religiosity?
c) What percent of the variability is accounted for by the relation of these two variables?
7. The data given below are obtained from student records.( Grade Point Average (x) and Graduate
Record exam score (y)) Calculate the regression equation and compute the estimated GRE scores
for GPA = 7.5 and 8.5..
Subject 11 12 13 14 15 16 17 18 19 20
X 8.3 8.6 9.2 9.8 8.0 7.8 9.4 9.0 7.2 8.6
y 2300 2250 2380 2400 2000 2100 2360 2350 2000 2260
8. A horse was subject to the test of how many minutes it takes to reach a point from the starting
point. The horse was made to carry luggage of various weights on 10 trials.. The data collected are
presented below in the table. Find the regression equation between the load and the time taken to
reach the goal. Estimate the time taken for the loads of 35 Kgs , 23 Kgs, and 9 Kgs. Are the
answers in agreement with your intuitive feelings? Justify.
Trial Number 1 2 3 4 5 6 8 8 9 10
Weight (in Kgs) 11 23 16 32 12 28 29 19 25 20
Time taken (in mins) 13 22 16 47 13 39 43 21 32 22
9. A study was conducted to find whether there is any relationship between the weight and blood
pressure of an individual. The following set of data was arrived at from a clinical study.
Serial Number 1 2 3 4 5 6 8 8 9 10
Weight 78 86 72 822 80 86 84 89 68 71
Blood Pressure 140 160 134 144 180 176 174 178 128 132
10. It is assumed that achievement test scores should be correlated with student's classroom
performance. One would expect that students who consistently perform well in the classroom
(tests, quizes, etc.) would also perform well on a standardized achievement test (0 - 100 with 100
indicating high achievement (x)). A teacher decides to examine this hypothesis. At the end of the
academic year, she computes a correlation between the students achievement test scores (she
purposefully did not look at this data until after she submitted students grades) and the overall
48
 g.p.a.(y) for each student computed over the entire year. The data for her class are provided
below.
X 98 96 94 88 01 77 86 71 59 63 84 79 75 72 86 85 71 93 90 62
y 3.6 2.7 3.1 4.0 3.2 3.0 3.8 2.6 3.0 2.2 1.7 3.1 2.6 2.9 2.4 3.4 2.8 3.7 3.2 1.6
a) Compute the correlation coefficient.
b) What does this statistic mean concerning the relationship between achievement test
prformance and g.p.a.?
c) What percent of the variability is accounted for by the relationship between the two variables
and what does this statistic mean?
d) What would be the slope and y-intercept for a regression line based on this data?
e) If a student scored a 93 on the achievement test, what would be their predicted G.P.A.? If they
scored a 74? A 88?
11. With the growth of internet service providers, a researcher decides to examine whether there is a
correlation between cost of internet service per month (rounded to the nearest dollar) and degree
of customer satisfaction (on a scale of 1 - 10 with a 1 being not at all satisfied and a 10 being
extremely satisfied). The researcher only includes programs with comparable types of services. A
sample of the data is provided below.
Cost of internet (in $) 11 18 17 15 9 5 12 19 22 25
satisfaction 6 8 10 4 9 6 3 5 2 10
a) Compute the correlation coefficient.
b) What does this statistic mean concerning the relationship between amount of money spent
per month on internet provider service and level of customer satisfaction?
c) What percent of the variability is accounted for by the relationship between the two variables
and what does this statistic mean?
12. It is hypothesized that there are fluctuations in norepinephrine (NE) levels which accompany
fluctuations in affect with bipolar affective disorder (manic-depressive illness). Thus, during
depressive states, NE levels drop; during manic states, NE levels increase. To test this
relationship, researchers measured the level of NE by measuring the metabolite 3-methoxy-4-
hydroxyphenylglycol (MHPG in micro gram per 24 hour) in the patient's urine experiencing
varying levels of mania/depression. Increased levels of MHPG are correlated with increased
metabolism (thus higher levels) of central nervous system NE. Levels of mania/depression were
also recorded on a scale with a low score indicating increased mania and a high score increased
depression. The data is provided below.
MHPG 980 1209 1403 1950 1814 1280 1073 1066 880 776
Affect 22 26 8 10 5 19 26 12 23 28
a) Compute the correlation coefficient.
b) What does this statistic mean concerning the relationship between MHPG levels and affect?
c) What percent of the variability is accounted for by the relationship between the two variables?
d) What would be the slope and y-intercept for a regression line based on this data?
e) What would be the predicted affect score if the individual had an MHPG level of 1100? of
950? of 700?
13. The table below contains 25 cases -- the mother’s weight in kilograms and the infant's birth
weight in grams. Does this data suggest some relationship between the mother's weight and the
infant's birth weight Why would such a relationship be important

. Prepregnancy Weights of Mothers and Birthweights of their Infants

Case Number Mother's Weight (kg) Infant's Birthweight (g)
1 49.4 3515
2 63.5 3742
3 68.0 3629

49
 4 52.2 2680
5 54.4 3006
6 70.3 4068
7 50.8 3373
8 73.9 4124
9 65.8 3572
10 54.4 3359
11 73.5 3230
12 59.0 3572
13 61.2 3062
14 52.2 3374
15 63.1 2722
16 65.8 3345
17 61.2 3714
18 55.8 2991
19 61.2 4026
20 56.7 2920
21 63.5 4152
22 59.0 2977
23 49.9 2764
24 65.8 2920
25 43.1 2693

5 PROBABILITY
5.1 What is Probability?
Probability theory is the branch of mathematics that studies the possible outcomes of given
events together with the outcomes' relative likelihoods and distributions. In common usage,
the word “probability" is used to mean the chance that a particular event (or set of events) will
occur expressed on a linear scale from 0 (impossibility) to 1 (certainty). Factually, It is the
50
study of random or indeterministic experiments eg tossing a coin or rolling a die. If we roll a
die, we are certain it will come down but we are uncertain which face will show up. Ie the
face showing up is indeterministic. Probability is a way of summarizing the uncertainty of
statements or events. It gives a numerical measure for the degree of certainty (or degree of
uncertainty) of the occurrence of an event.

We often use P to represent a probability Eg P(rain) would be the probability that it rains. In
other cases Pr(.) is used instead of just P(.).

Definitions
 Experiment: A process by which an observation or measurement is obtained. Eg tossing a
coin or rolling a die.
 Outcome: Possible result of a random experiment. Eg a 6 when a die is rolled once or a
head when a coin is tossed.
 Sample space: Also called the probability space and it is a collection or set of all possible
outcomes of a random experiment. Sample space is usually denoted by S or  or U
 Event: it’s a subset of the sample space. Events are usually denoted by upper case letters.

Suppose the sample space S consists of n(S) equally likely outcomes and n(E) of those are
favourable for an event E then probability of an event E is the ratio of the number of
favourable outcomes n(E) to the total number of all possible outcomes n(S) ie
number of favourable outcomes n(E)
P(E)  
total number of possible outcomes n(S)

5.2 Approaches to Probability

There are three ways to define probability, namely classical, empirical and subjective
probability.
5.2.1 Classical probability
Classical or theoretical probability is used when each outcome in a sample space is equally
likely to occur. The underlying idea behind this view of probability is symmetry. Ie if the
sample space contains n outcomes that are fairly likely then P(one outcome)=1/n.
The classical probability for an event A is given by
Number of outcomes in A n(A)
P(A)  
Total number of outcomes in S n(S)
Eg Roll a die and observe that P(A)  P(rolling a 3) = 16 .

Example
A fair die, with faces numbered 1 to 6, is rolled once, write down the saple space S hence find
the probability that the score showing up is ; a) a multiple of 3 b) a prime number.
Solution
S  1, 2, 3, 4, 5, 6 Multiples of 3 are 3 and 6 while prime numbers are 2, 3 and 5
Thus P(Multiple of 3)  62  13 and P(prime number)  63  12

5.2.2 Frequentist or Empirical probability

When the outcomes of an experiment are not equally likely, we can conduct experiments to
give us some idea of how likely the different outcomes are. For example, suppose we are
interested in measuring the probability of producing a defective item in a manufacturing
process. The probability could be measured by monitoring the process over a reasonably long
period of time and calculating the proportion of defective items.
51
In a nut shell Empirical (or frequrntist or statistical) probability is based on observed data.
The empirical probability of an event A is the relative frequency of event A, that is
Frequency of event A
P(A) 
Total number of observatio ns
Example 1
The following are the counts of fish of each type that you have caught before.
Fish Types Blue gill Red gill Crappy Total
No of times caught 13 17 10 40
Estimate the probability that the next fish you catch will be a Blue gill.
P(Blue gill) = 1340  0.325

Example 2
A girl lists the number of male and female children her parent and her parent’s brothers and
sisters have. Her results were as tabulated below
Males Females
d) Find the probability that, if the girl
Her parents 2 5
has children of her own, the 1st
Her mother's sisters 6 8
born will be a girl.
Her mother's brothers 4 8 e) If the girl eventually has 10
Her father's sisters 5 8 children, how many are likely to
Her father's brothers 7 7 be males?
Totals 24 36
Solution
a) Following the family pattern, P(1st born will be a girl)  36
60
 0.6
b) 60% of the children will be females  40% will be males. Thus 4 out of 10 children are
likely to be males.
Remark: The empirical probability definition has a weakness that it depends on the results of
a particular experiment. The next time this experiment is repeated, you are likely to get a
somewhat different result. However, as an experiment is repeated many times, the empirical
probability of an event, based on the combined results, approaches the theoretical probability
of the event.

5.2.3 Subjective Probability:

Subjective probabilities result from intuition, educated guesses, and estimates. For example:
given a patient's health and extent of injuries a doctor may feel that the patient has a 90%
chance of a full recovery.. Subjectivity means two people can assign different probabilities to
the same event.

Regardless of the way probabilities are defined, they always follow the same laws, which we
will explore in the following Section.

Exercise
1) What is the probability of getting a total of 7 or 11, when two dice are rolled?
2) Two cards are drawn from a pack, without replacement. What is the probability that both
are greater than 2 and less than 8?
3) A permutation of the word "white" is chosen at random. Find the probability that it begins
with a vowel. Also find the probability that it ends with a consonant.
4) Find the probability that a leap year will have 53 Sundays.

52
 5) Two tetrahedral (4-sided) symmetrical dice are rolled, one after the other. Find the
probability that;
a) both dice will land on the same number.
b) each die will land on a number less than 3.
c) the two numbers will differ by at most 1.
Will the answers change if we rolled the dice simultaneously?

Ways to represent probabilities:

1) Venn diagram; We may write the probabilities inside the elementary pieces within a Venn
diagram. For example, P(AB' )  0.32 and P(A)  P(AB)  P(AB' ) = 0.58 [why?] The
relative sizes of the pieces do not have to match the numbers.

2) Two-way table; This is a popular way to represent statistical data. The cells of the table
correspond to the intersections of row and column events. Note that the contents of the table
add up accross rows and columns of the table. The bottom-right corner of the table contains
P(S) = 1
B B' Total
A 0.26 0.32 0.58
A' 0.11 ? 0.42
Total 0.37 0.63 1
Tree diagram; Tree diagrams or probability trees are simper clear ways of representing
probabilistic information. A tree diagram may be used to show the sequence of choices that
lead to the complete description of outcomes. For example, when tossing two coins, we may
represent this as follows

A tree diagram is also often useful for representing conditional probabilities

5.3 Review of set notation

Complement: The complement of event A, (denoted A' ), is the set of all outcomes in a
sample that are not included in the event A.
Intersection of events: The event A  B (or simply AB) read as ‘A intersection B’ consists of
outcomes that are contained within both events A and B. The probability of this event is the

53
probability that both events A and B occur [but not necessarily at the same time]. Here after
we will abbreviate intersection as AB.
Unions of Events: The event A  B read as ‘A union B’ consists of the outcomes that are
contained within at least one of the events A and B. The probability of this event P(A  B) ; is
the probability that events A and/or B occurs.

Set notation
Suppose a set S consists of points labelled 1, 2, 3 and 4. We denote this by S = 1, 2, 3, 4 . . If
A = 1, 2 and B  2, 3, 4 , then A and B are subsets of S, denoted by
A  S and B  S (B is contained in S). We denote the fact that 2 is an element of A by
2A .
The union of A and B, A  B  1, 2, 3, 4. If C   4 , then A  C  1, 2, 4. The
intersection A  B  AB  2: The complement of A, is A'  3, 4 .

Distributive laws; A  (B  C)  AB  AC and A  (B  C)  (A  B)  (A  C)

De Morgan's Law; (A  B)'  A' B' and (AB)'  A'B'

Venn diagram
Venn diagram is a diagram that shows all possible logical relations between a finite collection
of sets.
Venn diagram is often used to illustrate the relations between sets (events).
The sets A and B are represented as circles; operations between them (intersections, unions
and complements) can also be represented as parts of thevdiagram. The entire sample space S
is the bounding box. See Figure 2.1

Exercise
1) Use the Venn diagrams to illustrate Distributive laws and De Morgan's law.
2) Simplify the following (Draw the Venn diagrams to visualize)
a) (A' )' b) (AB)'A c) AB  AB' d) (A  B)  B
3) Represent by set notation and exhibit on a Venn diagram the following events
a) both A and B occur
b) exactly one of A, B occurs
c) A and B but not C occur s
d) at least one of A, B, C occurs
e) at most one of A, B, C occurs
4) The sample space consists of eight capital letters (outcomes), A, B, C ,...,H. Let V be the
event that the letter represents a vowel, and L be the event that the letter is made of
straight lines. Describe the outcomes that comprise
a) VL b) V  L' c) V' L'

54
5) Out of all items sent for refurbishing, 40% had mechanical defects, 50% had electrical
defects, and 25% had both. Denoting A = fan item has a mechanical defect and B = fan
item has an electrical defect, fill the probabilities into the Venn diagram and determine the
quantities listed below. a) P(A) b) P(AB) c) P(A' B) d) P(A' B' ) e) P(A  B) f)
P(A'B' ) g) P(A  B)'
6) A sample of mutual funds was classified according to whether a fund was up or down last
year ( A and A' ) and whether it was investing in international stocks ( B and B' ). The
probabilities of these events and their intersections are represented in the two-way table
below. Fill in all the question marks hence find the probability of A  B
B B'
A 0.33 ? ?
A' ? ? 0.52
0.64 ? 1

5.4 Rules of Probability

1) For an experiment with a sample space S = E1 , E 2 , . . . , E n  we can assign probabilities
n
P(E1 ), P(E 2 ), . . . , P(E n ) provided that 0  P(E i )  1 and P(S)   P(E i )  1
i 1
Remark:
k
a) If a set (event) A consists of outcomes E1 , E 2 , . . . , E k , then P(A)   P(E i )
i 1

b) If E = S then P(E)  P(S )  1 and If E is has no elements, (ie if E is empty

 E   or  ), then P(E)  P( )  0
2) If E is an event in the sample space S, then E' (called the complement of E) is an event in
S but outside E. P(E)  P(E' )  1  P(E' )  1  P(E)
3) If the sample space S contains n disjoint events E1 , E 2 , . . . , E n , then
n
P(E1 )  P(E 2 ), . . . P(E n )   P(Ei )  1
i 1

4) Let A and B be two events such that A  B, then P(A)  P(B)

5) For any two events A and B, P(A  B)  P(A)  P(B)  P(AB) where P(AB)  P(A  B) .
Extension of this rule leads to the Inclusion–Exclusion Principle. This principle is a way
to extend the general addition rule to 3 or more events. Here we will limit it to 3 events.
P(A  B  C)  P(A)  P(B)  P(AB)  P(AC)  P(BC)  P(ABC)
6) Law of Partitions: The law of partitions is a way to calculate the probability of an event.
Let A1 , A 2 , . . . , A k form a partition of the sample space  .then, for any events B,
k
P(B)  P(A1B)  P(A 2 B)  ....  P(A k B)   P(A i B)
i 1

55
Example 1
The Probability that John passes a Maths exam is 4/5 and that he passes a Chemistry exam is
5/6. If the probability that he passes both exams is 3/4, find the probability that he will pass at
least one exam.
Solution
Let M be the event thet John passes Math exam, and C be the event thet John passes
Chemistry exam.
P(John passes at least one exam) = P(M  C) = P(M)  P(C) - P(MC)  54  56  34  60
53

Example 2
A fair die, with faces numbered 1 to 6, is rolled twice and the sum of the scores showing up
noted. Let A be the event that the sum of the scores is greater than 7, B be the event that the
sum of the scores is is a multiple of 3 and C be the event that the sum of the scores is a prime
number. Show that P(A  B)  P(A)  P(B)  P(AB) and also find P(A  C), P(BC) and
P(BC' )
Solution
+ 1 2 3 4 5 6 A  B means the set of all multiples of 3
that are greater than 7. Clearly
1 2 3 4 5 6 7
2 3 4 5 6 7 8 P(A  B)  365
3 4 5 6 7 8 9 A  B means the set of all values that are
4 5 6 7 8 9 10 multiples of 3 and/or greater than 7.
Clearly
5 6 7 8 9 10 11
6 7 8 9 10 11 12 P(A  B)  3622
 18
11
 P(A)  P(B)  P(AB)
1. P(A)  15
36
 125 and P(B)  12
36
 13
2. P(C)  15
36
 125
A  C means the set of all multiples of 3 that are prime number. Clearly P(A  C)  2
36
 181
P(A  C)  P(A)  P(C)  P(AC)  125  125  181  79
B  C means the set of all greater than 7 that are prime numbers. Clearly P(BC)  2
36
 181
B C' means the set of all greater than 7 that are not prime numbers. Clearly P(BC )  11
36
|

Exercise
1) Which of the following is a probability function defined on S = E1 , E 2 , E 3 
a) P(E1 )  14 , P(E 2 )  13 and P(E3 )  12 b) P(E1 )  13 , P(E2 )  16 and P(E3 )  12
c) P(E1 )  23 , P(E 2 )   13 and P(E3 )  23 d) P(E1 )  0, P(E2 )  13 and P(E3 )  23
2) As a foreign language, 40% of the students took Spanish and 30% took French, while
60% took at least one of these languages. What percent of students took both Spanish and
French?
3) In a class of 100 students, 30 are in mathematics. Moreover, of the 40 females in the
class, 10 are in Mathematics. If a student is selected at random from the class, what is the
probability that the student will be a male or be in mathematics?
4) The probability that a car stopped at a road brook will have faulty breaks is 0.23, the
probability that it will have badly worn out tyres is 0.24 and the probability that it will
have faulty breaks and/or badly worn out tyres is 0.38. Find the probability that a car
which has just been stopped will have both faulty breaks and badly worn out tyres.

56
5) Given two events A and B in the sane sample space such that P(A)  0.59 , P(B)  0.3
and P(AB)  0.21 . Find; a) P(A  B) b) P(A' B) c) P(AB' ) d) P(A  B' )
6) Let A and B be two events in the sane sample space such that P(A  B)  34 , P(B' )  23
and P(AB)  14 .Find P(B), P(A) and P(AB' )
7) Suppose that P(A)  0.4 , P(B)  0.5 and P(A  B)  0.2 Find; a) P(A  B) b) P(A' B' )
c) PA'( A  B) d) PA  (A' B)
8) A die is loaded such that even numbers are twice as likely as odd numbers. Find the
probability that for a single toss of this die the spot showing up is greater than 3
9) A point is selected at random inside an equilateral triangle of sides 3 units. Find the
probability that its distance to any corner is greater than 1 unit.

Definition: (Odds of an event)

It’s the ratio of the probability of an event occurring to that of the event not happening. If A
)  1-P(A)
P(A) P(A)
is an event then the odds of A is given by P(A'
Example
Find P(A) and P(A' ) if the odds of event A is 54
Solution
P(A)
1- P(A)
 54  4(1  P(A))  5P(A)  4  9P(A)  P(A)  94 and P(A' )  95
3 a
Question: Find P(E) and P(E' ) if the odds of event E is (i) 4 (ii) b

5.5 Relationship between Events

 Compound event: Two or more events combined together. Eg AB is a compound event
 Mutually exclusive events: Two events A and B are said to be mutually exclusive if they
cannot occur simultaneously. That is if the occurrence of A totally excludes the
occurrence of B. Effectively events A and B are said to be mutually exclusive if they
disjoint. ie A  B    P(AB)  0
 Exhaustive events: Events whose union equals the sample space.
 Independent events: Two events A and B are said to be independent if the occurrence of
A does not affect the occurrence of B. If events A and B are independent then
P(AB)  P(A)  P(B)

Remark: Three events A, B and C are said to be jointly independent if and only if
(i) P(AB)  P(A)  P(B) , P(AC)  P(A)  P(C) and P(BC)  P(B)  P(C) (ie they are
pairwise independent ) and
(ii) if P(ABC)  P(A)  P(B)  P(C)
Note it does not necessarily mean that if events A,B and C are pairwise independent then
they are jointly independent

Example 1
Roll a fair die twice and define A to be the event that the sum of the scores showing up is
greater than 7, B be the event that the sum of the scores showing up is a multiple of 3 and C
be the event that the sum of the scores showing up is a prime number. Which of the events A,
B and C are independent? Are the 3 events jointly independent?

Solution
57
 From the above example, P(A)  P(C)  125 , P(B)  13 , P(AB)  365 and P(AC)  P(BC)  181
Since P(AB)  0 events A and B are not mutually exclusive. Similarly events A & C and B
and C are not mutually exclusive
P(A)  P(B)  125  13  365  P(AB)  A and B are independent events.
P(A)  P(C)  125  125  144
25
 P(AC)  A and C are dependent events.
P(B)  P(C)  3  12  36  P(BC)  B and C are dependent events.
1 5 5

The 3 events are not jointly independent since pairwise independence is not satisfied.

Example 2
Three different machines in a factory have the following probabilities of breaking down
during a shift.
Machine A B C Find the probability that in a particular shift,;
probability 15
4 3
10
2
11
a) All the machines will break down
b) None of the machines will break down.
Solution
Since the events of breaking down of machines are independent, probability that all the
machines will break down is given by P(ABC)  154  103  112  275
4

The probability that none of the machines will break down is given by
P(ABC)  1511
 107  119  50
21

Exercises
1) In a game of archery the probability that A hits the target is 13 and the probability that B
hits the target is 2 5 . What is the probability that the target will be hit?
2) Toss a fair coin 3 times and let A be the event that two or more heads appears, B be the
event that all outcomes are the same and C be the event that at most two tails appears.
Which of the events A, B and C are independent? Are the 3 events jointly independent?
3) A fair coin and a fair die are rolled together once. Let A be the event that a head and an
even number appears, B be the event that a prime number appears and C be the event that
a tail and an odd number appears.
a) Express explicitly the event that i) A and B occurs ii) Only B occurs iii) B and C
occur
b) Which of the events A, B and C are independent and which ones are mutually
exclusive?
4) A die is loaded so that the probability of a face showing up is proportional to the face
number. Write down the probability of each sample point. If A is the event that an even
number appears, B is the event that a prime number appears and C is the event that an odd
number appears.
a) Find the probability that: i) A and/or B occurs ii) A but not B occurs iii) B and C
occurs d) A and/or C occurs
b) Which of the events A, B and C are independent and which ones are mutually
exclusive?

Theorem 1: If events A and B are independent, then A and B' are also independent
Proof

58
Decomposing A into two disjoint events AB and AB' . We can write
P(A)  P(AB)  P(AB' )  P(AB' )  P(A)  P(AB)  P(A)  P(A)  P(B) since events A and B
are independent. Thus P(AB' )  P(A)1  P(B)  P(A)  P(B' )  A and B' are independent

Theorem 2: If events A and B are independent, then A' and B' are also independent
Proof
Decomposing B' into two disjoint events AB' and A' B' . We can write
P(B' )  P(AB' )  P(A' B' )  P(A' B' )  P(B' )  P(AB' )  P(B' )  P(A)  P(B' ) since events A
and B' are independent. (From theorem 1 above) Thus
P(A' B' )  1  P(A)P(B' )  P(A' )  P(B' )  A' and B' are also independent

5.6 Counting Rules useful in Probability

In some experiments it is helpful to list the elements of the sample space systematically by
means of a tree diagram,. In many cases, we shall be able to solve a probability problem by
counting the number of points in the sample space without actually listing each element.

Theorem (Multiplication principle)

If one operation can be performed in n1 ways, and if for each of these a second operation can
be performed in n 2 ways, then the two operations can be performed together in n 1n 2 ways.
Eg How large is the sample space when a pair of dice is thrown?
Solution; The first die can be thrown in n 1  6 ways and the second in
n 2  6 Ways. Therefore, the pair of dice can land in n 1 n 2  36 possible ways.
The above theorem can naturally be extended to more than two operations: if we have
n 1 , n 2 , .... , n k consequent choices, then the total number of ways is n 1 n 2  ....n k

Permutations
Permutations refer to an arrangement of objects when the order matters (for example, letters
n!
in a word).The number of permutations of n distinct objects taken r at a time is n P r 
(n  r )!
Example
From among ten employees, three are to be selected to travel to three out-of-town plants A, B,
and C, one to each plant. Since the plants are located in different cities, the order in which the
employees are assigned to the plants is an important consideration. In how many ways can the
assignments be made?
Solution;
Because order is important, the number of possible distinct assignments is 10 P 3  720
In other words, there are ten choices for plant A, but then only nine for plant B, and eight for
plant C. This gives a total of 10(9)(8) ways of assigning employees to the plants.

Combinations
The term combination refers to the arrangement of objects when order does not matter. For
example, choosing 4 books to buy at the store in any order will leave you with the same set of
books. The number of distinct subsets or combinations of size r that can be selected from n
n!
distinct objects, (r _ n), is given by n C r 
r!(n  r )!
Example 1
59
In the previous example, suppose that three employees are to be selected from among the ten
available to go to the same plant. In how many ways can this selection be made?
Solution
Here, order is not important; we want to know how many subsets of size r = 3 can be selected
from n = 10 people. The result is 10 C 3  120
Example 2
In a poker hand consisting of 5 cards, find the probability of holding 2 aces and 3 jacks.
Solution
The number of ways of being dealt 2 aces from 4 is 4 C 2  6 and the number of ways of being
dealt 3 jacks from 4 is 4 C 3  4
The total number of 5-card poker hands, all of which are equally likely is 52 C 5  2, 598, 960
Hence, the probability of getting 2 aces and 3 jacks in a 5-card poker hand is P(C) =
6 4
P(C) 
2, 598, 960

Example 3
A university warehouse has received a shipment of 25 printers, of which 10 are laser printers
and 15 are inkjet models. If 6 of these 25 are selected at random to be checked by a particular
technician, what is the probability that; a) exactly 3 of these selected are laser printers? b) at
least 3 inkjet printers?
Solution
First choose 3 of the 15 inkjet and then 3 of the 10 laser printers.
There are 15 C 3 and 10 C 3 ways to do it, and therefore
C3 10 C3
P(exactly 3 of the 6) = 15
 0.3083
25 C 6

C  C C  C C  C C  C
P(at least 3) = 15 3 10 3  15 4 10 2  15 5 10 1  15 6 10 0  0.8530
25 C 6 25 C 6 25 C 6 25 C 6
Exercises
1) An incoming lot of silicon wafers is to be inspected for defectives by an engineer in a
microchip manufacturing plant. Suppose that, in a tray containing 20 wafers, 4 are
defective. Two wafers are to be selected randomly for inspection. Find the probability that
neither is defective.
2) A person draws 5 cards from a shuffled pack of cards. Find the probability that the person
has at least 3 aces. Find the probability that the person has at least 4 cards of the same
suit.
3) A California licence plate consists of a sequence of seven symbols: number, letter, letter,
letter, number, number, number, where a letter is any one of 26 letters and a number is
one among 0, 1,… 9. Assume that all licence plates are equally likely. What is the
probability that;
a) all symbols are different?
b) b) all symbols are different and the first number is the largest among the
numbers?
4) A bag contains 80 balls numbered 1…. 80. Before the game starts, you choose 10
different numbers from amongst 1…. 80 and write them on a piece of paper. Then 20 balls
are selected (without replacement) out of the bag at random. What is the probability that;
a) all your numbers are selected?
b) none of your numbers is selected?

60
 c) exactly 4 of your numbers are selected?
5) A full deck of 52 cards contains 13 hearts. Pick 8 cards from the deck at random (a)
without replacement and (b) with replacement. In each case compute the probability that
you get no hearts.
6) Three people enter the elevator on the basement level. The building has 7 floors. Find the
probability that all three get off at different floors.
7) In a group of 7 people, each person shakes hands with every other person. How many
handshakes did occur?
8) A marketing director considers that there's “overwhelming agreement" in a 5-member
focus group when either 4 or 5 people like or dislike the product. If, in fact, the product's
popularity is 50% (so that all outcomes are equally likely), what is the probability that the
focus group will be in “overwhelming agreement" about it? Is the marketing director
making a judgement error in declaring such agreement “overwhelming"?
9) A die is tossed 5 times. Find the probability that we will have 4 of a kind.
10) A tennis tournament has 2n participants, n Swedes and n Norwegians. First, n people are
chosen at random from the 2n (with no regard to nationality) and then paired randomly
with the other n people. Each pair proceeds to play one match. An outcome is a set of n
(ordered) pairs, giving the winner and the loser in each of the n matches. (a) Determine
the number of outcomes. (b) What do you need to assume to conclude that all outcomes
are equally likely? (c) Under this assumption, compute the probability that all Swedes are
the winners.
11) A group of 18 Scandinavians consists of 5 Norwegians, 6 Swedes, and 7 Finns. They are
seated at random around a table. Compute the following probabilities: (a) that all the
Norwegians sit together, (b) that all the Norwegians and all the Swedes sit together, and
(c) that all theNorwegians, all the Swedes, and all the Finns sit together.
12) In a lottery, 6 numbers are drawn out of 45. You hit a jackpot if you guess all 6 numbers
correctly, and get $400 if you guess 5 numbers out of 6. What are the probabilities of each
of those events?
13) There are 21 Bachelor of Science programs at New Mexico Tech. Given 21 areas from
which to choose, in how many ways can a student select:
a) A major area and a minor area?
b) A major area and first and second minor?
14) From a box containing 5 chocolates and 4 hard candies, a child takes ahandful of 4 (at
random). What is the probability that exactly 3 of the 4 arechocolates?
15) If a group consist of 8 men and 6 women, in how many ways can a committee of 5 be
selected if:
a) The committee is to consist of 3 men and 3 women.
b) There are no restrictions on the number of men and women on the committee.
c) There must at least one man.
d) There must be at least one of each sex.
16) Suppose we have a lot of 40 transistors of which 8 are defective. If we sample without
replacement, what is the probability that we get 4 good transistors in the first 5 draws?
17) A housewife is asked to rank four brands A, B, C, and D of household cleaner according
to her preference, number one being the one she prefers most, etc. she really has no
preference among the four brands. Hence, any ordering is equally likely to occur.
a) Find the probability that brand A is ranked number one.
b) Find the probability that brand C is number one D is number 2 in the rankings.
c) Find the probability that brand A is ranked number one or number 2.
18) How many ways can one arrange the letters of the word ADVANTAGE so
19) that the three As are adjacent to each other?

61
20) Eight tires of different brands are ranked 1 to 8 (best to worst) according to mileage
performance. If four of these tires are chosen at random by a customer, find the
probability that the best tire among the four selected by the customer is actually ranked
third among the original eight.

5.7 Conditional Probability and Independence

Humans often have to act based on incomplete information. If your boss has looked at you
gloomily, you might conclude that something's wrong with your job performance. However,
if you know that she just suffered some losses in the stock market, this extra information may
change your assessment of the situation. Conditional probability is a tool for dealing with
additional information like this.
Conditional probability is the probability of an event occurring given the knowledge that
another event has occurred. The conditional probability of event A occurring, given that event
B has occurred is denoted by P(A/B) and is read “probability of A given B” and is given by
P(AB) P(AB)
P(A/B)  provided P(B)  0 Similarly P(B/A)  provided P(A)  0
P(B) P(A)
 P(AB)  P(A/B)  P(B)  P(B/A)  P(A)
Remark: Another way to express independence is to say that the knowledge of B occurring
does not change our assessment of P(A). This means that if A and B are independent then
P(A/B)  P(A) and P(B/A)  P(B)

Example
In a large metropolitant area, the probability of a family owning a colour T.V , a computer or
both 0.86, 0.35 and 0.29 respectively. What is the probability that a family chosen at random
during a survey will own a colour T.V and/or a computer? Given that the family chosen at
random during a survey owns a colour T.V, what is the probability that it will own a
computer?
Solution
Let T and C be the event of owning a colour T.V and a computer respectively. Then
P(T  C)  P(T)  P(C)  P(TC)  0.86  0.35 - 0.29  0.92
P(TC) 0.29
P(C/T)    0.
P(T) 0.86
Reduced sample space approach
In case when all the outcomes are equally likely, it is sometimes easier to find conditional
probabilities directly, without having to apply the above equation. If we already know that B
has happened, we need only to consider outcomes in B, thus reducing our sample space to B.
Number of outcomes in AB
Then, P(A/B) 
Number of outcomes in B
For example, P(a die is 3 / a die is odd) = 13 and P(a die is 4 / a die is odd) = 0

Example
Let A = {a family has two boys} and B = {a family of two has at least one boy} Find P(A/B)
Solution
The event B contains the following outcomes: B  (B, B) , (B, G) , (G; B) and. Only one of
these is in A. Thus, P(A/B)  13 . However, if I know that the family has two children, and I
see one of the children and it's a boy, then the probability suddenly changes to 1/2. There is a
subtle difference in the language and this changes the conditional probability

62
5.7.1 Tree Diagrams in conditional probability
Suppose we are drawing marbles from a bag that initially contains 7 red and 3 green marbles.
The drawing is without replacement that is after we draw the first marble, we do not put it
back. Let's denote the events R 1 = the first marble is red R 2 = the second marble is red
G1 = tthe first marble is green and so on. Let's fill out the tree representing the consecutive
choices.

The conditional probability P(R 2 /R 1 ) can be obtained directly from reasoning that after we
took the first red marble, there remain 6 red and 3 green marbles. On the other hand, we could
P(R 1R 2 ) 42 90 2
use the formula to get P(R 2 /R 1 )    where the probability P(R 2 R 1 )
P(R 1 ) 7
10 3
C 7
{same as P(R 1R 2 ) } can be obtained from counting the outcomes P(R1R 2 )  7 2 
10 C 2 15
Question: Find P(R 2 ) and P(R1 /R 2 ) .
Example 2
Suppose that of all individuals buying a certain digital camera, 60% include an optional
memory card in their purchase, 40% include a set of batteries, and 30% include both a card
and batteries. Consider randomly selecting a buyer and let A = memory card purchased and
B = battery purchased. Then find P(A/B) and P(B/A).
Solution
From given information, we have P(A)  0.60 , P(B)  0.40 and
P(both purchased) = P(A/B) = 0.30
Given that the selected individual purchased an extra battery, the probability that an optional
P(AB) 0.30
card was also purchased is P(A/B)    0.75
P(B) 0.40
That is, of all those purchasing an extra battery, 75% purchased an optional memory card.
P(AB) 0.30
Similarly P(battery j memory card) = P(B/A) =   0.5
P(A) 0.60
Notice that P(A/B)  P(A) and P(B/A)  P(B) , that is, the events A and B are dependent.
Remark: The tree diagram may become tedious especially when the tree grows beyond 4
stages. In such a case we can make use of binomial formular which is applicable when:
i) The experiment’s outcome can be classified into 2 categories success and failure with
probabilities p and 1-p respectively
ii) The experiment is to be repeated n independent times
iii) Our interest is the number of successes
63
The probability of observing x successes out of n trials is given by:-
P(x) n C x  p x (1  p) n  x for x = 0,1 , 2, . . . . , n
Example
A fair coin is tossed 10 times.Whai is the probability of observing exactly 8 heads?
Solution
n  10 p  0.5 5 and x = 8 successes Therefore
P(X  8)10 C8  0.58 (1  0.5)10 8 10 C8  0.510  0.044
Exercises
1) A pair of fair dice is rolled once. If the sum of the scores showing up is 6 , find the
probability that one of the dice shows a 2.
2) A consumer research organisation has studied the services and warranty provided by 50
new car dealers in a certain city. It’s findings are as follows
In Business for Good services and a Poor services and a
warranty warranty
At least 10 yeras 16 4
Less than 10 years 10 20
If a person randomly selects one of these new car dealers ;
a) What is the probability that he gets one who provides good services and a warranty
b) Who has been in business for at least 10 years, what is the probability that hehe
provides good services and a warranty
c) What is the probability that one of these new car dealers who has been in business for
less than 10 years will provide good services and a warranty?
3) Three machines A, B and C produces 50%, 30% and 20% respectively of the total number
of items in a factory. The percentage of defective outputs of these machines are 3%, 4%
and 5% respectively. If an item is selected at random:-
a) Find the probability that it is defective
b) And found to be defective, what is the probability that it was produced by machine A?
4) A year has 53 Sundays. What is the conditional probability that it is a leap year?
5) The probability that a majority of the stockholders of a company will attend a special
meeting is 0.5. If the majority attends, then the probability that an important merger will
be approved is 0.9. What is the probability that a majority will attend and the merger will
be approved?
6) Let events A, B have positive probabilities. Show that, if P(A/B) = P(A) then also P(B/A)
= P(B).
7) The cards numbered 1 through 10 are placed in a hat, mixed up, and then one of the cards
is drawn. If we are told that the number on the drawn card is at least five, what is the
probability that it is ten?
8) In the roll of a fair die, consider the events A = {2, 4, 6} = “even numbers" and B = {4, 5,
6} =”high scores". Find the probability that die showing an even number given that it is a
high score.
9) There are two urns. In the first urn there are 3 white and 2 black balls and in the second
urn there 1 white and 4 black balls. From a randomly chosen urn, one ball is drawn. What
is the probability that the ball is white?
10) The level of college attainment of US population by racial and ethnic group in 1998 is
given in the following table
Racial or No of Adults %age with %age With %age with Graduate or
Ethnic Group (Millions) Associate's Degree Bachelor's Degree Professional Degree
Native 1.1 6.4 6.1 3.3
Americans
64
 Blacks 16.8 5.3 7.5 3.8
Asians 4.3 7.7 22.7 13.9
Hispanics 11.2 4.8 5.9 3.3
Whites 132.0 6.3 13.9 7.7
The percentages given in the right three columns are conditional percentages.
a) How many Asians have had a graduate or professional degree in 1998?
b) What percent of all adult Americans has had a Bachelor's degree?
c) Given that the person had an Associate's degree, what is the probability that the
person was Hispanic?
11) The dealer's lot contains 40 cars arranged in 5 rows and 8 columns. We pick one car at
random. Are the events A = {the car comes from an odd-numbered row} and B = {the car
comes from one of the last 4 columns} independent? Prove your point of view.
12) You have sent applications to two colleges. If you are considering your chances to be
accepted to either college as 60%, and believe the results are statistically independent,
what is the probability that you'll be accepted to at least one? How will your answer
change if you applied to 5 colleges?
13) In a high school class, 50% of the students took Spanish, 25% took French and 30% of the
students took neither. Let A be the event that a randomly chosen student took Spanish,
and B be the event that a student took French. Fill in either the Venn diagram or a 2-way
table and answer the questions:

a) Describe in words the meaning of the event AB' . Find the probability of this event.
b) Are the events A, B independent? Explain with numbers why or why not.
c) If it is known that the student took Spanish, what are the chances that she also took
French?
14) One half of all female physicists are married. Among those married, 50% are married to
other physicists, 29% to scientists other than physicists and 21% to non-scientists’.
Among male physicists, 74% are married. Among them, 7% are married to other
physicists, 11% to scientists other than physicists and 82% to non-scientists What percent
of all physicists are female? [Hint: This problem can be solved as is, but if you want to,
assume that physicists comprise 1% of all population.]
15) Error-correcting codes are designed to withstand errors in data being sent over
communication lines. Suppose we are sending a binary signal (consisting of a sequence of
0's and 1's), and during transmission, any bit may get flipped with probability p,
independently of any other bit. However, we might choose to repeat each bit 3 times. For
example, if we want to send a sequence 010, we will code it as 000111000. If one of the
three bits flips, say, the receiver gets the sequence 001111000, he will still be able to
decode it as 010 by majority voting. That is, reading the first three bits, 001, he will
interpret it as an attempt to send 000. However, if two of the three bits are flipped, for
example 011, this will be interpreted as an attempt to send 111, and thus decoded
incorrectly. What is the probability of a bit being decoded incorrectly under this scheme?

5.8 Bayes’ Rule

65
Events B1 , B2 , . . . , BK are said to be a partition of the sample space S if the following two
conditions are satisfied. i) Bi BJ   for each pair i , j and ii) B1  B2  . . .  BK = S
This situation often arises when the statistics are available in subgroups of a population. For
example, an insurance company might know accident rates for each age group Bi . This will
give the company conditional probabilities P(A/Bi ) (if we denote A = {event of accident}.
Question: if we know all the conditional probabilities P(A/Bi ) , how do we find the
unconditional P(A)?
Consider a case when k = 2:
The event A can be written as the union of mutually exclusive events AB1 and AB 2 , that is
A = AB1  AB2 it follows that P(A) = P(AB1 )  P(AB2 )
If the conditional probabilities P(A/B1 )) and P(A/B2 ) ) are known, that is
P(AB1 ) P(AB2 )
P(A/B1 )  and P(A/B2 )  then P(A)  P(B1 )  P(A/B1 )  P(B2 )  P(A/B2 )
P(B1 ) P(B2 )
Suppose we want to find probability of the form P(Bi /A) , which can be written as
P(ABi ) P(Bi )  P(A/Bi ) P(Bi )  P(A/Bi )
P(Bi /A)   
P(A) P(A) P(B1 )  P(A/B1 )  P(B2 )  P(A/B 2 )
This calculation generalizes to k > 2 events as follows.
Theorem
If B1 , B2 , . . . , BK form a partition of the sample space S such that P(Bi )  0 for i = 1, 2, . . . k;
then for any event A  S ,
k k
P(ABi ) P(B )  P(A/Bi )
P(A)   P(ABi )   P(Bi )  P(A/Bi ) Subsequent ly, P(Bi /A)   k i
 P(Bi )  P(A/Bi )
i 1 i 1 P(A)
i 1

This last equation is often called Law of Total Probability.

Example 1
A rare genetic disease (occuring in 1 out of 1000 people) is diagnosed using a DNA screening
test. The test has false positive rate of 0.5%, meaning that P(test positive / no disease) = 0.005 .
Given that a person has tested positive, what is the probability that this person actually has the
disease? First, guess the answer, then read on.
Solution
Let's reason in terms of actual numbers of people, for a change.
Imagine 1000 people, 1 of them having the disease. How many out of 1000 will test positive?
One that actually has the disease, and about 5 disease-free people who would test false
positive. Thus, P(disease/test positive)  16 .
It is left as an exercise for the reader to write down the formal probability calculation.
Example 2
At a certain assembly plant, three machines make 30%, 45%, and 25%, respectively, of the
products. It is known from the past experience that 2%, 3% and 2% of the products made by
each machine, respectively, are defective. Now, suppose that a finished product is randomly
selected.
a) What is the probability that it is defective?
b) If a product were chosen randomly and found to be defective, what is the probability
that it was made by machine 3?
Solution
Consider the following events:
66
A: the product is defective and B i : the product is made by machine i=1, 2, 3,
Applying additive and multiplicative rules, we can write
(a) P(A)  P(B1 )  P(A/B1 )  P(B2 )  P(A/B2 )  P(B3 )  P(A/B3 )
 (0.3)(0.02) + (0.45)(0.03) + (0.25)(0.02) = 0.006 + 0.0135 + 0.005 = 0.0245
P(B3 )  P(A/B3 ) 0.005
(b) Using Bayes' rule P(B3 /A )    0.2041
P(A) 0.0245
This calculation can also be represented using a tree diagram as follows

Here, the first branching represents probabilities of the events B1 and the second branching
represents conditional probabilities P(A/B1 ) . The probabilities of intersections, given by the
products, are on the right. P(A) is their sum.

Exercises
1) Lucy is undecided as to whether to take a Math course or a Chemistry course. She
estimates that her probability of receiving an A grade would be 0.5 in a math course, and
2
3 in a chemistry course. If Lucy decides to base her decision on the flip of a fair coin,
what is the probability that she gets an A?
2) Of the customers at a gas station, 70% use regular gas, and 30% use diesel. Of the
customers who use regular gas, 60% will fill the tank completely, and of those who use
diesel, 80% will fill the tank completely.
a) What percent of all customers will fill the tank completely?
b) If a customer has filled up completely, what is the probability it was a customer
buying diesel?
3) In 2004, 57% of White households directly and/or indirectly owned stocks, compared to
26% of Black households and 19% of Hispanic households. The data for Asian
households is not given, but let's assume the same rate as for Whites. Additionally, 77%
of households are classifieds as either White or Asian, 12% as African American, and
11% as Hispanic.
a) What proportion of all families owned stocks?
b) If a family owned stock, what is the probability it was White/Asian?
4) Drawer one has five pairs of white and three pairs of red socks, while drawer two has
three pairs of white and seven pairs of red socks. One drawer is selected at random a pair
of socks is selected at random from that drawer.
a) What is the probability that it is a white pair of socks?
b) Suppose a white pair of socks is obtained. What is the probability that it came from
drawer two?
5) For an on-line electronics retailer, 5% of customers who buy Zony digital cameras will
return them, 3% of customers who buy Lucky Star digital cameras will return them, and
8% of customers who buy any other brand will return them. Also, among all digital
67
 cameras bought, there are 20% Zony's and 30% Lucky Stars. Fill in the tree diagram and
answer the questions.
a) What percent of all cameras are returned?
b) If the camera was just returned, what is the probability it is a Lucky Star?
c) What percent of all cameras sold were Zony and were not returned?

6) Three newspapers, A, B, and C are published in a certain city. It is estimated from a

survey that that of the adult population: 20% read A, 16% read B, 14% read C, 8% read
both A and B, 5% read both A and C, 4% read both B and C, 2% read all three. What
percentage reads at least one of the papers? Of those that read at least one, what
percentage reads both A and B?
7) Suppose P(A/B)  0.3 , P(B)  0.4 and P(B/A)  0.6 . Find P(A) and P(A  B)
8) This is the famous Monty Hall problem f A contestant on a game show is asked to choose
among 3 doors. There is a prize behind one door and nothing behind the other two. You
(the contestant) have chosen one door. Then, the host is flinging one other door open, and
there's nothing behind it. What is the best strategy? Should you switch to the remaining
door, or just stay with the door you have chosen? What is your probability of success
(getting the prize) for either strategy?
9) There are two children in a family. We overheard about one of them referred to as a boy.
a) Find the probability that there are 2 boys in the family.
b) Suppose that the oldest child is a boy. Again, find the probability that there are 2 boys
in the family.g [Why is it different from part (a)?]
10) At a university, two students were doing well for the entire semester but failed to show up
for a final exam. Their excuse was that they travelled out of state and had a at tire. The
professor gave them the exam in separate rooms, with one question worth 95 points:
\which tire was it?". Find the probability that both students mentioned the same tire.
11) In firing the company's CEO, the argument was that during the six years of her tenure, for
the last three years the company's market share was lower than for the first three years.
The CEO claims bad luck. Find the probability that, given six random numbers, the last
three are the lowest among six.

68