BIOSTATISTICS

WEEK 8
LEVEL 2
KASU
THE STUDENTS T TEST

• It is applied to find the significance of difference

between two
• means as:
• 1. Unpaired t-test
• 2. Paired t-test.
UNPAIRED T TEST

• This test is applied to unpaired data of independent

observations made on individuals of two different or
separate groups or samples drawn from two
populations, to test if the difference between the two
means is real or it can be attributed to sampling
variability such as between means of the control and
experimental groups.
Determine the pooled
degrees of freedom from the
formula
df = (n1 – 1) + (n2 – 1) = n1
+ n2 – 2.
• Compare calculated value with the table value at particular
degrees of freedom to find the level of significance in two-
tailed test. In one-tailed test, compare your result with value
given under P = 0.10 and P = 0.02. If it is higher, it is
significant at 5% level (P = 0.05) and 1% level (P = 0.01),
respectively, i
DEGREES OF FREEDOM

• Degrees of freedom is a statistical term that indicates the

number of observations that are free to vary. In other words,
we could change 29 of the values in the sample and still get
the same mean of 10.7 as long as we control or set the last
value so that the mean is 10.7.
CHI – SQUARE TEST

• Chi-square test, unlike other tests of significance such as

‘Z’ and ‘t’ tests discussed previously, is a nonparametric
test not based on any assumption or distribution of any
variable.
• The test of association between two events in binomial or
multinomial samples is the most important application of
the test in statistical methods
• It measures the probability of association between two
• discrete attributes
• The test involves the calculation of a quantity, called chi-square
(χ)2 from the Greek letter ‘chi’ (χ) and pronounced as ‘kye’. It was
developed by Karl Pearson and has got the following three
common but very important applications in medical statistics as
test of:
• 1. Proportion
• 2. Association
• 3. Goodness of fit.
EXAMPLE OF NOMINAL ASSOCIATIONS

• SMOKING AND CANCER,

• TREATMENT AND OUTCOME OF A DISEASE,
• VACCINATION AND IMMUNITY,
• NUTRITION AND INTELLIGENCE
EXAMPLE OF MULTINOMIAL ASSOCIATIONS

• To test the association between number of cigarettes, equal

to 10, more than 10, 11–20, 21–30 and more than 30 smoked
per day and the incidence of lung cancer
• Association between incidence of filariasis and social classes
—very rich, middle and poor
• Association between parity of the mother 1st, 2nd, 3rd, 4th,
5th and above and the weight of the newborn
CONTINGENCY TABLE FOR CHOLERA

Social class Positive Negative Total Percentage

number number positive
Elite 4 76 80 5
Upper Class 20 180 200 10
Middle Class 60 440 500 12
Lower Class 144 576 750 20
Total 228 1272 1500 15.2
• Chi-square (χ2) test is also applied as a test of “goodness of
fit”, to determine if actual numbers are similar to the
expected or theoretical numbers—goodness of fit to a theory.
K is the number of classes for χ2 in goodness of fit test.
• It can be observed whether a frequency distribution fits or
differs from a *hypothetical or theoretical or assumed
distribution of qualitative data
• It may be 0.1, 0.01, 0.001 or may be somewhere in between
any of these two. The level of significance of the χ2 value may
be stated in percentages as 5%, 1% and so on, instead of as
probability of occurrence by chance out of unity given in the
table (P = 0.05 or 0.01 and so on).
• To rule out chance variation in any observed value, apply χ2
test for which three essential requirements are:
• 1. A random sample
• 2. Qualitative data
• 3. Lowest expected frequency not less than 5. If Yates
correction is applied test of proportion can be applied even if
expected value is below 30.
• Determine the expected (E) in each group of the sample or
the cell of table on the assumption of null hypothesis, i.e. no
difference or variation in the proportions of the group from
that of the universe
• Calculate for the negative class
• E2 = Total – E1 = 80 – 12.16 = 67.84; or by the same method
• Then calculate for all the classes: hat is high, middle and low
• Then find the difference between the observed and the
expected frequencies in each cell (O – E). Then calculate the
chi by the formula below
• Then sum up all the chi to get the total chi-square value

Degrees of freedom (df) = (c – 1)(r – 1)

Where c is the number of vertical columns or
classes and r is the
number of horizontal rows or groups or
classes.
 SAMPLING

• In medical studies, the sampling data are collected from a

• population or universe sufficiently large and representative of
the
• population under study.
• It is chosen by a standard sampling technique
• The population or universe must clearly be defined before
drawing a sample
• A value calculated from a sample is called a statistic such as
mean (X—), standard deviation (s) and proportion (p).
The two main objectives of sampling are:
1. Estimation of population parameters (mean, proportion,
etc.) from the sample statistics.
• 2. To test the hypothesis about the population from which the
sample or samples are drawn.
• There are two main characteristics of a representative
sample.
• 1. Precision which implies the size of the sample
• 2. Unbiased character.
TYPES OF SAMPLING TECHNIQUES

• Probability
• Non probability :
this is when the researcher chooses a set of units based on
availability or subjective judgment.
It usually does not represent the entire population objectively.
EXAMPLES OF NONPROBABILITY
SAMPLING
• Convenience Sampling involves selecting units from available
but partial lists or selecting people who are passing by a location
such as a supermarket.
• Purposive Sampling means that the researcher uses
knowledge of the field to select units to be studied.
• Snowball Sampling refers to picking an initial set of units, then
a second round of units that are nearby or have links to the first-
round selections. There may be additional rounds.
• Nonprobability sampling has the advantage of being quick and inexpensive
to
• implement.
• It is often used with qualitative research focused on in-depth exploration of
a topic on a relatively small number of observations.
• Qualitative research can complement quantitative surveys in several ways.
• It can be carried out before a random-sample survey to identify key issues,
contributing to the design of the questionnaire.
PROBABILITY SAMPLING

• It is defined as sampling in which the probability of selection can be

calculated because the selection is made randomly from a complete
list of units (indeed, it is also known as random sampling).
• The researcher defines the population and the selection method but
does not have any discretion in deciding which individual units will
be included in the sample.
• If a random sample is well-designed and large enough, it will be
representative of the population
ADVANTAGES

• It represents an Unbiased estimate

• The sampling error of sample-based averages can be
estimated
TYPES OF PROBABILITY SAMPLING

• Simple Random Sampling:

in which the entire sample is based on a draw from the
sampling frame, where each sampling unit has an equal
probability of being selected. The probability of selecting each
unit is n/N, where n is the number of units to be selected and N
is the total number of units in the sampling frame
SYSTEMATIC RANDOM SAMPLING

• there is a fixed interval between selected units. First, a unit is

randomly selected from among the first N/n units in the
sampling frame. Subsequently, units are selected every N/n
units.
• For example, a systematic random sample of 20 households
from a list of 200 households. It will start randomly from N/n
units = 10, then after every 10 households
MULTISTAGE SAMPLING

• This refers to a selection process in which the selection

occurs in two or more steps (this is also called cluster
sampling
• For example, suppose we are carrying out a national survey.
The researcher may randomly select 10 of the 50 states, 5
counties in each state, and 100 households in each county,
for a total sample of 5,000 households.
•Thank you for listening