UNIT-II

SAMPLE AND SAMPLING

DESIGN

SAMPLING
A part of population is known as sample. The method consisting of the selecting for study, a portion of the Universe with a view to draw conclusions about the Universe and/or Population is known as sampling. Sampling helps in time and cost saving. It also helps in checking their accuracy.

BASIC TERMS IN SAMPLING

POPULATION- the aggregate of all similar units

under consideration is called population. SAMPLE- it may be defined as apart of the whole under consideration. Sample is a part of the population. SAMPLE SIZE- the total number of units/observations in a sample SAMPLING METHOD TECHNIQUE OF SAMPLING

Advantage of Sampling Survey

The Size of the Population.- If the population

Amount of funds budgeted for the study. Facilities.-The extent of facilities available- staff,
Sampling is opted when the amount of money budgeted is smaller than the anticipated cost of census survey.

to be studied is quite large, sampling is warranted. However, the size is a relative matter. Whether a population is large or small depends upon the nature of the study, the purpose for which it is undertaken, and the time and other resources available for it.

access to computer facility and accessibility to population elements. When the availability of these facilities is limited, sampling is preferable. Time.- The time limit within the study should be completed in another important factor to be considered in deciding the question of sample survey.

Disadvantages of sampling
Sample selected, always is not a true representative of
the entire population. Bias is inherent feature of non- probability sampling method. Often all the characteristics of the entire population to be studied are not included in the sample. Chances of occurrence of sampling error can not be completely removed, however it can be minimized. In order to achieve near to actual results, researchers often choose large sample size, without considering the cost/budget and at times its vice-versa.

PROCESS OF SAMPLING
The decision process of sampling is complicated one. The various criteria governing the choice of the sampling technique. Purpose of the Survey: What does the researcher aim at? If he
intends to generalize the finding based on the sample survey to the population, then an appropriate probability sampling method must be selected. Measurability: The application of statistical inference theory requires computation of the sampling error from the sample itself. Degree of Precision (clarity): Should the results of the survey be very precise or even rough results could serve the purpose? The desired level of precision as one of the criteria of sampling method selected.

Sampling process (contd.)

Information about Population:
How much information is available about the population to be studied? Where no list of population and no information about its nature are available, it is difficult to apply a probability sampling method. Then exploratory study with non-probability sampling may be made to gain a better idea of population. The nature of the Population: In terms of the variables to be studied, is the population homogenous or heterogeneous? In the case of a homogenous population, even a simple random sampling will give a representative sample. If the population is heterogeneous, stratified random sampling is appropriate.

Sampling process (contd.)

Geographical Area of the study and the Size of the
Population: If the area covered by a survey is very large and the size of the population is quite large, multi-stage cluster sampling would be appropriate. But if the area and the size of the population are small, stage probability sampling methods could be used. Financial resources: If the available finance is limited, it may become necessary to choose a less costly sampling plan like multistage cluster sampling. Time Limitation: The time limit within which research should be completed restricts the choice of a sampling method. Economy: The desired level of research at minimum cost.

TYPES OF SAMPLING DESIGNS

PROBABILTY SAMPLING METHOD
Also known as chance sampling. Every item in the universe has an equal chance of inclusion in the sample. it is blind chance only that determines whether one item or other is selected. NON- PROBABILTY METHOD It does not afford any basis for estimating the probability that each item in the population has of being included in the sample. The items for the sample are selected deliberately by the researcher; his choice concerning the items remains supreme.

PROBABILITY SAMPLING

1. Probability or Random Sampling.- It is a method

wherein each sampling unit in the population has a known chance of being selected for answering a survey. It is of two types , unbiased and easy to employ. A draw of lots is a typical example. For example if a marketer wants to reward his loyal customers , he puts all their purchase order in a bag and extract one order out of it. Systematic random sampling if an interviewer wants to take a sample of ten students from a class of 60 , he can get a list of all students and then he select every 6th student out of this.

a) Simple random sampling the method is scientific

a)

PROBABILITY SAMPLING METHODS/TECHNIQUES (Contd.)

Stratified random sampling the population is first divided into various strata , usually represented by different demographic parameters such as age , gender and income. this is usually done when the strata is not equally represented in the population and study of each strata is important for marketer. i) Proportionate stratified sampling it involves the draw of an equal ratio of the sampling units as the ratio among the strata size of the population. Ii) Disproportionate stratified sampling it is the selection of sampling units in a ratio that is not equal to the ratio among the strata size in the population.

PROBABILITY SAMPLING METHODS/TECHNIQUES (Contd.)

CLUSTER SAMPLING in this the population
is divided into clusters . The division of clusters is usually done on the basis of geographic areas. Each cluster is non over-lapping. A member of one cluster cannot be the member of other cluster.

NON-PROBABILITY SAMPLING METHODS/TECHNIQUES

CONVENIENCE SAMPLING- it involves
selecting the sampling units according to the convenience of the researcher. A small intercept study is a typical example. JUDGEMENT SAMPLING- here the judgment of the researcher is used to identify the sampling unit . This is often done during the course of exploratory research wherein expert opinions and their insights is considered. QUOTA SAMPLING quotas are fixed for various sub groups in the population . The sub groups generally represent various demographic characteristics of the population, such as gender, income and educational qualification.

HYPOTHESIS TESTING

HYPOTHESIS
A Hypothesis is an assumption made by researcher that is tested to find its logical or empirical consequence. It is statement of relationship between two or more variables. The need for hypothesis is as follows: Hypothesis indicates the point of the investigation and gives direction on the study. A Hypothesis specifies the source of data, that shall be studied and in what context they shall be studied. It determines data needs. It suggests the type of research that is likely to be most appropriate. A Hypothesis contributes to the development of theory.

HYPOTHESIS
For example : 1.Students who receive counseling will

show a greater increase in creativity than students not receiving counseling.

2.The automobile A is performing as well as

automobile B.

CRITERIA FOR HYPOTHESIS CONSTRUCTION

It should be empirically testable, whether
it is right or wrong. It should be specific and precise. The statements in Hypothesis should not be contradictory. It should specify variables between which the relationship is to be established. It should describe one issue only.

TYPES OF HYPOTHESIS
We generally consider two types of Hypothesis.
1 Null Hypothesis (Ho ) While comparing two different methods in terms of their superiority, wherein the assumption is that both the methods are equally good is called null hypothesis. 2 Alternative Hypothesis (Ha ) - While comparing two different methods in terms of their superiority, wherein, stating a particular method to be good or bad as compared to the another one is called alternate hypothesis.

COMPARISION
Null Hypothesis is always specific, while

alternate hypothesis gives an approximate value. The rejection of null hypothesis involves great risk, which is not in case of alternate hypothesis. Null hypothesis is more frequently used in statistics then alternate hypothesis because it is specific and it is not based on probabilities.

HYPOTHESIS TESTING
(FLOW DIAGRAM)

IMPORTANT TERMS IN HYPOTHESIS TESTING

STANDARD ERROR- it gives an idea about the
unreliability of a sample. The greater the S.E., greater will be the departure of actual distribution from the expected one and hence greater will be unreliability. LEVEL OF SIGNIFICANCE (alpha)- a fixed %tage usually 5% which implies that a researcher is ready to take risk of rejecting null hypothesis when it was true. it is usually determined in advance before testing of hypothesis.

 CONFIDENCE INTERVAL- the max. and min.

within which the parameters of the populations are expected to lie, with the degree of confidence. TWO TAILED TEST- A two tailed test rejects the null hypothesis if, the sample mean is significantly higher or lower than the hypothesized value of the mean of the population.

TWO TAILED TEST CONTINUED

FOR EXAMPLE:- IF we take population mean (hypothesiszed)=100, if our sample mean deviates significantly from 100 in either direction then we shall reject null hypothesis and vice versa. ONE TAILED TEST- This test is used when the population mean is either lower or higher than some hypothesized value.

 TYPES OF ERRORS IN HYPOTHESIS TESTING

TYPE I error (alpha) - rejection of null hypothesis, even when it is true TYPE II error (beta) - acceptance of null hypothesis even when it is false. It is always desirable to minimize both these errors. The tests of significance are so designed so as to limit the probability of Type I error to a specified value (usually 5% or 1%).

A FORMAL STATEMENT STATING Ho AS WELL AS Ha

SPECIFY THE LEVEL OF SIGNIFICANCE

DECIDE THE CORRECT SAMPLING DISTRIBUTION SELECTING A RANDOM SAMPLE AND WORKOUT AN APPROPRIATE VALUE (on basis of empirical data collected) CALCULATE THE PROBABILITY THAT SAMPLE RESULT WOULD DIVERGE AS WIDELY AS IT HAS FROM EXPECTATIONS,IF Ho WERE TRUE IS THIS PROBABILITY EQUAL TO OR SMALLER THAN alpha VALUE IN CASE OF ONE-TAILED TEST AND alpha/2 IN CASE OF TWO-TAILED TEST YES NO Reject Ho Accept Ho

DETERMINING THE SAMPLE SIZE

The main point to be kept in mind while determining a sample size is that it should be of optimum size i.e. it should give confidence interval of the desired width and should be chosen through some logical process.

Following points should be kept in mind while determining the sample size

Nature of population Number of classes proposed Nature of study- general/technical Type of sampling Standard of accuracy and acceptable confidence level Availability of finance Other considerations

SAMPLING DISTRIBUTION OF MEAN

Sampling distribution of mean refers to
the probability of distribution of all possible means of random samples of a given size that we take from a population. some sample means will be above be the population mean, while some will be below the population mean. Thereby making up the sampling distribution.

Some important terms

Sample mean- the mean or centre of the sampling

distribution should ideally be equal to population mean. sample mean to be above or below the population mean (even if the population is not normal), hence the mean of sampling distribution is an unbiased estimate of population mean, it will be correct on average in many samples.

Mean of sampling distribution- There is a tendency for a

Standard deviation of sampling distribution- it measures

how much the sample static (mean) varies from sample to sample

 If samples are taken from a normal population,

the sampling distribution of mean would also be normal. The standard deviation (sigma) is calculated. And z-test is used to calculate the degree of reliability. than we assume that the sample mean should be as near as possible to the population mean, thereby assuming it to be near normal however we have to take a sample size more than 30.(hence normal variate z is calculated) but if we are using sample size less than 30, t-test is used.

In case of populations that are not normal, even

CONCEPT OF ATTITUDES
An attitude is a predisposition to respond
in a positive or negative way to someone or something in ones environment. It is an evaluative statement either favorable or unfavorable concerning the objects, people or events. They represent our feelings, thoughts and behavioral tendencies.

Characteristics of attitudes
An attitude has to be favorable or unfavorable. Attitudes are firmly embedded in a complex
psychological structure of beliefs. They are different from Values. Attitude is a predisposition to respond to a certain set of facts. Attitudes are always in form of evaluative statements.

Functions of attitude
Utilitarian function Value-expressive function Ego-defensive function Knowledge function

Components of attitude
Cognitive component (knowledge, beliefs)
Affective component (emotions, feelings) Conative component (behavioral aspect)

Measurement of attitudes
Attitudes are very difficult to measure
because of the nature of attitudes. since they deal with the concept of feelings and beliefs therefore they cannot be quantified as the other physical aspects such as height, weight etc. Therefore we use scales to measure different types of consumer attitudes.

Difficulty of measurement process

We use different scales depending upon the
type of item to be measured. Eg. ~Length is a weight item and hence can be easily measured with a RATIO SCALE ~Preference is an attitude and therefore can not be easily measured hence we need an interval scale. ~Happiness on the other hand is a creative perception/attitude and hence is most difficult to measure therefore we use a ordinal/nominal scale to measure it.

TYPES OF SCALES USED IN MEASURING ATTITUDES

Scaling refers to the process of measuring of attitudes. Choice of scale depends upon the type of attitude being measured. NON COMPARITIVE SCALES 1. NOMINAL SCALE- These scales are comprised of numbers that are used to categorize objects or events. A nominally scaled number serves only as a label for a class or category. It is the procedure of assigning facts/choice to various alternative categories which as usually exhaustive as well as mutually exclusive. EXAMPLE: How do you keep the books at present in the library? ( ) Subject wise ( ) Department wise ( ) Alphabetically ( ) Author wise

2. ORDINAL SCALE- These scales represent numbers,

letters or other symbols used to rank items. It can be classified not only on the basis of whether they share some characteristic with another item but also whether they have more or less of this characteristic than some other object. The ranking of certain attributes/benefits as deemed important by the respondent is obtained through this scale. EXAMPLE Rank the following attributes on a scale of 1-5 according to their importance to a washing machine. a) Company image b) Functions c) Price d) Comfort e) Design

3. INTERVAL SCALE- This scale represent

A. B. C.

numbers used to rank items such that numerically equal distances on the scale represent equal distances in the property being measured. It involves qualities description of number of respects of a product or traits of a person. EXAMPLE: Product design Poor 1 2 3 4 5 Good Brand name Lesser known 1 2 3 4 5 Well known Price Low 1 2 3 4 5 High

 RATIO SCALE- This scale consists of

numbers that rank items such that numerically equal distances on the scale represent equal distances in the property being measured and have a meaningful zero. All descriptive measures and inferential techniques are applicable to ratio scaled data. EXAMPLE: Same as previous but the scale used -3 -2 -1 0 1 2 3

COMPARITIVE SCALES

The respondents directly compare two or

more objects and may choose among them. ~ PAIRED COMPARISION SCALE- In this scale
respondents are asked to choose from among the alternatives. Attributes to be tested as well as every possible pair combination are listed out in the questionnaire and from each pair of alternatives the respondent makes his choice. For e.g.- comparing 5 brands of coffee with respect to flavors. The respondent is required to indicate his preference in pairs.

 THE TURSTONE SCALE- This scale is also called the

method of equally appearing intervals. The researcher first assembles a large number of statements, say 150 to 200 statements, about the attitude under study. Several judges are then asked to classify these statements independently into 5 piles, ranging from extremely negative to extremely positive. The sixth pile is to represent the neutral point. A final selection of 20-25 items is made based on those statements. These statements are finally presented to respondents who are asked to check those statements they agree with.

Following may be few statements I like the layout of book The book does not contain any examples The book is confusing . The language of book is simple.

 THE LIKERT SCALE- This scale is also

called the method of summated ratings is a 5-point scale ranging from strong agreement to strong disagreement. In this method, a judging group is not involved. The respondent is given a statement about a topic and he records the degree of his agreement or disagreement with the statement.

 THE SEMANTIC DIFFERENTIAL- This scale

associates probes into the intensity and content of a respondents attitude towards, say, 4 companies operating in a particular sector on parameters such as: -3 -2 -1 0 +1 +2+3 Brand name not known _ _ _ _ _ _ _ well known Not reliable _ _ _ _ _ _ _ reliable The evaluation is made using a scale of adjectives which are polar opposites, ranging from one extreme position to the opposite extreme, separated by seven equal intervals.

 MULTI- DIMENSINAL SCALING- This is a

complicated scale but can be used to scale objects, individuals with minimum information. It is characterized by a set of procedures for portraying perceptual or effective dimensions of substantive interest. It is used when all the variables in a study are to be analyzed simultaneously and all such variables happen to be independent.

To determine the perception of a consumer for 5 insurance companies to be evaluated on 2 attributes viz. convenient locality and courteous personal service
Inconvenient
B A

courteous
C E D

Not courteous

convenient

Criteria for good test

There are two criteria to decide whether
the scale selected is good or not. 1) Reliability 2) Validity

Reliability

Reliability means the extent to which the

measurement process is free from errors. It deals with accuracy and consistency. The scale is said to be reliable, if it yields the same results when repeated measurements are made under constant conditions. Criteria used: Test- retest- respondents are administered scales at 2 different times under nearly equivalent conditions

Reliability (contd.)
Alternative form reliability- 2 equivalent
forms of a scale are constructed, then tested with the same respondents at 2 different times. Internal- consistency reliability- split half reliability.

Validity Success of a scale lies in measuring what

we intend to measure. Validity tries to measure the same and therefore has more importance than reliability. Criteria used: Construct validity Content validity Predictive validity Criterion validity Convergent validity- positive correlation with other measures of same construct Discriminant validity- opposite to above

Uses of scales in Marketing Research

Helps in collecting reliable quantifiable
data for otherwise uncountable attributes such as attitudes. Help in understanding not only the presence or absence of a particular attribute but also helps in knowing the degree of the same. Scaled data is easy to analyze & interpret. Scaling also helps in segregation of data