What points must be considered by a researcher while selecting an appropriate statistical test to be used in marketing research?
What points must be considered by a researcher while selecting an appropriate statistical test to be used in marketing research?
Ans.
While deciding which statistical test is to be used for the purpose of analysis, the researcher must first consider the following points:
1. Purpose of Using a Test. Whether the intention is to determine that there exits significant differences between groups of data, or it is aimed to find out whether there exists association between two or more variables ?
2. Measurement Scales Involved. Whether a nominal, ordinal, interval, or ratio scale is used? The nature of the numbers involved determines which tests can be used in analysing the data collected.
3. Number of Samples Involved. There may be a comparison being made between the sample and the population. For example, comparing the living standard of employees of Bank X with the general citizens of a particular city.
In another, case, a comparision may be made between two or more samples. For example, comparing the movie going habits of morning and evening college students of a particular university.
In still another case, a comparison may be intended to be made between units within a sample. For example, comparing the shopping habits of males and females working in a particular factory.
4. Inter-relationship of Samples. When two or more than two samples are involved, it is essential to see whether these samples are independent or are they related? For example, it will be an event of independent samples when Group I is given “50 paise off” coupons and Group II is given “75 paise off’ coupons. But, if the same group is used twice before and after giving the coupons, it is an event of related samples.
5. Many Numbers of Variables Involved. When many numbers of variables are involved, it is a test for association when criterion (dependent) and predicator (independent) variables are involved. When only two variables are involved, the techniques used are called as bivariate. On the other hand, if more than two variables are involved in the tests, the multivariate techniques are used.
When we measure the job performance (dependent variable) of management executives with their scores in the recruitment test (independent variable) is an example of bivariate technique.
Sales of shoes (dependent variable is related to style and income of people (independent variable) is an example of multivariate techniques.
These five points provide a key to the researcher in selecting the appropriate statistical test.
