Recently, we wrote about how to prepare a questionnaire for a BA thesis. Today we will explain how many questions to ask and how to choose a test sample. Thanks to this, the questionnaire for the master’s thesis (and bachelor’s thesis) will be properly prepared and the research results will be reliable and used to confirm the hypothesis.
Number of questions
How many questions we put in the survey is an individual matter. The questions should refer to variables (dependent and independent) as well as indicators characterizing them. In principle, the number of variables also affects the number of questions. One or more questions should be assigned to one variable.
This is best illustrated by an example
For example, we examine the “influence of pathology on the breakdown of marriages”. One of the variables will be the type of pathology. Now you need to set indicators that will allow you to know what kind of pathology occurred in the family of the subjects. We select the indicator “opinion of the subject about problems that appeared in the family”.
The next step is to formulate the question:
Please, indicate which phenomena appeared in your family?
- a) physical violence
- b) alcoholism of a family member
- c) criminality of a family member
- d) drugs
- e) workaholism
- f) mental neglect
We select one or more questions for each indicator. However, do not put too many questions, because there may be a situation in which the respondent will get bored, he will not read them, he will only read random answers. The optional number of questions in the survey is 15-20 plus 4-5 questions to the number (age, sex, education, etc.).
What methods of selecting the sample to use?
After preparing the questionnaire, you should consider who and how many people should fill it. This is called selection of the sample.
For the purposes of graduation work, random selection is usually used, which consists in giving questionnaires to random persons. However, it is much more advantageous to use the selection of a representative sample. The sample will be representative when the results of the study can be generalized to the population as a whole.
Therefore, instead of thinking about how many people should complete the survey, we should focus more on whether the respondents represent the community we want to study.
For example, when examining the causes of educational difficulties of pre-school children, the questionnaires should be presented to people who deal with children at this stage of development – parents, carers and pre-school educators.
The simplest method is non-random selection. For example, select people from the immediate environment for the study or select individuals that represent the general population that we want to study. Random selection can be used when the study concerns the general population, without any division into age, education or occupation. An example can be research on the “impact of store exposure on purchasing decisions”. Then you can invite accidentally met people in the store or on the street to fill in the survey.
The size of the sample
As a rule, it is recommended that a sample with a population of over 200 should be selected for testing. For the purposes of the master’s thesis, this number may be smaller, but it should not be less than 80-100. In the case when we are examining a precisely defined population, for example teachers, the test may be even smaller, because reaching 100 employees of education can be quite a challenge for the student.