Hi
I have attached a sample draft of a master level thesis. The data collection is through questionnaire and the first hypothesis is simple an average i.e yes/no question. The data analysis method intended for other hypotheses is student t test to determine the differences with regards to gender, education and occupation.
Does this combination of an average and student t test seems appropriate for this kind of setting at masters level? or is there any other test i can run on this data?
Thanks in advance.
Hi Hafiz Hasnat Ahmad, I have a few remarks about your research.
Your target group is the people of Pakistan who have/had a client relationship with any Pakistan bank.
1) Response rate. You plan to launch an online survey via Facebook and WhatsApp. This considerably narrows the response rate further dawn. You get only respondents who are willing to answer the questions from an already selected group. You would be satisfied with an N greater than 100 respondents. This seems to be a very low percentage of the target group. You should be very careful in drawing conclusions from such study.
2) Hypothesis H0 (a): The digital financial services available in Pakistan are sufficient to cater the market needs. You plan to take the average of the yes/no answers. I guess you want to test that the average is larger than .50; that the majority would say yes. You could test this with a one-sample t-test. If you have only a few questions (say some 10 questions) in your survey (related to the question whether the financial services are sufficient) you could also run a chi-square test for the goodness of fit. You could e.g. test whether the proportion all are .50 (on all questions).
3) For the remaining hypotheses, I would suggest you run Between x Between x Between factorial ANOVAs instead of multiple t-tests; the design would be: Gender x Education level x Occupational status.
Sorry, in your study your Factors are not independent. Yet, run ANOVAs instead of t-tests when there are more than 2 levels of Education and or Occupation. Check the distributions of the dependent measures with histograms and Shapiro tests.
If obviously not normal, run nonparametric tests such as the Kruskal-Wallis test.
- Badges
- Science topic