I am willing to start my PhD research analysis once I finish collecting the data, and I've been asked to prepare a proposal data analysis.
My research is about the determinants of entrepreneurial spirit and my data is mainly qualitative so i am not sure which statistical analysis to use for my results? Any ideas ? Thank you
You say "determinants of entrepreneurial spirit", that suggests pretty quant stuff. Do you mean that you have used open text boxes in your questionnaire? Within areas like psychology these are sometimes termed 'qualitative' but they don't mean that in the same sense as say an ethnography.
I think people here will need a bit more info before they can give suggestions -what did you ask, which variables are you using, what were your hypothesis etc. Clearly you can run basic descriptive stats but for a PhD I suspect your supervisor will expect modelling, of some sort.
I think you can use the multivariate analyzes usually used in ecology or other fields (CAH, AFC, DCA, ... etc.).
Hafid Elkhaddar asked a very good question that cannot be answered by one supervisor/academic since there are numerous qualitative data analysis programs out there. What I recommend is to use one that is most user-friendly. Here are a list of FREE programs that you can use: https://www.predictiveanalyticstoday.com/top-free-qualitative-data-analysis-software/
To my eyes, CAT looks good enough.
I can imagine your questionnaire will have qualities/attributes given in some scale (5 - point Likert say), Then I suppose you have many factors from which you want to choose the most significant determinants. Principle Component Analysis is a good procedure that you can employ. Some statistical software can be employed.
The normal first step in setting out to do a PhD is to get immersed in the literature. Try doing that.
For a reliable answer, one needs to know i) the research design, ii) instrument, iii) research questions, and iv) idea of hypothesis formulated, if any. The research focus sounds like a descriptive survey research which could adopt a descriptive or inferential statistical tool but which specific model depends on your answer to the four questions raised above.
I feel depending on your data set you can go for Parametric and Non parametric statistics. Sample size also matter while using any staistical tool. Qualtative data can be converted to quantitative data for using inferential statistics. So, lot of options are there in Para and Non parametric stat. I am also agree with Dr.Wilfred Omwansa Orori and Dr. Joseph Chukwuma.
It's depending how your questionnaire is structured.
Your answers are dichotomic or in scale (Likert-type)? A good study design include to measure the reliability (Cronbach's alpha, McDonald's omega, etc.), and also a multivariate analysis (if your questionnaire is new, or not validated, is recommended exploratory factor analysis. If you are studying associated factors you should think about use a logistic regression too. But, all depends how is structured your questionnaire.
I think that the statistical analysis of the data and determining the appropriate statistical techniques depends on:
· the types of variables in the questionnaire.
· the aims of the study.
· Test hypotheses of the study.
· and previous studies related to the study.
In my opinion, i think the relevant analysis you can do is descriptive statistics analysis.
Data structure differs from research to research. You need to tell us the objectives and hypothesis of the research; so as to be able to recommend the appropriate statistical tools and methods to use.
I agree with Mohamed Aboraya :
Mohamed Aboraya added a reply 7 days ago:
"I think that the statistical analysis of the data and determining the appropriate statistical techniques depends on:·(1) The types of variables in the questionnaire; (2)· The aims of the study; (3)· The test hypotheses of the study; · and (4) The previous studies related to the study."
Recommended / Share
- 1 Recommendation
I suggest that you can rely on “constant comparative analysis” (Glaser and Strauss). Constant comparison analysis is well suited to grounded theory because this design is specifically used to study those human phenomena for which the researcher assumes that fundamental social processes explain something of human behavior and experience.
Depending on the way your data are structured, you will be able to use more sophisticated tools, some of which are suggested here. In any case, however, I would suggest starting with descriptive statistical analysis, which will give the correct intuitions to follow further steps. Simple tables, comparison of means and percentages will help a lot in the preliminaries.
Depends on type of data in your questionnaire. You have dichotomous answer or quantitative data? you can do multivariate analysis but for dichotomous type of answers (like only yes or no), better to do logistic regression.
Hafid Elkhaddar,
One can speculate a lot about your goals and respond accordingly. Surely you have already read bibliographic references on the subject and have found some similarities (and dissimilarities) with yours. There you find ideas that allow you to ask something more precise. Look at the responses from Dan E Kelley and Joseph Ndozianyichukwu Chukwuma
Whenever I use a Likert type questionnaire, I prefer the frequency distribution of means and std. deviation, Students 't', and Fiducial limits. These are my preferred evaluations, but I at times also the reference "Applied Statistics and Probability for engineers (Montgomery & Runger). Remember, the statistical analysis
always depends on the type of data that you use.
If your data are qualitative I don't understand why you want to use statistical analysis that does not seem appropriate. You might rather look to analysis software such as in vivo if you really want to use a "tool", otherwise, Braun & Clarke (2006) is always a good start for thematic analysis.
- Badges
- Science topic
- Similar topics
- Biological Science
- Neuroscience