How to Choose the Correct Statistics for Your
Dissertation, Thesis or Research Project


Several questions about your study need to be answered in order to select the proper statistics for your dissertation or thesis.  What type of data has been collected?  Does the data fit a Gaussian or what is called a “normal” distribution (bell-shaped curve)?  What questions, what types of comparisons will you be making in your dissertation or study?

Decisions based on these answers will help pave the way in choosing the appropriate statistical tests needed to analyze your data.  One decision that needs to be made includes whether or not to use parametric or nonparametric tests.  If your data is sampled from a Gaussian distribution, parametric statistical tests such as the frequently used t-test and analysis of variance (ANOVA) may be used.  If the data you collected comes in the form of categories or rankings, nonparametric statistical tests such as the chi-square, Mann-Whitney or Kruskal-Wallis tests may be used. 
Request Dissertation Statistics Help 

Another major question that guides your choice of statistical test also depends on the types of questions you are answering with your dissertation or thesis.  You probably will start with a basic description of your data.  What are the average or mean values of your distributions; what was the range of values, the mode and spread of scores as seen in the standard deviation? 

Next, how many independent variables are being manipulated in the study?  How many dependent variables were measurements or data collected for?  This will aid in clarifying what type of comparisons are you making: Are you comparing one group of scores to a hypothetical score; two groups of scores; or three or more?  Are you looking for how the scores are related to each other?  Are you attempting to make predictions in your study?  These answers will help you determine exactly which parametric or nonparametric statistical test is most appropriate for your dissertation or thesis. 

At times, these issues may not be as clear-cut.  For example, if your data is more unusually distributed, there may be exceptions to what tests are used.  While this process can seem rather complex, it isn’t impossible. Contact one of our dissertation statistics consultants for further help in analyzing data and reviewing your research design.
Request Dissertation Statistics Help