The final dissertation defense, presented after all the data is analyzed and discussed, is the end of the journey for a Ph.D., and the road is never without speed bumps.  However, I find that too many candidates focus on the end of the journey and not the road map, the dissertation proposal.

A properly planned proposal, typically the first three chapters of, (1) Introduction, (2) Literature Review, and (3) Methods, can make for a much more pleasant journey, or it can pave the way for detours and major hazards on the road to success.

A dissertation proposal should be thought of as a contract between the candidate and committee.

The proposal, especially the Methods chapter, details the specific steps in how the research will be conducted. Well, at least it details the plan. Things change, but it is good to have a detailed plan.

I like to think of the Methods chapter as a recipe. Sure, maybe one will not be able to get vanilla extract and will need to substitute maple syrup. But in the end we are expecting a chocolate chip cookie, not just a generic dessert.

I often see approved proposals, signed by the committee, with a data analysis section that simply, and inadequately, reads:

The data will be analyzed using SPSS v.23 software. Descriptive statistics will include means and standard deviations of the study variables. The tests will include t-tests, ANOVA, and regression.

Not defining in detail what data will be collected, how it will be collected, the variables that will be used for analysis, the coding of the variables for the analysis, and the specific statistical tests that will be used is a recipe for adding months, maybe years, to your journey.

Why does this happen? I am sure there are many reasons. But often, especially in the online schools, committee members are overwhelmed with their workload and they do not take the time needed during the proposal phase to critically read and review a student’s proposal. Often, one committee member is designated as the methodologist or statistics expert, but they only know a bit about methodology, enough to be dangerous, and all other committee members follow his/her lead on the methodology. And too often, this results in the committee signing off on the proposal without a proper and thorough review.

This of course, makes the dissertation candidate elated, because, well, the proposal was approved! GREAT NEWS!

The happy (and unsuspecting) candidate then collects their data and then runs some numbers. But which of those statistical tests will answer the research questions? Maybe more than one will. Maybe none of them will. This is usually when a statistician gets a call to help, and sometimes there is not much that can be done. Then the methods must be re-worked to match the data collected…or it is just a wash. In either case, this delays the process, sometimes substantially.

You should consider your Methods chapter as incomplete if it  does not include:

• Specific details about the participants that will be included
• The sampling plan
• The operationalization (coding) of each variable that will be included in the tests
• The hypotheses that will be tested to address the research questions
• The specific statistical tests that will be used to test those hypotheses and specifications for those statistical models.
• For more detailed studies, tables of variable levels and operationalizations, and the tests to be used, for each research question are a definite plus!

A complete methods section with all of the details, signed by your committee and other powers that be (the AQR, IRB, etc.)  will keep these things from happening:

• The new committee chair thinks you should concentrate on a larger or smaller group of participants. Perhaps they would like you to recruit only African American women instead of all African American students for your study.
• A committee member read a journal article over Summer Break and now wants you to add social economic status to your study, but you’ve spent your Summer Break collecting data that only included gender, race/ethnicity, and marital status for your participants.
• The committee doesn’t understand ANOVA and wants you to do 10 independent samples t-tests instead.

Oh, I have dozens of stories…but you get the idea.

My advice? Details, details, details!  Operationalize all of your variables, mention specific tests you will use. Detail, ad nauseam, every step in the process. Yes, it is a lot of work and most of it will involve statistical method and theory…

But it will be worth it when you submit your completed dissertation with your Results chapter and Discussion to your committee or IRB and then they suggest (demand?) that another variable should be added, or that another test should be performed.

With a detailed and signed proposal, you can take a deep breath, smile, point to your signed proposal (contract!) and say, “Wow, that is a great idea for my next study! But for now, let’s get this one finished according to our plan.”

Your proposal is your contract. Make sure you’ve covered all of your bases and you’ll be glad you took the extra time in your travels on the road to success.