More Important Points for Choosing a Good Research Proposal
In attempting to find a good proposal topic, here are some helpful suggestions:
1. Think about an outcome (dependent variable) that really interests you.
2. Think about a key predictor or predictors (independent variables) that really interest you, in relation to your outcome measure.
3. Comb through the codebooks to make sure that you are able to find – within a single codebook – variables that measure both your potential independent variable(s) AND dependent variable(s). If both of your potential variables are not to be found within a single codebook, then you can’t do that for your research proposal. The only exception is a longitudinal dataset like AddHealth or MIDUS, where the same people are tracked over time. Different datasets can be linked together from longitudinal datasets because the unit of analysis is the same, i.e., the same person in each different dataset. So you could have an Independent variable from an earlier dataset in that longitudinal series, and then a Dependent variable within a later dataset in that longitudinal series. So you are linking the same individual to him/herself, just at different points in time. You can also combine different years from the same dataset, such as the BRFSS, but you would only do this to amp up your sample size for a particular variable of interest, e.g., transgender individuals. Both your Independent variable and your Dependent variable would have to be in every separate year that you were planning to combine together from the different datasets – this type of data combining is *solely* to increase the sample size of either your Independent variable or your dependent variable or both.
4. Do a literature search of previous research using your potential independent variable and dependent variable as search terms. Quick searches may be done using scholar.google.com or using a wide variety of health based databases. These databases are accessible for free through the Sage Library. To access them:
• Go to www.sage.edu
• Click on Academics
• Click on Libraries
• Click on Databases
• Select “Health Sciences” from the “Browse databases by subject” drop down box
• There are a variety of health science databases, including CINAHL, Medline, ProQuest Health & Medicine, Public Health Database, etc…
• Comb through these databases to look for articles using search terms that include your potential independent variable and your potential dependent variable, in order to see what research (if any) has been done linking these two variables together.
• If you find articles that specifically research the connection between your independent variable(s) and dependent variable(s), see how the research was done. What were the units of analysis (individuals, organizations like hospitals, counties, cities, states, countries, etc…). If one study looks at the association between variables X and Y using individuals, and your study is planning to look at the association between the same variables X and Y but using counties, then it is sufficiently unique. For example, my mask/COVID study is unique because NOBODY has looked at the association between mask usage and COVID case rates or death rates using counties as the unit of analysis (Is the percentage of the population that wears masks across various counties associated with differences in the COVID case rate and the COVID death rate within those various counties, controlling for all other relevant factors). So even if there were 5 different mask studies out there, if they all used individuals in their study for the units of analysis, and I’m using counties, it is sufficiently different to be deemed a unique contribution. So if you DO find articles that have looked at the same independent variable(s) and dependent variable(s) that you plan to look at, pay particular attention as to HOW they analyzed it. And then ask yourself how yours would be different. If you do this step while you are still in the selection process for your proposal topic, it will save you a LOT of time later on.