Your final statistics assignment is to write a short research paper that uses multivariate regression to analyze a research topic of interest to you, preferably one related to your capstone research project.
I. Research question:
Write a one or two paragraph introduction, and introduce your research question at the end of it.
II. Theories and hypotheses:
o State and explain some causal theories that may answer your research question, and how you will operationalize these into variables using your data set. Don't describe the variables in detail yet - leave that for the next section.
o You don't have to write a literature review for this assignment (you just did one), but if you know of research that relates to your topic, feel free to cite it. You will receive extra credit if you cite the research correctly and if you connect your theories and hypotheses with the existing literature.
III. Data and methods:
o First, name and describe the data set in a paragraph or two. Make sure you state the year of the survey, the nature of the sample, the population sampled from, the response rate, and any obvious limits to generalizability.
o Second, describe the variables you will use to operationalize your hypotheses.
Include a description of the wording of the questions and the answer categories, and any recoding you did.
Descriptive statistics belong in this section: Every paper should have a table that reports the minimum, maximum, mean, and standard deviation for ordinal and interval variables, and a table that reports the frequencies for nominal variables and dichotomies.
o Third, describe in a paragraph the bivariate methods you will use to analyze your data, i.e., t-tests, crosstabs and Chi Square, maximum percentage difference and/or gamma, and Pearson's R, as appropriate. (To decide what tests to use, see the table on page 1 of Statistics Guide A - Bivariate Tables and Measures of Association.)
o Fourth, give the multiple regression equation and state how you will use multiple regression to control for one or more variables.
Typically, people control for basic demographic variables such as race, income, education, and sex.
Researchers often also control to test for spuriousness and interaction among the variables they use to test their hypotheses.
IV. Findings:
o First describe the results of the bivariate hypothesis tests for statistical and substantive significance.
o Then give the complete results of the multiple regression, as found in the three SPSS output tables.
V. Discussion:
o Analyze how your results relate to your theory and hypotheses.
o If you found what you expected, say what you would do next.
o If you found unexpected results, speculate on what they might mean, and what you might do in the future to try to explain them.
Note re Assistance: This project is like a real research project, so you can get help from each other or from me. Feel free to e-mail questions or talk to me on the phone or in the online room if you want to consult with me.