Research Methods Survey
I. Short Introduction (paragraph or so) 5 points
• What are you trying to find out? Why is it important to study this topic? What makes this topic interesting, different? What are the variables? How did your findings from your interviews help to narrow or expand the scope of your topic, and/or theory, and/or hypothesis, and/or variables? You can also include what others have said as well (previous lit. review). What is your research question/hypothesis? The RQ/hypothesis should end this section of the paper.
II. Methods (1 page) 10 points
• Who are you targeting in your survey and why? (referent population). What sampling design did you use? (self-select, purposive, accidental, etc.) Why this design?
• Describe the nature of your survey questions (background, IV questions, DV question, level of measurement). You do not need to provide the actual questions or explain each question, just an overview of the questions you chose and why?
• How did you administer the survey? (online, paper) How many people did you ask? How many actually participated? (response rate) What is your final sample size (N)?
• How did you code your independent and dependent variables? What variables did you collapse, if any? How did you collapse this variable(s) and why in this manner? (average of responses, variance of responses) If your DV was created by using multiple questions, which questions (indicators) did you use to measure your DV (concept).
• How you code can be changed to your liking, usually by using theory after looking at your responses. For example if you are using the Likert scale, you may want to change it from strongly agree…strongly disagree to grouping the agrees into one and the disagrees (or neutral) into another (1=agree, 0=did not agree). You could also change it to 1=strongly agree, 2=other. Using your sociological imagination will determine the best way for you to code.
III. Findings (3+ pages) 20 points (10 for univariate and 10 for bivariate)
• Univariate Analysis—create a frequency table summarizing the mean, median, mode, and standard deviation of the independent and dependent variable. One for each variable. You should also include a graph (bar graph, pie chart) for each. SPSS can do all of this for you. The variables you use for the univariate should be the same two you will use for the bivariate.
• Are there any interesting patterns? Is there variance and are the responses what you would expect? What do you think influences, or causes, the variance in the responses?
• Bivariate Analysis—provide a cross-tabulation chart (2×2 or 3×3 or 2×3 or 3×2) with the corresponding chi-square and p-value you get from SPSS, using the same two variables from above. When you cross tabulate your data, label the rows and columns and include column percentages.
• What is your p-value? (0.05). Is your chi-square significant? What is the null hypothesis? Do you reject, or fail to reject, the null hypothesis? Are the two variables associated? What does it mean to achieve statistical significance? Are you committing an error of commission (type I) or an error of omission (type II), if any? Are there any patterns/trends?
• Integrate the tables and charts into the essay; don’t just tack them on at the end. You should label them as well (ex. Table A: Race and Income or Figure 1: Income by Race, etc.).
IV. Discussion (about a page) 10 points
• Do your findings confirm or refute your original hypothesis stated in the introduction? What do your findings mean? Here you can relate back to the literature review from earlier in class and/or what you mention in the introduction. What were the limitations of your study? (sample size, sampling design, study population) If you were to continue with this topic, how would you revise your question and/or design? Better sampling design? New research question?
• Overall, what did you learn about your research topic and what are some things you could do to improve your survey?
V. Codebook (4 points)
• Your codebook is basically an explanation of how you coded the responses for each of your questions. The codebook is not written in essay form. It should appear in an appendix-like section. It is the information you would input in the “values” column in variable view in SPSS. You should be able to do this for every (or almost every) question. Often when you code you use (0) for the negative, and (2 or higher) for other possible responses. It might look something like this:
Age 1=under 18, 2=19-25, 3=25+
Sex 1=male, 0=female (not male)/ or 2=female
Generation 1=first, 2=second or more
GPA 1=below 2.5, 2=above 2.5
Ethnicity 1=black, 0=not black/ or 2=white, 3=Asian, etc.
Church Attendance 1=often, 2=sometimes, 3=rarely, 4=never
Capital Punishment 1=strongly agree, 2=agree, 3=disagree, 4=strongly disagree/ or 1=agree, 0=did not agree
VI. Survey (1 point)
• Provide one blank copy of your survey. Please do not submit hard copies of all the questionnaires you distributed and retrieved.