Methodology: Sample Question 1

Based on your reading list, I would like you to summarize and discuss what is already known about the methods you will employ in your dissertation. How did you choose them, and what methods did you reject. Why?

Compare the methods you have chosen to other reasonable methodological alternatives, and indicate why you believe your methods are better for this problem. Your answer should draw upon the readings and upon your previous work in this area. Also, discuss whether your methodology is different from the usual methods employed to study newsrooms, and if so, why.

Finally, address the weaknesses of your methodology. Are there ways it could be improved, or new methods that could be employed, to study newsroom processes?

Methodology: Sample Question 2

Take your dissertation proposal or another proposal in your area of interest and discuss the methods you would use in the project. Compare those methods to other reasonable alternatives and indicate why you believe your methods are better for this problem. Your answer should draw upon other work that has been completed in the area and should point out how your methods are the same or different from the usual methods used.

Methodology: Sample Question 3

Answer both questions:

  • You are interested in conducting a content analysis examining photographic coverage of George Bush prior to and after the attacks of Sept. 11, 2001. Explain the key decisions that you would need to address to ensure a valid and reliable study. What types of statistical tests would you incorporate into the study? Why?
  • You are interested in conducting an experiment examining the effects of different photographic depictions of George Bush prior to and after the attacks of Sept. 11. Explain the threats to validity that you would face with such a study. What can you do to limit these threats?

Methodology: Sample Question 4

Propose a research question and design the experiment to investigate the impact of any feature(s) of media on cognitive/emotional processing. Your experiment must include at least one between subjects factor and at least one within subjects factor as well as both psychophysiological and self-report or behavioral measures of dependent variables. Make sure you provide conceptual and operational definitions of all dependent variables as well as details concerning how you will manipulate your independent variable(s).

  • Discuss the strengths and weaknesses of physiological and self-report or behavioral measures used in your experiment. Specifically discuss reliability and validity issues associated with each measure.
  • How do you think the media feature(s) investigated in this experiment will affect the physiological measure(s)? Why do you believe this effect will occur, and what is the conceptual meaning of the effect for your dependent variable? For instance, what level of your independent variable (IV) might lead to cardiac deceleration? Why do you believe this level of your IV will lead to cardiac deceleration and what does cardiac deceleration mean for the dependent variable heart rate is being used to measure?

Methodology: Sample Question 5

In survey research response rates have been declining steadily across the last several years. Why is the response rate so important? What can be done to increase the response rate for (a) phone surveys, and (b) Internet surveys? What other major problems do survey researchers face in ensuring that their data are valid and reliable?

Methodology: Sample Question 6

You are conducting a survey of newspaper sports editors to examine their job satisfaction. Explain:

  • How you would select a sample of respondents.
  • What variables you might want to include as control variables.
  • The logic behind using control variables and why you chose to use them in this study.
  • Statistical tests that you may want to use in the data analysis.

Methodology: Sample Question 7

(2 hours) Describe two research projects that address the following features. One study should be about an experiment in which several groups are compared. The other study should be descriptive in nature. For both studies identify the IV (independent variable) and the DV (dependent variable). If there is more than one IV and/or more than one DV for either study, please identify this, too. Tell briefly, what each study is about, what statistical procedures (including assumptions) could be employed to address the research question(s) and what are some likely results. You can pretend that you have no logistical problems (e.g., IRB, obtaining research subjects, etc.)

Methodology: Sample Question 8

(2 hours) Testing has relied upon classical test theory (CTT) for nearly 100 years. Tests are developed and scores are reported according to this theory. More recently, developments in statistics have advanced item response theory (IRT) to the fore—in fact, its prominence is such that most national testing programs (e.g., SAT, GRE, MCAT, LSAT, etc., and many school district and state testing programs) are built on IRT. Describe and contrast these two measurement theories. In particular, cite some (1) characteristic features [reliability and validity], (2) advantages, and (3) disadvantages. You may also relate some contexts where one theory or the other may be appropriate.

Methodology: Sample Question 9

The first part of this question deals with the distinction with moderating and mediating variables. Given that some of the most useful theoretical models in mass communication research involve some sort of intervening variables, the way such variables mediate and/or moderate effects is important.

  • Define mediating and moderating variables. How are they different?
  • Explain how you would statistically test for a single moderator.
  • Explain how you would statistically test for a single mediator.

The second part of this question deals with multiple linear regression. Provide an example, perhaps from your own research, of a hierarchical regression model. For the purpose of this question, your model must have at least 4 IVs (i.e., predictors). What does a hierarchical regression model give you that standard or stepwise regressions don’t? Suppose you had a case in which, for a single model, predictor A’s unstandardized regression coefficient was larger (absolute value) than the unstandardized regression coefficient of predictor B, and yet, B’s contribution to the DV was statistically significant and A’s contribution was not. Explain how that could be (or could it be?).