Mass Communication Research Research Procedures

1. Select A Topic
2. Determine Topic Relevance
3. Review the Literature
4. Develop hypotheses or research questions
5. Data Analysis and Interpretation
6. Present the results
7. Replicate the study

Topic selection is perhaps the most important part of the process. Without a good topic research will have little or no significance. What you are looking for is a small piece of the puzzle that will help you understand the larger picture. Many academics focus on small area of research for their entire career, either a certain type of research (media effects) or methodological approach (content analysis, survey research, historical analysis, etc.). They become experts in that one area. Private researchers, on the other hand, usually have a topic given to them to study.

Where do you find ideas for topics?

1. Academic journals (like Public Opinion Quarterly; watch for a section in the end on "further research").
2. Trade publications (like American Journalism Review).
3. Magazines and periodicals (good ideas here for research questions).
4. Everyday situations (what would you like to know about how the world works?)
5. The Internet (brings the world to your fingertips, e.g. The Pew Center for Civic Journalism).
6. Archived data (news text and public opinion surveys, one example being the Lexis-Nexis database).
7. Secondary analysis (reusing old data, reanalyzing something in a different way). Data is readily available and can be inexpensive, allowing you more time for analysis, although there are limitations. You may run into bad data.

1. Ask yourself who cares? Is the problem significant for our knowledge of the world of communications, society or an industry? Does it have practical or theoretical value?
2. Narrow it down.
3. Can it be investigated? Is it an empirical question? Can we gather the data?
4. Can the data be analyzed reliably?
5. Does it have external validity? Can the results be generalized from your findings to other situations, like a larger population in survey research? Is the sample large enough and representative enough to actually report what public opinion is on a topic at a given time?
6. Are the Costs and Time worth it?
7. Is the approach right?
8. Are there ethical problems that might harm subjects psychologically?

1. Learn from previous studies.
2. Has another study been done or your topic?
3. It helps in designing method and questions.

Variable: An attribute of observations that takes on different values in different observations. We'll examine variables including class rank (freshmen, sophomores, juniors, seniors) and GPA, among others. These variables represent nominal and ordinal data, a subject we'll examine in detail in the future.

Hypothesis: A statement of a relationship among (two or more) observable variables derived from theory that can be tested directly. — Think of an example we might have on the class survey project, e.g. seniors are more interested in academics than sophomores. This relationship can be directly tested from answers provided by the survey.

Research Question:A question posed about the relationship among (two or more) observable variables. — They are used when you are unsure about a problem. What might we speculate on and ask about for the class project? Are students more interested in academics or athletics?

5. Data Analysis and Interpretation

After collecting the data, we'll use inferential and descriptive statistics to interpret the data. From this interpretation, our hypotheses will either be supported or not supported. Before announcing our results, we must be concerned with questions of validity, i.e. the degree to which our observations mean what we think they do.

Internal validity is concerned whether the study really investigates the proposed research question. Though there are 15 threats to internal validity, you should be concerned with the following:

1. History - external events that may influence a subject, e.g. taking a survey before an election, party conventions, etc.
2. Maturation - internal events that may influence a subject, e.g. people pay more attention to saliency of issues as election approaches
3. Instrumentation - does the device used to measure depend on observation? If so, observers may be affected by previous events (becoming more efficient) or fatigue (becoming less efficient).
4. Mortality - subject drops out of study
5. Experimenter Bias - mistakes made in observation, data recording, mathematical computations, and interpretation.
6. Evaluation Apprehension - subjects are essentially afraid of being tested, or reluctant to tell real answers.

External validity refers to the degree to which we can generalize from data gathered in one situation to other populations in other situations. There are three threats to external validity:

1. Subject selection - was there a representative sample? to what populations can it be generalized (e.g. all UT students or only communications majors?)
2. Testing - subjects have knowledge of test or being tested, want to perform well for surveyor. Pretesting may increase or decrease the sensitivity of the subject to the testing device.
3. Experimental conditions - if the experimental setting changes, will the results? does the presence of observers, instrumentation or

7. Replication
The ideal research can be repeated and with the same results. As Kerlinger and Lee (2000) note, when the original relation between data is studied again, it might be studied with different participants, under somewhat different conditions, and with more or fewer variables. However, the relationship should still be present to be supported. REMEMBER: A hypothesis can be supported by the available data, but it is not proven.

If you don't understand something in this Web note, please e-mail Dr. Sitton.

INDEX

SYLLABUS

SCHEDULE

e-MEDIA

COMM-STOP