Sociology 357—Content Analysis Assignment
Due Monday, October 13, in class
For this assignment, your goal is to use content analysis to examine how one published news
article portrays candidates or issues related to the upcoming Presidential election. Content
analysis as a method of research is useful for showing how some aspect of reality is portrayed by
some media outlet. Content analysis provides a sustained, systematic way to observe and
measure the portrayal of that reality, as opposed to the quick, impressionistic way that we
normally read news articles.
Even though content analysis is usually based on a large sample of articles, for this assignment,
you are only being asked to analyze a single article. Because you are only analyzing one article,
your results cannot be easily generalized to a larger area of interest—like how all news outlets
portray the candidates. So the goal of this assignment is to give you experience of how you
would do content analysis by using a single article as an example.
Below, I list the different steps for the assignment, and at the end, I list the components of your
report that you will write up and hand in for a grade. I recommend that you write the report as
you go. This is not the kind of assignment in which you need to know what results you get before
you write everything. It is more honest to write the report as you go; then you can avoid the
unethical temptation to write your hypothesis and research procedures in retrospect so it makes
your results look better.
Steps 1-4 and Step 6 will be done with your partner. Steps 5, 7, and 8 are to be done
individually. I will make time in class next week for you and your partner to work on this
together. On Wednesday, I will assign you your partner for this assignment. Over the weekend,
you and your partner should do Steps 1 and 2. Next Monday, I will give you and your partner
time to do Steps 3 and 4 in class. For next Wednesday, you should do Step 5 so that you and
your partner can do Step 6 together in class.
What to Do
Step 1: With your partner, find an article to analyze. You should find an article published in a
newspaper or a news magazine (like Time, Newsweek, Wisconsin State Journal, etc.) about
some aspect of the Presidential election. The article could be about a candidate, an issue, an
event—anything as long as it is related to the Presidential election. It should be published in a
somewhat “mainstream” news source, and it should have been published no earlier than June
2008. It should be between 750 and 1500 words in length. Read the article once so you can get a
sense of what the article is about, its tone, its message, etc. This will help you decide on what
variables you want to choose and what your hypothesis will be.
Step 2: With your partner, prepare the text of the article for analysis. Put the text of the article
into a word processing document and leave at least a 4” margin on the right-hand side. You leave
that part of the paper blank so you can manually code the article. Print out the article.
Step 3: With your partner, decide on your unit of analysis, and choose 3 variables and their
attributes for coding. Talk together about what variables you want to code for, what the attributes
will be, and what is the best unit of analysis to use. For example, if you have an article about
John McCain, one variable you could code for would be “valence,” and the attributes could be
“positive,” “negative,” “mixed,” and “neutral.” Or, you could choose the variable “speaker,” and
the attributes of the variable would be determined inductively, based on whether the journalist
wrote the words, or whether someone is being quoted. In the process of choosing your variables
and attributes, decide on what is the best unit of analysis: do you want to code every paragraph?
Every sentence? Every independent clause? You should be able to justify your decision in a
Step 4: With your partner, label one independent and one dependent variable and state a
hypothesis. Based on your initial reading of the article, and on the variables and attributes you
have selected, state one hypothesis about what you expect to find. Your hypothesis should be
about the relationship between two of your variables. There should be a plausible (though not
absolutely certain) temporal ordering between your variables. It is important that your
independent variable could logically “come before” your dependent variable. Remember, your
hypothesis should be testable! Based on the coding that you are about to do, you should be able
to state what results would “support the hypothesis” and what results would “not support the
Step 5: Individually, go through and code the article for your 3 chosen variables. Each unit of
analysis should be coded for each variable. You decide whether you want to do one variable at a
time, or whether you want to code for all three variables at once. I recommend clearly marking
divisions between units of analysis at first, and then using a different color of pen to write down
the attribute of each variable. Each unit should receive only one code for each variable.
Step 6: With your partner, compare your coding with your partner’s. For each of your 3
variables, you should calculate each of the following figures:
N = total number of units coded (if each unit was coded once for each variable, this would be the
same for all 3 variables)
A = number of units that you and your partner coded the same
C = number of units that you and your partner coded differently
A/N = “intercoder reliability” (the proportion of units that were coded the same)
Step 7: Individually, write the report. Each of you will write a report independently; only
sections B and C should be the same for both you and your partner. The report should
contain the sections listed below. Pay particular attention to sections E, F, and G because
those are the only sections of the report that are not listed in the “Steps” above, but they
are especially important sections!
A) Introduction: In your introduction, you should introduce the article that you analyzed.
Give the citation information for the article (author, title, source, date, page numbers, # of
words) and describe briefly what the article is about. Also, briefly talk about the context
in which the article is published. For example, you could talk about what kind of source it
is, who the intended audience is, what was going on in the campaign or in the news at the
time when the article was published, or anything else you think it is important to keep in
mind when trying to understand the meaning or significance of the article.
B) Unit of Analysis, Variables, and Attributes: Explain what your unit of analysis is and
what variables you chose for the coding. Explain why you selected those particular
variables and unit of analysis. In addition, explain how you decided what the attributes of
the variables are and justify your decision. (In other words, explain why you decided to
code for those variables in the way that you did.)
C) Hypothesis: State your hypothesis and label the independent and dependent variables.
Justify why you think the hypothesis will be true. Also write specifically what criteria
you will use to decide whether the hypothesis is “supported by the data” or “not
supported by the data.”
D) Coding Method: Describe how you coded the article—the process by which you coded
for your 3 variables. State the extent to which each variable was coded inductively and
the extent to which each variable was coded deductively. Then present the reliability
figures that you calculated with your partner and comment on what these figures tell you
about how reliable or unreliable your coding scheme was.
E) Univariate Description of Results: In this section, you should present the raw results of
your coding. Each variable should be described fully according to its attributes, and the
attributes should be indicated in both numerical and percentage terms (summing to
100%). Depending on your variables, it probably makes sense to present this in tabular
form. Each table would display the full list of attributes for one variable and the number
and % of times that the attribute was identified as a code. Be sure that you also tell the
reader in the text what the tables say. The tables do not speak for themselves! In
essence, you present this information twice: both in the text and in tabular form.
F) Bivariate Description of Results: In this section, present the raw results of a cross-
tabulation between the two variables that make up your hypothesis. Construct a
contingency table (see p. 422-423 in the textbook) that shows how the independent
variable relates to the dependent variable. Report only percentages, but make sure you
report the number of observations for each category of the independent variable in the
“Total” column. The percentages for each attribute of the independent variable should
roughly equal 100. Describe these results in the text.
G) Analysis of Results: In this section, you should interpret your results—say what is
significant about them. First, explain what the bivariate description above means, and
state why the results “support your hypothesis” or “do not support your hypothesis.”
Second, talk about the “manifest content” of the article—what we can conclude about the
meaning of the article strictly from the words that are there and the coded variables that
were unambiguous. (Hint: You should have high intercoder reliability on these variables.)
Third, talk about the “latent content” of the article—what we can conclude about the
meaning of the article if you include your interpretive “reading between the lines” and
other contextual factors that might shape the meaning of the article. (Hint: You should
have lower intercoder reliability on these variables.) Finally, state the limitations of your
analysis and what should be done differently in future research.
H) Appendix: Attach the text of your article, visibly marked with all the codes from the
Step 8: Individually, submit a Group Process Report. In it, explain how you and your partner
worked together in the steps that required cooperation, and whether or not you think you
and your partner worked well together and shared the workload equally. You and your
partner must NOT see what each other wrote, and I will not share your report with your
partner. They will be kept strictly confidential.
I will not give you a grade until after you have submitted your Group Process Report!
It doesn’t have to be long—just a few sentences that tell me about how well you and your
partner worked together. But you do have to do it!