INTRODUCTION
    starts out very broad, ends up narrowly focusing on your specific study and its hypotheses

METHOD
    very narrow, detailed, technical, and specific.

RESULTS
    still narrowly focused on the specific results of the specific tests you performed to test your hypotheses.

DISCUSSION
    starts out very narrow, summarizing your results; then becomes broader as you discuss the implications
     and limitations to your research; ends up broadly conveying the 2-4 most important things you want
     your readers to remember from your research.

INTRO
I. Broad statement of topic.

Examples of initial sentences:
"Racism pervades American life."
"Because of nuclear weapons, the human race is perpetually on the verge of extinction."
Or, feel free to start with a broad question:
"How do social beliefs relate to social reality?"
"Are stereotypes generally inaccurate?"
"Why do people so frequently fall prey to self-serving illusions?"

You want a catchy, grabby initial sentence, that is backed up by a paragraph or
two that conveys the importance and power of your topic.

This very first section of the intro should end with a brief foreshadowing of your paper:
"This paper reports a /study/survey/experiment/etc. that ... [examines whatever your
study examines].

II.  General background to your specific topic. 
-- The best introductions develop some sort of mini-theory which
serves as a context for and basis of your hypotheses.

-- discussion of research relevant to your hypotheses/questions.

-- if you are replicating an old study, describe it in detail.  If any other research
   has tried to replicate that old study, describe it, too.
-- research on the same broad topic, but without clear relationship to your topic,
   probably does not need to be included.  For example, let's say you are performing
   a study of the role of sex stereotypes in person perception.  A study of the role
   of some other stereotype (race, social class, etc.) in person perception might be useful;
   a study evaluating the effectiveness of forced busing to equalize racial educational
   disparities probably would not be useful.

III.  Somewhere, either embedded in your general background discussion, or perhaps
at the end, in its own section, should be a clear statement of your hypotheses and
research questions.
 

METHOD
Different studies will require different things to be included in the method.  Here are
the types of things that should be included:
Sample (overall, broken down by demographic groups).
Refusal rate in a survey.
Experimental design (only for true experiments).
Procedures (how did you collect the data).
Questionnaires.

RESULTS
I. Preliminary analyses
This includes things like manipulation checks, deception checks, reliabilities.
It may also include descriptive statistics (means, frequencies) for all variables
upon which more sophisticated analyses will be performed (e.g., ANOVAs, correlations, etc.).

II. Main analyses
This is where you report the analyses that test your hypotheses.
Be sure to report both the result and whether the result supported the hypothesis.
If there are clearly separate hypotheses, to be tested by different analyses,
you might consider using section headings to mark off and organize
your presentation of the analyses focusing on each specific hypothesis.

DISCUSSION
I. Summarize the upshot of your main analyses.

II. Discuss limitations of your research.
-- "discuss" them does not mean
    "throw up your hands in futility and commence self-flagellating over the completely useless
     and uninformative nature of your, and indeed any, social science research because, of course,
     all research has some limitations."

      "discuss" means something like "commence a thoughtful evaluation of not only what cannot
      be concluded on the basis of your research, but also of just what (most likely) can be
      concluded on the basis of your research."

     Be especially alert to supposed weaknesses that actually strengthen your conclusions.
     Real life examples:

        1. In an experiment, finding a difference between groups with a dependent measure with poor
            reliability *strengthens* the conclusion that the groups differ on the construct being measured.
            Why?  Because low reliability makes it *more difficult* to find such differences.

        2. You perform a correlational study suggesting that stereotypes have a weak tendency to
             produce bias.  "Aha," says your stat professor, "correlation does not mean causation."
            Does this mean your study should be ignored?  Not at all, because there are only two likely
            interpretations of your study: 1) You are right, and stereotypes have only weak effects;
            or 2) You are wrong, and stereotypes have *NO* effect!  Although we might not be able to
             distinguish between 1 + 2, your study clearly shows that stereotypes do not lead to powerful
             biases, and not even to moderate ones.

        3. You perform an experiment demonstrating flaws and biases in decision-making.  "Gotcha,"
            your roommates scream, "your study was only on intro psych students."  Implicit in this
            criticism is the idea that this reduces your ability to generalize your results to most other
            people.  This is possible.  However, if anything, even your roommates would probably have
            to agree that smarter people, on average, are less likely to make dumb decisions than are
            less smart people.  If students in college, and especially students at Rutgers are, on average,
            smarter than the general population, then if anything, decision-making flaws and biases are
            likely to be *even more* prevalent among the general population than among Rutgers students.

III. Discuss directions for future research
This can be integrated with your limitations section (limitations naturally lead to directions
for future research).
Again, the more thoughtful the better.  "My sample was small, so the study should be replicated
with a larger sample" does not take a lot of thought.  If you have some idea why you did (or did not)
get the predicted pattern, how might subsequent research test that idea?  When might your
predicted pattern be larger or smaller, or even reversed (and how might research test for that)? Etc.

IV. Implications
Why is your study important?  Why should anyone care?

V. Conclusion
This should be sharp, tight, and punchy.  "This research yielded three main findings:
1. .... ;  2. .......;  3. ........."   "When taken together, this pattern strongly suggests that ...."

What is the big message or two that you want readers to come away with? Tell 'em here!

Return to Lee Jussim's Home Page

Go to the Rutgers Psychology Home Page
Go to the Rutgers Home Page