How to Write a Discussion Section

How to Write a Discussion Section

Larry Steinberg

Most new writers have difficulty writing a discussion that does more the restates the results, which, of course, is not what a discussion section should do (that is why they call it a discussion, and not a summary). Over the years, I have developed a quasi-formula for doing one that works reasonably well, but that can be easily modified if necessary. Much credit for this goes to my old collaborator, Ellen Greenberger, who taught me how to do it well. The specific examples here are drawn from an article I co-authored with Nina Mounts, on peer influence on adolescent achievement and deviance.

1. Start with a strong lead paragraph that asserts the most important finding(s) as a point of fact (i.e., "Social scientists have long questioned whether the observed similarity between adolescents and their friends is actually due to the influence of teenagers on each other. The results of the present analyses indicate quite clearly that, at least in the realms of drug use, academic achievement, and educational expectations, adolescents are affected by the company they keep. Even after taking into account initial predispositions that draw similar adolescents together, we find that teenagers become more like their close friends over time. Over a one-year interval, adolescents who are close to drug-using friends report more of an increase in their own drug use than their peers who began the year with similar levels of drug use but whose friends are not as drug-using. Similarly, adolescents with friends who are successful in school over time improve their own academic records than do their counterparts with initially similar levels of achievement but less academically-oriented friends."

2. Follow with a paragraph that brings in additional findings that make this study unique (here, discuss the fact that authoritative parenting moderates peer influence and give the specifics. i.e., "Despite the evidence here for peer influence effects, we also find that some adolescents are more susceptible to peer influence than others. Specifically....").

3. Next comes a paragraph that ties these findings back into the literature (i.e., that deals with the selection versus socialization issue and the need to look at more than one context). It usually helps to say how the findings expand on or clarify earlier work. ("This study supports the contentions of Kandel, Epstein, and others that adolescents are genuinely influenced by their friends, and that similarity between them is not simply due to selection. Moreover....")

4. Next, a paragraph that reminds the reader of the methodological strengths of the study and re-emphasizes why the findings should therefore be taken seriously. ("Among the strengths of this study were.... These features give us greater confidence that...")

5. Now, a paragraph that reminds the reader of the limitations of the study and cautions the reader against getting too excited. ("Despite these strengths, the study is limited by... Future research might address these issues by...")

6. Then comes a paragraph that deals with hypotheses that were not confirmed. (i.e., the lack of effects for delinquency and school misconduct). "Although many of the study's hypotheses received support, we did not find that.... This reminds us that peer influence may be domain specific and reaffirms the importance of examining peer selection and socialization effects separately fvor different aspects of adolescent behavior. In the realm of deviance, for example, it is interesting to note that despite the strong intercorrelations among the three measures of deviance -- delinquency, drug use, and school misconduct -- only in the domain of drug use did we find true influence effects once selection effects were taken into account."

7. Then, a paragraph that deals with the most important oddity/oddities in the findings (I think the interactive effect on drug use for boys -- or the fact that the moderating effects are more evident for girls). ("One curious and unexpected finding concerned the differential moderating effects of authoritative parenting among girls versus boys. Our results suggest that.... One possible interpretation of this is...")

8. Conclude with a paragraph that has the take-home message. ("Perhaps the most important lesson to be learned from this study is that.... An ecological perspective on peer influence requires that we....")

You should be able to do this in about five pages.

You may find that some other sequence of paragraphs 4, 5, 6, and 7 works better for a particular article, but I would make sure that 1, 2, 3, and 8 are where they are supposed to be.

This is usually enough to satisfy an editor and reviewers, although my experience is that they invariably ask to scale down the speculation and add more cautions in a revision.