Writes of Passage: Writing an Empirical Journal Article

Transcription

1 LYNN WHITE University of Nebraska Lincoln Writes of Passage: Writing an Empirical Journal Article This article provies avice about preparing research reports for submission to professional journals in general an Journal of Marriage an Family in particular. In aition to working through all the major parts of a research paper, I provie some general avice about writing, eiting, an revising. The article is intene to help new professionals improve the quality of their journal submissions an the likelihoo of successful publication. Writing research articles for professional journals is an art requiring goo research skills, a clear sense of problem, an strong writing an eiting skills. Assuming that years of grauate school have provie goo research skills, I focus on the other requirements of writing a research article. My avice reflects the issues I most often raise when I review articles an 30 years of experience writing (an revising) research articles. I review guielines for the major sections of the typical empirical research report an conclue with some suggestions about writing professionally. The emphasis is on writing for Journal of Marriage an Family (JMF), but the general principles apply across journals an substantive areas. WORKING THROUGH A RESEARCH PAPER The format for a research paper is not set in stone. Each research problem is ifferent, an Department of Sociology, University of Nebraska Lincoln, Lincoln, NE (lwhite3@unl.eu). Key Wors: research, theory, writing. the organization of the paper will epen on whether it is exploratory research rather than theory testing. In aition, authors have some latitue in eveloping a personal style. Generally, however, each article nees an introuction, a literature review, a statement of the problem, escription of metho, results, an conclusion. The organization of the piece, the titles of various sections, an the relative weight of these sections vary from paper to paper an from journal to journal, but some general guielines apply to reports of qualitative an quantitative research. Abstract An abstract shoul summarize your stuy. In a few short sentences, it shoul state the research hypothesis, the sample, sample size, ata use, an the finings. A starting sentence such as Using ata from a national sample of n women interviewe by telephone in 2002, we examine the relationship between x an y will allow you to squeeze a lot of information into a few wors. In a bare-bones fashion, without hyperbole or exaggeration, state the finings of the stuy. Examine prior issues of your target journal for abstract style an be sure to comply with the maximum length specifie by the journal (120 wors for JMF). Introuction The introuction is critical to capturing the reaer s attention an setting the tone for the paper. In approximately a single page, it shoul specify the research question, the ata to be use, an the strengths of the esign, an it Journal of Marriage an Family 67 (November 2005):

2 792 Journal of Marriage an Family shoul answer the so what? question. Too many authors wait until p. 13 to tell the reaer whether they have 20 or 10,000 cases. It is important to be specific in the introuction about what you are going to o. Implying that you are going to o the efinitive test of exchange theory in the introuction will cause trouble if the actual stuy is much narrower. By the en of the secon paragraph, the reaer shoul have an accurate iea of what is coming. It is not a place for extensive citations or literature review. Literature Review This section inclues a review of both the theoretical or conceptual framework an of the prior empirical literature relevant to the topic. This material helps the reaer unerstan the problem an how it fits with prior work. The relative weight of these two aspects varies from paper to paper, but you generally o nee both. It helps to think of the literature review as a funnel: You begin with a general overview of relevant theoretical or conceptual ieas then move through prior empirical work to a iscussion of your paper. Thus, the empirical literature review is usually narrower than the theoretical framework. The literature review shoul lea inexorably to the research questions or statement of the problem section. By the time you get to the en of the literature review, the research questions shoul be obvious to the reaer because they have been the focus of the review. Unlike a issertation, the literature review in a journal article sets up the particular research problem instea of proviing a comprehensive review. You nee neither to escribe the evolution of the theoretical perspective you are using nor to review all the work that has ever been one on topics relate to yours. You o not nee to be exhaustive. There is no extra creit for listing five citations after every statement or for ragging Bourieu into the theory. You only nee to cite an article once, not each time it is relevant to a sentence in your review. When you are on well troen groun (e.g., gener an housework), cite one or two major reviews an focus only on what is critical to your paper. Unless you are giving the material a new spin, you o not nee to review prior conceptualizing. Important criteria are relevance an balance. Make sure you cover both sies of the story, for example, work that fins no relationship as well as work that oes. The literature review shoul support the paper that actually evelops rather than the paper that was planne. If your ata lea you own an unexpecte trail, it may be necessary to examine new theoretical frameworks an new literature to try to make sense of the finings. If your original review oes not prepare the reaer for the analysis actually presente, it shoul be omitte no matter how elegant or well written. Because so much material goes into a literature review, a clear structure is essential. It may be helpful to introuce it with a statement such as three boies of research bear on this question: x, y, an z. Then review each in turn with a heaing. (Because they provie structure to the material, use heaings freely.) Statement of the Problem This section also might be calle research questions or hypotheses, but it is the crux of the paper. It acts as the link between the literature review an the metho section by laying out the issues to be examine, introucing the stuy esign, an efening the esign s appropriateness for the problem. It is both the conclusion of the literature review an the introuction to methoology. It shoul not introuce new ieas but shoul be a summary of issues you have raise in the literature review an a transition to your own stuy. Hypotheses or research questions shoul be fairly explicit, though they nee not be state formally. The etails of this section will epen on the type of research. For quantitative analysis, this shoul inclue stating the basic research questions, explaining any expectations about meiating or moerating influences, an explaining briefly what control variables will be inclue. For qualitative work, this section will lay out the questions to be explore. The etails of your stuy esign go in the metho section (below), but why you mae your major esign ecisions goes in the statement of the problem. This section is where you introuce the strengths an weaknesses of your esign an provie preliminary justification for your ecisions. The aim is not only to introuce your esign but also to help your reaer reach a positive jugment about your esign ecisions. Especially if some aspect of your stuy esign is going to raise eyebrows (e.g., you inclue only women or measurement of a key concept

3 Writing an Empirical Journal Article 793 is unusual), introuce the issue an explain it here. Explain the likely consequences of your ecision an suggest what the reaer shoul make of this. Thus, you might say, or To answer these questions, we use the ABCD ata set from Although this stuy lacks measures of E an F, it oes inclue unique ata on G an H the focus of this paper. To answer these questions, we conucte in-epth interviews with 12 welfare mothers living in the San Diego metropolitan area. Although the sample is small an geographically restricte, it... Metho The metho section provies the bare bones of the sample, measures use, a statement about missing values, an an analysis plan. In some journals (e.g., in psychology), this material is printe in a tiny font in stanarize blocks. Although JMF oes not o this, the principle is the same: This is the place for a terse statement of what you i rather than a iscussion of why you i it. You nee to provie enough information so that others with access to the same ata (an increasingly likely phenomenon) can replicate your work an get the same results. Data. This section shoul inclue sample size, how responents were selecte or recruite, an from what population responents were rawn. Information shoul be provie about levels of nonresponse or refusals, escription of patterns of nonresponse, any oversampling proceures, an a brief iscussion of any biases introuce because of the esign or nonparticipation rates. In the case of a subsample rawn from a larger sample, the criteria for selection an the number omitte shoul be specifie. Systematic assessment an reporting of selection bias is necessary. For example, are these families selective of those with better parent-chil relationships or chilren with higher attenance? The statistical sophistication of this report will epen on metho, but all stuies nee such a section. Either at the en of the ata section or at the beginning of results, provie some escription of your sample so the reaer unerstans the context of your ata. For example, are these mile-age, mile-class, marrie, White iniviuals with aolescent chilren? Measures. This is a place for simple eclarative statements with minimum justification. Specify the source of the item or scale, exact question woring, the answer categories, an how they were coe. For example, it is sufficient to state, Following Smith an Jones (2002), marital happiness is measure by 5 statements in a 4 ¼ strongly agree to 1 ¼ strongly isagree format with on t know coe 2.5. The statements were, I am happy with my husban as a frien to o things with, sexual partner,... For scales, an inicator of reliability an perhaps uniimensionality is important. If your measures are extensively use, you can provie a citation an omit extensive information about specific properties such as factor analysis. For qualitative reports, this section is less terse an stylize, but the information provie is similar: How were the concepts coe, what reasons o we have for juging the results reliable an vali, an what metho was use to analyze the ata (incluing software, if any)? Missing values. The bane of empirical stuies is missing values. Stanars for ealing with missing ata in quantitative stuies are increasingly high, but the critical step in all research is a escription of the extent of the problem, how you hanle it, an an assessment of the consequences for your finings. If the sample has 10,000 cases, but you are analyzing only 7,800, we nee to know why. If you use an imputation process, escribe this. The issue is not restricte to analysis of seconary ata sets. Whether you rop 3 of 25 observations or 2,200 out of 10,000 cases, we nee to know who was omitte, why, an the consequences of this omission for the creibility of the research. Analytic approach. If you are using a novel statistical technique, it is useful to escribe the technique briefly an give some references. If the metho has been use recently in the journal, then this probably shoul not be more than a paragraph. Results The results section is the most technical part of a research article, yet it still requires strong writing skills to hol it together. The basic principle is to tell a story an evelop the plot as you go along. Start with basic escriptive finings

4 794 Journal of Marriage an Family (orient the reaer an escribe the major characters), test the basic hypothesis, an then elaborate on it. Throughout, you nee to keep the reaer s eye on the central threa an explain how the next proceure will help you follow that threa. This means remining the reaer of the hypotheses an incluing summaries an transitions as you move through this section. When you have complete testing a specific hypothesis, briefly tell the reaer whether your supposition was supporte an what the next step is in making sense of the finings. Although the iscussion an the integration of the finings into the literature are reserve for the en of the paper, summarize as you go along. A major ecision that confronts qualitative an quantitative researchers is eciing how much to inclue an where to stop. The best guie to this shoul be your introuction an statement of the problem, which shoul have ientifie the major question an story line. This shoul help you avoi getting tangle up in interesting (but not central) issues such as relationships with control variables. If it is not set up in your theory section, it is probably a ifferent paper. Some pointers for quantitative reports. Although specifics will epen on whether you are using regression or structural equations, some problem areas that come up in many reports are worth noting: Significant or meaningful? In any ata set, but especially those with very large samples, substantively unimportant ifferences may be statistically significant at the conventional.05 level. You can solve some of this problem mechanically by using a more emaning probability level. At the other en of the spectrum, a very small sample may lack the power to fin that a substantively important effect is statistically significant. Regarless of sample size, it is important to consier the substantive importance of the finings instea of relying on statistical significance. For example, you might calculate effect sizes or report preicte scores for those with low an high values of key inepenent variables. Deal an Anerson (1995) provie a goo iscussion of this an a variety of issues relating to reporting research results. Interaction terms. Make sure that your interaction terms were hypothesize in the statement of the problem an have a reasonable justification in theory or prior work. Too often interactions appear to be a hoc ata reging. Report how many tests you ran (an how), an how many were significant. Make your text comprehensible to reaers inepenent of the tables. Not all reaers will want to stuy the tables, an the report shoul rea clearly by itself. Special pointers for qualitative reports. Because my own research is quantitative an I am rarely aske to review qualitative work, I refer the reaer to three useful articles that will help with qualitative reports. In the 1995 special issue on Paraigms an Designs, Ambert, Aler, Aler, an Detzner (1995) iscuss common errors in reporting that make qualitative reports har to evaluate an make reviewers less incline to trust them. More recently, Belgrave, Zablotsky, an Guaagno (2002) provie avice to qualitative researchers who are writing for (or at least being reviewe by) quantitative researchers. Finally, Sarah Matthews s article in this issue provies tips on organizing an writing qualitative reports. Summary, Discussion, Conclusion Although heaings may vary, articles nee to en with a summary, a iscussion, an a conclusion. The summary oes just that: It summarizes the key finings. This summary section (one or two paragraphs) can serve as the basis for the abstract, an many reaers will rely on it to get the gist of your paper. The iscussion section links the finings back to theory an prior empirical work, pointing out where your results support Smith or contraict Jones an where you explore new groun. The answer to the So what? question raise in the introuction shoul appear here. Explain how your research avances the fiel an what we know now that we i not know earlier. The iscussion also assesses the strengths an weaknesses of the stuy esign: How confient can we be of these finings? Where shoul we be cautious? At this point, you shoul bring back the issues you raise in your statement of the problem, for example, sample restrictions, causal orer, an so on. You nee to be honest, but you o not nee to make these issues the major point of the section. The iscussion often inclues the ubiquitous call for future research. Although it

5 Writing an Empirical Journal Article 795 can be pretentious to suggest that your work can set a research agena for the fiel, it is as critical to specify what you have left unanswere as it is to specify what you have learne. Finally, you nee a conclusion, a paragraph that wraps it all up, a soun bite if you will. Make this a positive statement about how your work avances the fiel an why the paper is important. This shoul be about the same topic as the introuction. References Pages of references count against a journal s limite number of pages an against the manuscript length guielines (30 pages for JMF), so citations shoul be limite to those that are necessary as a rough approximation, 4 oublespace pages for a 20-page manuscript. Think about the following issues when checking this section: (a) All citations in the text shoul be in the reference section, an no citations shoul be liste that are not in the text. Reviewers may try to follow up your references, an sloppiness here is likely to raise suspicions of sloppy work in places where it is not so visible. (b) Use the format preferre by the target journal (JMF requires APA style). (c) Reviewing the reference section is a goo check on the suitability of the manuscript for a particular journal. If your reference section inclues few or no references to JMF, for example, JMF might not be a suitable journal for your manuscript. () Scan the ates of your references to see whether your literature review has become ate since it was first written. If there are no references to the past 2 3 years, it is probably wise to pause an review the most recent work. Acknowlegments You probably i not o this without help. In a sentence or two on the title page, give creit to people who fune, avise, or reviewe you. If the paper has been given at a professional meeting, this shoul be reporte. Tables Table format varies epening on the questions an the methoology, but stanars also exist. JMF has an especially transparent set of table requirements on its Web site jmf. For other journals, check their Web sites or, failing that, follow the format of publishe versions. A few other guielines are liste below: Tables shoul be unerstanable without scanning the text. This means clear, substantive titles an variables name so they are immeiately comprehensible. Variable names such as race or gener are useless. Name ummy-coe variables with the category that is coe 1, for example Latino or Female, or put notes at the bottom of the table explaining the coing. Names shoul match the highest score. For example, if 4 ¼ often cannot pay the bills an 1 ¼ never has trouble paying the bills, call the variable Economic Harship, not Economic Well-Being. Space is at a premium. This generally means three or four tables maximum. In escriptive tables, put the epenent variable at the top. Discuss the tables from top to bottom an from left to right. If the same variables are in two tables, they shoul be in the same orer on both. Do not use programming labels in the tables or in the text, an, in general, avoi acronyms as well. Develop short variable names, such as Marital Satisfaction instea of MARSAT, an use them in text an tables. Even if you nee to say young aults 50 times, o not resort to YAs. Double-check tables for reliable formatting in electronic submission. Tables that print out neatly on your printer may be all askew when printe by the reviewers. Saving them as a pf file is one sure way to solve the problem. Supplemental Materials Generally, you shoul strive to prouce a manuscript that stans on its own without the nee for appenices or supplements. A scan of JMF shows that few issues inclue any articles with appenices. In the case of questionnaires or etaile reports of survey methoology, reference to a Web site is the best strategy. GENERAL ADVICE Focus, Focus, Focus A journal article shoul tell a story, preferably just one story. The focus shoul be state

6 796 Journal of Marriage an Family clearly on the first page an shoul remain central an obvious throughout the paper. Although it shoul go without saying, experience suggests it bears repeating: The focus shoul be the same in the abstract, the introuction, the literature review, the analysis, an the conclusion. Length Journal space is at a premium, an most journals have suggeste page limits for the entire manuscript (JMF suggests a 30-page limit). This translates to something like pages of text, 4 pages of references, an four tables. Within this general guieline, a complex paper with a complex stuy esign may require 35 pages, an a narrowly focuse paper may be too long at 25 pages. Although it may be painful, a succinct presentation will help you avoi extensive theory iscussions, unnecessary literature reviews, an tangential analysis. If you are cutting own a larger report (e.g., a issertation), start writing from scratch rather than trying to save paragraphs or even sentences. Often, the article will be narrower in scope than the issertation, so the theory an literature review will nee to be refocuse. In aition, the terse format of a journal article requires an entirely ifferent approach than the more iscursive an comprehensive writing of a issertation. Have Others Rea It an Hee Their Avice When your manuscript is reay, ask two or three colleagues, preferably those with publishing experience, to rea it. Ask them to write a review of it or at least to talk to you about it. Despite how umb you think their comments are, pay attention to them. If they i not unerstan something that you think is perfectly clear, the reviewers probably will not either. If they i not get the point, the reviewers are also likely to miss it. Even if you i explain it, obviously you i not explain it well enough. Your colleagues will unoubtely be more patient with you an nicer than the reviewers, an they may rea it more carefully. Aressing your colleagues critiques is probably just the tip of the iceberg, however, an you shoul not be surprise if journal reviewers have ifferent or more extensive comments. Eit Your Own Work Thoughtfully Learning to eit your own work is a har-won skill. Although there are many guie books, I continue to rely on Strunk, White, an Angell s (2000) The Elements of Style (or any eition). Rea it carefully an often until such rules as Use the active voice, Omit neeless wors, an Avoi a succession of loose sentences are ingraine. The American Psychological Association s (2001) Publication Manual is also a useful reference. Subject/verb agreement an correct use of commas are simple but critical requirements. More than a few mistakes give the impression of carelessness that can color the reaer s opinion about your research work. Remember that 80% of articles submitte to top journals such as JMF are rejecte. You want to make it clear in every way that you are in the other 20%. In aition to checking for technical correctness, scan your work for flabby writing an for unnecessary paragraphs, sections, an wors. The classic topic sentence remains a goo iea, an a paragraph shoul hang together an cover a single topic. Look for coherence within a paragraph an examine the transition from one sentence, one paragraph to the next. Follow the Guielines Establishe by the Journal Each journal has guielines for the text, heaings, tables, figures, an references. These may be inclue in the first or last issue of each year or, increasingly, on the journal s Web site. Following these guielines carefully is important to putting your work in the top 20%. In aition to reucing problems at the eiting stage, it emonstrates a professional approach to writing, an it signals to the eitor an reviewers that you have mae a thoughtful ecision about crafting this piece for this journal (as oppose to sening it to a half ozen journals in succession without revision). JMF has an unusually helpful Web site. Follow its links to avice on table guielines, information for authors, APA style guielines, an preparation of figures. In aition to these technical matters, it is often useful to scan recent issues of your target journal carefully for style an content.

7 Writing an Empirical Journal Article 797 Learn From Your Mistakes Learn from eitors rather than fight them. If you are lucky enough to have your work copyeite by a professional or if one of your colleagues is a goo copyeitor, o not just make the changes they suggest, but think about why they mae the changes an make a note to yourself to avoi the error in the future. The same is true with rejecte manuscripts. Although it is tempting just to file that fat envelope of ense criticism because it is too painful to rea, you really must consier carefully what you have been criticize for an figure out how to avoi it in the future. Probably, these are senior scholars who have given hours of their time to assess your work. If you can avoi efensiveness, their avice shoul be valuable. In aition, you may be able to aress the concerns an have your work accepte in another journal. Revising Manuscripts Some papers are rejecte outright. You may, however, get some version of a we ll be willing to look at it again if you revise it as suggeste. These phrases may be har to ecoe, an you shoul go over them with a more experience colleague to figure out whether this means Take care of these few things an publication is practically guarantee or We are willing to look at a revision, but we o not have much faith in your ability to make all these changes an it harly seems worth your effort. You can talk to the eitor to help you figure out this message. Even scholars with 10-page lists of publications can probably count on the fingers of one han the number of manuscripts that have been accepte outright without a request for some revision. This means that revising is a central skill. Some general avice from someone who oes not nee all five fingers to count manuscripts that have been accepte outright: The first reaing of the reviews almost always is more aunting than it really is. Give yourself a few ays to get over your initial ismay, an then analyze the reviews an the eitor s letter systematically. I make a photocopy of the reviews, an then put a number in the margin for each concrete suggestion. After I o this, I consier whether it is possible to make each change an, if it is not possible, whether a stronger explanation an more references will allow me to make a compelling argument for the way I i it. My experience is that reviewers suggestions substantially improve my papers. Although some revision requests are minor, eitors selom ask for or are foole by cosmetic revisions. Revisions often require a lot of work perhaps rebuiling the ata set, reanalyzing it, an refocusing the literature review. The en result is usually a much better paper. Occasionally, you may continue to believe that the way you i it first was just as goo if not better than the reviewer suggests. In this case, you may have to ecie whether you woul rather be right or be publishe, at least in this journal. Finally, although persistence in revision often pays off, it is vital to recognize when the major critique centers on a fatal flaw that cannot be fixe (e.g., there is too little variance on the inepenent variable, the ata are too ol). In this case, the most sensible response is sometimes to chalk this up to experience an move on to a more fruitful project. If you get a revise an resubmit that leans towar the encouraging en of the continuum, o it now! Do not wait for the journal to change eitors, the eitor to change her min, or someone else to publish it before you. Regarless of whether you ecie to revise for the original journal or to submit it to another journal, o not put it off. The longer you wait, the harer it will be to pick it up again, an the less likely you will be to get to it. Responing to a revise an resubmit inclues two parts, of almost equal weight: the revise manuscript an the response to the eitor an reviewers. Most new professionals have zero training on this aspect of research, so ask senior colleagues to let you see some of their response letters before you start. The response shoul get the same amount of careful attention as the revision itself, so eit it carefully an have someone else rea it before you sen it. This response shoul not repeat all the material change in the revise manuscript. Instea, it shoul summarize the issues an give the page number where the new material is foun. Conclusion Writing is a craft, an each paper you write will make the next one easier to write. Nevertheless, even the most senior scholars are familiar with the revise an resubmit letter, an occasionally, their papers are rejecte out of han. Sometimes,

8 798 Journal of Marriage an Family a paper one journal rejects will get accepte by another journal (after revision, maybe even a better journal), but we all have papers that become permanent resients of the file rawer. It is not the experience of rejection that istinguishes the successful from the unsuccessful writers, but the response to it. Happy writing. NOTE Alan Booth an Gay Kitson provie helpful comments uring the preparation of this paper. REFERENCES Ambert, A.-M., Aler, P. A., Aler, P., & Detzner, D. F. (1995). Unerstaning an evaluating qualitative research. Journal of Marriage an the Family, 57, American Psychological Association. (2001). Publication manual of the American Psychological Association. Washington, DC: Author. Belgrave, L. L., Zablotsky, D., & Guaagno, M. A. (2002). How o we talk to each other? Writing qualitative research for quantitative reaers. Qualitative Health Research, 12, Deal, J. E., & Anerson, E. R. (1995). Reporting an interpreting results in family research. Journal of Marriage an the Family, 57, Matthews, S. H. (2005). Crafting qualitative research articles on marriages an families. Journal of Marriage an Family, 67, Strunk, W., Jr., White, E. G., & Angell, R. (2000). The elements of style. New York: Allyn & Bacon.