Article on Writing
Family Practice Research Journal
Vol. 8. No. 1. Fall/winter 1988, pp3-16.
Writing and Publishing
Maurice A. Hitchcock, Ed.D.
The purpose of this article is to assist tardily medicine researchers in writing organized and concise articles. While the basis of arty good article is the quality of the study it describes, the chances of the article being published and communicating results effectively to readers can be improved through effective writing. The IMRAD (Introduction. Methods, Results, and Discussion) format is used to instruct authors as to the contents of an article. Suggestions for improving the writing and publishing of articles include: outline the article, format the article for submission in the style specified by the journal, obtain reviews from coauthors and colleagues, and use an editor.
Research, Family Medicine, Faculty Development, Writing Skills
Each day researchers have their articles rejected by professional journals. These articles represent the culmination of one or more years of work designing the project, collecting and analyzing data, and writing the article. Many of these authors are in university tenure track positions that require publications for promotion and tenure. The rejections of their papers represent failures for these authors, not just in terms of lost time, but also in decisions about their salary, promotion, and tenure statuses.
Why is it that so many articles are rejected by journals? Some say that a high percentage of rejections is inevitable since journals have limited space and receive more articles than can be published. This does not explain, however, the fact that some authors get almost all of their articles published. Journals do not reject good articles because they have too many. Good articles describing quality studies are accepted for future publication. Articles that get rejected: 1) describe poorly designed or conducted studies, 2) are poorly written, or 3) do not conform to the journal's guidelines or areas of focus. Two of the reasons, then, that many articles get rejected are that they are poorly written or submitted inappropriately to a journal.
Researchers need to write better articles and submit them so that they have the best chance of being accepted for publication. Research has indicated that many successful authors learned to write and publish articles by collaborating with research mentors.1 Most researchers in family medicine, however, have not had an opportunity to work with a mentor because there are not yet sufficient umbers of research mentors in the discipline. Culpepper and Franks (1984)2 found only 43 physicians spending 50% of their time in research activities at family medicine residencies and departments nationwide. Many family medicine researchers have had to teach themselves about writing and submitting articles for publication.
The purpose of this article is to assist family medicine researchers to White research articles that communicate effectively myth readers and have a good chance of being accepted for publication by a journal. To clarify the instructions provided, examples are drawn from drafts of a manuscript being written by the author entitled “Factors Influencing Student Selection of Family Practice Residency Programs in Texas.” It should be noted that quality articles are written in numerous ways. This article describes one such method, the author’s. Other good sources for Whiting articles are listed in the “References” section of this article.3-13
Contents of a Research Article
The most popular format used for writing research articles today is the IMRAD method, named for the sections of the article (Introduction, Methods, Results, and Discussion). It may be helpful to think of the sections as analogous to the system model (Input, Process, Output, Feedback). The “Introduction section of the article describes the “input” into the research project (What is the subject of the study? What do we already know about the subject of focus? What is the Purpose, question, or hypothesis of the study?). The “Methods” section describes the “process” of the research (Who were the subjects of the study? How was the project designed? What measurement instruments were used? How was the data analyzed?). The “output” of the project is described in the “Result” sections of the article. Finally, the “Discussion” section provides “feedback” about the subject of focus in the article (How do the findings of this study relate to findings of previous research on the topic? What issues should be researched next on this topic?). Authors should take care not to mix the contents of the sections (e g, do not discuss implications of a study’s findings in “Results” or tell methods in “Discussion”). A more complete description of the contents of each section of the research article follows. A summary of the contents and suggestions is provided in Table 1.
Table 1. Contents of a Research Article
Purpose: describe the topic, previous literature and purpose of the study
- Arouse the reader's interest
- Avoid truisms
- Stress the importance of the topic
- Expand upon the first sentence
- Describe the subject of focus in the article
Review of Literature
- Review what is known about topic from prior studies
- May be as short as one paragraph or as long as several
- Draw generalizations from the literature
- Provide a rationale for the study
- Lead into a statement of the purpose of the project
Purpose of the Study
- Can be questions, hypotheses, or simple declarative statement
- Often final paragraph in the “introduction” section
Purpose: allow readers to judge the validity of the study & its findings
- Purpose is to allow judging of whether sample was biased
- Include a description of the participants of the study
- Describe how the sample was selected
- If a survey, the response rate should be given
- Purpose is to judge if extraneous variables biased study
- May include formal name of design (Case-Control study)
- May include techniques used to eliminate bias (Blinding)
- Purpose is to describe the study instruments and their qualities
- Describe pilot-testing, reliability, validity, item types
- Describe how study data were analyzed and reported
Purpose present data outcome of the study
- Not necessary to review all data collected
- Review study questions to decide on data to include
- State data as facts without interpretation
- Delay comment on data until the “Discussion” section
Tables & Figures
- Use tables for data too lengthy for tent
- Use figures to illustrate data for clarity
- Tables & figures must be coherent on their own
- Tables & figures should supplement, not duplicate, the text
- Use tables & figures judiciously; they are costly to publish
- Journals often suggest eliminating tables & figures
Purpose:describe the meaning and implications of the study results
Sections: Content varies greatly from article to article
Interpretation of Results
- Authors often describe the meaning of study outcomes
- Be direct; avoid qualifying phrases such as: “It appears that…”
Limitations of Study
- Purpose is to assist reader in judging validity of study findings
- Describe any limitations or biases that affected study outcomes
Comparisons to Previous Research
- Authors often describe how outcomes relate to previous research
Summarize, Generalize or Conclude
- Often the closing paragraph of an article
- Authors often:
- Summarize the research findings
- Generalize results to the larger population of interest
- Draw conclusions about the topic of focus
- Suggest issues for further research
The first sentence of an article should arouse the reader s interest in the topic of focus. A good way to start is by stressing the importance of the topic to the reader. Avoid truisms such as: “Family medicine is a new and exciting discipline,” or “The sedative effect of antihistamines is well known.” Here is an example of how an article might begin:
The residency matching process may be the single most important event in family medicine education, both for students seeking positions and for residency programs recruiting applicants....
The first paragraph should expand upon the first sentence and precisely describe the subject of focus for the article.
... Residents who do not match the philosophy, expectations, and personnel of a program are at risk of moving to another resident or leaving family medicine for another specialty. This risk, along with a desire to acquire quality training, motivates students entering family medicine to apply to an average of 11 programs and interview at eight before ranking their program preferences for the National Residency Matching Program (NRMP).Ref x The staffs of family practice residency programs also spend large amounts of time each year interviewing and ranking applicants to acquire quality residents for their programs. Inordinate amounts of time and effort are invested in the matching process, yet little is known about the variables that affect the matching of students with programs. This matching could be improved by better understanding the factors that influence students’ selections of particular residency programs.
Review of Literature
A review of the literature may be as short as a paragraph or as long as several. Its purpose is to review what is known from prior studies about the topic of the project. The review should draw generalizations from this literature, rather than describing the research that has been done. It should provide a rationale for the study and lead into a statement of purpose for the project.
Factors that influence students’ selections of particular residency programs have been investigated in three previous studies, one each in psychiatry,Ref orthopaedics,Ref and family medicine.Ref All of these studies suffer from methodological problems that make generalization of the results troublesome.... Some of the common factors that influenced students’ selections of programs include: students’ contacts with residents of the program; the quality of the faculty and their commitment to clinical instruction of residents; students’ contacts with the program director or chairman; the geographical location of the program; and an organized schedule of interviews at the program. In family medicine the most influential factors were current residents’ opinions of their program, the residency director, and the geographical location of the program; salary, benefits, and university affiliation of the program were rated lower but influential in students’ decisions.
Purpose of Study
Whether done through study questions, hypotheses, or simple declarative statement, the function of this paragraph is to describe the purpose of the study. Often the statement of purpose is the final paragraph in the “Introduction” section of an article.
The purposes of this study were: 1) to identify the relative influence of various factors in students’ selections of family practice residency programs in Texas; and 2) to determine if the factors that influenced students’ selections of various types of residency programs differed.
The traditional purpose of the “Methods” section of a research report was to provide other researchers with a blueprint of the study so that it could be replicated and the findings vernier. It is doubtful that many studies published today could be replicated entirely from the “Methods” description. The more important use of this section today is to provide readers with the information necessary to judge the quality of the study design and procedures and hence the validity of the findings.
There are generally four areas that should be addressed in the “Methods” section of an ample.
The author should provide a complete description of the participants of the study, including how they were selected or assigned. If a survey was constructed, the rate of questionnaire return should be given and the follow-up procedure described. The reader should be provided sufficient detail to judge whether the study outcomes could have been biased by the sample studied.
The subjects of this study were the new residents (127) selected through the NRMP by the 22 Texas family practice residency programs in 1987. Three residency programs did not take residents through NRMP or did not indicate those taken through NRMP. Responses from these programs were eliminated from the study. This left usable responses from 19 of 22 family practice residency programs. A total of 116 residents taken through NRMP responded, yielding a response rate of ninety-one percent. The majority of responses (66) were received from 12 Medical School Affiliated programs, the largest group of program type in Texas. The next largest group of responses (32) were from the four Medical School Based programs. The smallest number of responses (18) were received from the three Medical School Administered programs. There are currently no Community Based programs in Texas as defined by the NRMP criteria.Ref
A description of the experimental design should be provided to allow assessment of whether bias could have been introduced into the study by way of extraneous variables. This description may include the formal name of the design (e.g. clinical trail, case-control study) and may include a descriptor of special techniques used to eliminate bias (e.g.. blocking, blinding).
Study questionnaires were mailed to directors of each program for administration to residence during the first week of July, 1987. This timing was considered crucial to capture residents’ reasons for selecting programs while still recent in their memories and before residents were influenced by other experiences at their programs. Directors were called after one week to encourage prompt reply to the study.
This section should contain a description of the measurement instruments used in the study and their quality. If an instrument was designed for the study, a description of its construction (e.g.. pilot-testing, reliability/validity, types of items, should be included. If a standardized instrument was used (e.g., Myers-Briggs Type Indicator), the author should describe the reliability and validity of the instrument for populations similar to the ones included in the present study.
The study questionnaire was constructed by the authors and consisted of 52 rating-scale items. Residents were asked to indicate the importance of each item in the selections of their residencies by circling numerical responses ranging from zero to six. A graphic rating scale was placed at the top of the questionnaire for residents to refer to in rating all items. The scale had descriptive anchors at 0 (“not an important factor in my selecting this residency”), 2 (“a slightly important factor . . . “), 4 (“an important factor . . . “), and 6 (“an extremely important factor . . .”). Questionnaire items were constructed primarily from the literature review.Ref Other items were suggested by faculty and residents in the pilot-testing of the questionnaire.
The questionnaire was pilot-tested on three first-year residents in each of three family practice residency programs of different types in Texas during the Spring of 1987. After completing the pilot questionnaire, residents were interviewed by the faculty member or director administering the instrument. Questions were asked about: 1) the content of the questionnaire (Which, if any, items are ambiguous?; Which items are redundant?; Are there some items that you considered important which were not included?); 2) the scale (Is the scale clear, easy to understand and use?); 3, the process (Was the questionnaire easy to complete?); and 4) general comments (Any comments for improvement of the questionnaire?). The final questionnaire was derived from revisions based on feedback from this pilot-testing.
A description of how the data were analyzed should be included in this section.
Two methods of analysis were used on the study data. First, mean scores of residents’ ratings of each item were computed, and items were rank-ordered in terms of relative importance in influencing students’ selections of programs. These data are reported for each program type and for all programs combined. Second, mean scores of each item rating were compared across different program types using Tukey’s Studentized Range (HSD) tests to determine if factors influencing students’ selections of types of programs differed.
The results section of the article should contain a presentation of the data, but not necessarily all of the data obtained in the study. It is often useful to review, the research questions as guides for selecting the data to include in the article. The data should be stated as facts without interpretive or qualifying information. Delay comment on the data until the “Discussion” section.
Bad Example: Clearly, students selected various types of resident programs for very different reasons. For example, students who selected Medical School Based programs were drawn by the university affiliation of the program and the expert training they would receive, whereas students who selected Medical School Affiliated programs depended more on recommendations of others about the program and were more concerned about the attitude of other specialists in the community toward family practice....
Better: Significant differences were found in the level of importance of seven items in students’ consideration of different program types. (See Table 1) While it was rated an important factor by students who selected all program types, the “philosophy of training” was rated significantly less important by students who selected Medical School Based programs as compared to students who selected Medical School Affiliated programs and all programs combined (F=3.l3; p<.05)....
Tables and Figures
Tables are used for presentation of data that would be difficult or lengthy to explain fully in text. Figures are used to illustrate data for purposes of clarity. Each table and figure, with its legend, must be coherent on its own, not dependent on the narrative for understanding. Tables and figures should supplement, not duplicate, the written text. If the data can be explained fully and clearly in text, there is no reason to use a table or figure. Journals often suggest elimination of as many tables and figures as possible, because they occupy considerable space and are expensive to publish.
The content of the “Discussion” section varies greatly from article to article. There are, however, some common elements in many articles.
Interpretation of Results
Authors often explain the meaning of the study results for the reader. Be direct; avoid qualifying phrases such as “it appears that ...” or “our data suggest the possibility ..."
The factor that attracted most students to particular family practice residency programs related to the opportunities for students to get quality family practice training (e.g., commitment of family medicine faculty to clinical instruction of residents, quality of the family practice clinic experience, morale of staff, competence of family medicine faculty, philosophy of training, overall curriculum of program). At the other extreme, factors that were least influential in attracting students to programs related to the extrinsic rewards that one might be more interested in when seeking employment (e.g., salary, benefits, cost of living).
Limitations of Study
To assist readers of an article in judging the validity of the research results, many authors describe any limitations or biases which may have affected the outcome of the study.
While this study produced the intended ranking of factors influential in students’ selections of family practice programs in Texas and a comparison of these factors across program types, several cautions are offered for interpreting the results. First, while the rating scale used in this study was developed from previous research and pilot-tested on first-year residents, the reliability of the instrument was not checked. Second, the present study was of students selecting family practice residency programs in Texas; therefore, the results may not be applicable to programs in other areas of the United States ...
Comparisons to Previous Research
Many authors describe how the outcomes of a study relate to previously published research on the topic. This previous research is likely to have been described in the “Introduction” section of the article.
In comparing findings from this study to previous research, some similarities and differences were noted. In terms of similarities, the importance of “commitment of faculty to clinical instruction of residents,” and the “morale of the staff” were each identified previously by Bunch et al.Ref. . . In terms of differences, Sacks et al.Ref noted the importance of females having “at least one woman interviewer” and the importance of an “organized day of interviews,” while Bunch et al.Ref pointed out that “future prospects for fellowship training” and the “opportunities of residents to do research” were important factors in students’ consideration of residency programs. These issues were rated as less important in the present study ...
Summarize, Generalize or Conclude
As a closing paragraph for an article, authors often summarize the research findings, generalize the results to the larger population of interest, or draw conclusions about the topic of focus. Sometimes authors suggest issues needing further research.
In summary, three major accomplishments were achieved by the present study in extending knowledge of students’ selections of family practice residency programs. First, the present study confirmed many of the factors previously identified in other specialty areas as influential in students’ selections of family practice residence programs but also identified some factors unique to students selecting family practice programs. Some of the characteristics that are influential in students’ selections of any residency program, regardless of the specialty, are: the commitment of the faculty to clinical instruction of residents; the morale of the staff of the program; students’ exposure to the residents of the program; the professional courtesy extended to applicants; and the leadership of the program. Factors unique to students’ decisions of family practice programs include: ...
The last section written is an abstract of the article. Many journals limit the length of the abstract to 100 or 150 words. Abstracts should describe the research that was done but also include the findings of the study. Specifically, the abstract should summarize the purpose, methods, and results of a study, as well as the conclusions drawn.
This study identified the relative influence of 52 factors on students’ selections of family practice residency programs in Texas. All students (127) selected through the National Residency Matching Program in 1987 by 22 residency programs in Texas were surveyed. Based on a 91% response rate, factors which influenced students’ selections of programs were rank-ordered according to program type (e.g., Medical School Based). Factors most influential in students’ selections of programs overall related to the quality of training they would receive: “commitment of family medicine faculty to clinical instruction of residents”; “quality of family practice clinic expenence”;“morale of staff”; and “competence of family medicine faculty.” The least important issues involved extrinsic rewards: “salary”; “benefits”; and “cost of living.” Significant differences were found in the level of importance placed on seven factors by residents who selected different program types, indicating that students selected various types of family practice residency programs for deferent reasons.
Format the references in the style specified by the journal to which you plan to submit to the article. References should not be included that have not been read by the authors or cited in the article. Where writing the first draft of the article, you may want to list the references in shorthand where they are used in the article (e.g., Johnson et al., 1985). Later, when the revisions have been made in the draft, the references can be numbered sequentially and a formal reference list constructed.
Writing and Submitting Articles for Publication
Several suggestions are offered below for improving the process of writing articles and submitting them for publication.
Outline the Article
Work from an outline when preparing the manuscript. Computer outlining programs allow the transfer of outlines directly into word processing. The first draft of the article can be constructed directly from the outline created without duplicating work. An article written from an outline is more likely to follow a specific logic and convey intended meanings to readers. A good way to start the outline is to write the headings of the sections of the report (Introduction, Methods, Results, Discussion), then come back to the “Introduction” section and, following the plan described here, outline the first sentence, first paragraph, review of literature, statement of purpose, and so on until the outline is complete. Prior to outlining the “Results” section, it is often useful to construct the tables to be included in the report. This may help you to describe the results of the study and draw conclusions.
Format the Article for Submission
As you conceptualize the article, scan journals that might be appropriate for your article. Look for similar topics, analyses, and length. Most journals clearly explain the types and topics of articles they are interested in publishing in the “Information for Authors” section, which is often included in issues of the journal or can be obtained by request. Begin formatting your article in accordance with the guidelines specified by the journal you select as you are drafting the article. By selecting a journal early, the author can write to the readership (audience) of the journal and avoid the extra work of cutting out content not allowed by a specific journal format or pertinent to its audience.
Obtain Reviews from Coauthors and Colleagues
When the first draft is complete, the article should be submitted to all authors for review. The principal author is usually responsible for writing the first draft, circulating drafts to coauthors, and integrating suggested changes into the article. This review process may take as many as three iterations to get the article to the point where all authors are satisfied.
After the authors have reviewed the article, submit it to external reviewers. These reviewers may include faculty in your program/institution, fellows, and experts on the topic selected from the review of literature.
Use an Editor
There is no certification or degree process for editors. Anyone can be an editor. Look for someone with a strong background in grammar, sentence structure, spelling, and punctuation. Also look for an editor with a successful background in reviewing articles for publication. Ask for a vita. You may find that such an editor is employed at your university to assist faculty with their publications.
An editor’s review of your paper may be limited to grammar and spelling or may extend to formatting and conceptual design of the article. Give your editor specific instructions concerning the type of review you want. After you work with an editor several times, this becomes less important. Each editor uses a slightly different marking system in reviewing papers; be sure to request a glossary of the proofreading markings used by your editor. As suggested for reviews by authors and colleagues, the principal author is usually responsible for liaison with the editor and integrating any suggested revisions into the draft of the article.
Select Coauthors of the Article
The decision as to who will be authors of the article you plan to publish is perhaps the most difficult decision you must make as a senior author, as it has potential for conflict. If you use statistical consultants or technical advisors and do not include them as authors, the next time you ask for their help they may not be available. Perhaps this helps explain the recent trend toward multiple-authored papers. The following guidelines were synthesized from the author’s experience and several articles on the topic of authorship of papers.14-17A thorough reading of these articles is recommended for a complete understanding of the issue.
- Problems over coauthorship are less likely to occur when decisions about authorship, or at least tentative decisions, are made before a study is started. These decisions should be initiated by the person who is to be most responsible for designing and executing the project (senior author). Any conflicts over authorship or content of the paper should be resolved among co-workers with the senior author chairing the discussion.
- A good place to clarify authorship and individual responsibilities for the project is in a pre-study protocol. It is an excellent idea to write a description of your project which includes the design of the project (e.g., research objective, sample selection, independent and dependent variables). This protocol can also describe the roles of individual researchers and authors’ names in tentative order for the eventual publication. It should be noted that individuals may change roles on the project or drop out completely. Changes in authorship may have to be made during the drafting of the article.
- A person should be included as coauthor if The contributes significantly to the scientific formulation (e.g., design) or execution of a study, or the writing of the paper. In addition to participation the study, other writers have included the criteria of “understanding—”. . . an author should be able to take public responsibility for the content of the paper. An author should be able to indicate why and how the observations were made, and how the conclusions follow from the observations.’’
- The author listed first should be the person who actually did most of the work on the project and wrote most of the paper. All other authors should be listed in order of decreasing contribution to the study.
- Individuals should be consulted as to whether their names should be included as coauthors.
In addition, all coauthors should receive a final draft of the paper before it is
submitted for publication.
As an alternative to coauthorship, persons who make lesser contributions to the study (e.g.. advising about the statistical analysis, arranging for research subjects) can be acknowledged for their assistance. Permission from persons to be acknowledged is necessary to avoid the appearance of endorsement by the acknowledged persons.
Submit Article for Publication
When submitting an article to a journal, include a cover letter to the editor. This letter should identify the type of article you are submitting (e.g.,”full-length,” “communications”) and the author who will correspond with the journal editor. Include a statement indicating that the contents of the article have not been published elsewhere before and that the article is not being considered by other journals at the time.
You should receive acknowledgement that your paper has been received by the journal within two weeks after its submission. If you do not receive such an acknowledgement, call or write the editor; your paper may have been lost in the mail. The review will take from six weeks to three months—typically six to eight weeks. If the review extends beyond the usual time, call the editor for a status report.
You will be notified by letter from the editor of the decision regarding publication of your article. If the article is rejected, the reasons for rejection will be specified. Most journals include a so am of the reviewer’s comments. These comments can be very helpful in revising the article to submit to another journal. You may receive a letter from the editor which states: “The editorial review board has decided not to publish your paper in its present form. If you will make the following changes, however, your article will be reconsidered.” This is a conditional acceptance letter. Often, if you make the changes and resubmit the article as suggested, it is accepted for publication without further peer review.
If an article is rejected flatly by a journal, do not resubmit the article to the same journal. Change the formatting, revise the article based on the reasons for rejection, and submit the article to another journal. Do not be discouraged. Some of the author’s colleagues have submitted articles to as many as eight journals before getting them accepted for publication.
Review Galley Prints
Some journals allow only 24 hours for review of galleys (preliminary copy of the article for the journal). This is you last opportunity for changes prior to publication. You are not usually allowed to make substantial changes during this review, only punctuation, spelling, typo corrections and significant errors of fact. Be sure you review the galleys well. It is embarrassing to have colleagues point out mistakes in your articles. Get more than one person to review the galleys. Often you are too close to the article to see the mistakes.
1. Bland DJ, Sch Ritz CC: Characteristics of the successful researcher and implications for faculty development. J Med Educ 1986: 6l:22-31.
2. Culpepper L. Franks P. Family medicine research: Status at the end of the first decade. JAMA 1984: 250:63-8.
3. Scherger JE. Taylor RB: Writing a medical article. In Taylor RB. Munning KA (eds.): Written Communication in Family Medicine. New York. Springer-Verlag, 1984:33-41.
4. International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. Ann Int Med 1982: 96 (Part 1):766-7.
5. Day RA How to Write and Publish a Scientific Paper, 2nd ed. Philadelphia. ISI Press. 1982.
6. CBE Style Manual Committee: CBE Style Manual: A Guide for Authors, Editors and Publishers in the Biological Sciences, 5th ed. Bethesda M.D., Council of Biology Editors, Inc., 1983.
7. Barclay WR. Southgate MT, Mayo RW: Manual for Authors and Editors: Editorial Style and Manuscript Preparation. 7th ed. Los Altos CA. Lange Medical Publications. 1981.
8. Sheen AP: Breathing Life Into Medical Writing. St. Louis. CV Mosbv Co., 1982.
9. Strunk W Jr.. White EB: The Elements at Style, 2nd ed. New York. Macmillan Co.. 1972.
10. King LS: Why not Say It Clearly: A Guide to Scientific writing. Boston, Little, Brown and Co., 1978.
11. Fishbein M: Medical Writing: The Technic and the Art, 4th ed., Springfield IL 1972.
12. Geyman JP. Bass MJ: Communication of results of research. J Fam Prac 1978: 1:113-20.
13. Howie JW. VWhimster WF, Paton A. et al.; Writing and speaking in medicine. Br Med J, 1976: 3:1113-25.
14. Fotion N. Conrad C: Authorship and other credits. Ann Intern Med 1984; 100:592-4.
15. Editor: Responsibilities of coauthorship. Ann Intern Med, 1983: 99:066-7.
16. Burman K: Hanging from the masthead: Reflections on authorship. Ann Intern Med 1982; 97:602-5.
17. American Psychological Association: Publication manual. 3rd ed. Washington D.C. American Psychological Association. 1983.