The researcher should decide where the manuscript will be submitted before writing the final paper. The expansion of the scope of practice in the health professions has been accompanied by a proliferation of publications serving specialized areas of practice. The choices are numerous and selection of the appropriate one deserves careful thought.
Some journals have a clearly defined focus with priorities explicitly stated. This focus is often stated in a journal's masthead or instructions to authors. For example, the Journal of Rehabilitation Research and Development (JRRD) has a complete statement that clarifies the kinds of papers that are appropriate, including the priorities:
JRRD responsibly reports the results of rehabilitation research relevant to veterans. Our goal is to publish cutting-edge research that enhances the quality and relevance of Department of Veterans Affairs rehabilitation research and disseminate biomedical and engineering advances. Priority areas are prosthetics, amputations, orthotics, and orthopedics; spinal cord injury and other neurological disorders (with particular interest in traumatic brain injury, multiple sclerosis, and restorative therapies); communication, sensory, and cognitive aids; geriatric rehabilitation; and functional outcome research. JRRD accepts national and international submissions.1
When a journal's focus is not so obvious, the contents of several issues of that journal should be read to determine if a particular study is consistent with the subject matter and type of research that the journal tends to publish. It is an unfortunate waste of time, effort and perhaps money to make the wrong choice and to have a manuscript returned because it "is not suitable for publication" in a particular journal. This is almost verbatim what the rejection letter will say.
Another consideration in selecting a journal is the readership. The product of research should reach the people who will best be able to use the information. If, for example, a study documents the functional outcome of an orthopedic surgical procedure, the report should be in a journal that orthopedists read. If, on the other hand, the study focuses on the postoperative physical or occupational therapy intervention, journals devoted to these professions will be more appropriate.
Every journal publishes Instructions to Authors that must be followed in the preparation of a manuscript. Although the general format of a research report is fairly consistent in medical and scientific writing,2 each journal has its own particular rules about organization and length of a manuscript; preparation of tables, illustrations, or graphs; and method of reference citation. Look carefully at articles published in the journal to follow the format. Failure to follow the instructions may be a reason for rejection; or, at least, the manuscript will be returned for corrections.
Authors must expect delays in responses when their article is being reviewed. Some journal editors are more rigorous about turnaround time than others. Standard policy for scientific journals states that authors should submit an article to only one journal at a time. This protects journals from conflicts in copyright. If an article is rejected, the author can then submit it to a different journal.
The content of this chapter is in keeping with the CONSORT3 and STARD4 statements that are designed to assure that papers related to intervention or analysis of diagnostic tools are complete. These statements were described more fully in Chapters 9 and 27 (see Tables 9.3 and 27.3 and Figures 9.1 and 27.4).
Structure and Content of the Written Research Report
The sections of a research report are the abstract, introduction, methods, results, discussion and conclusion, as shown in Table 33.1. The introduction and methods sections serve the same purpose as in the project proposal; that is, they describe the rationale for the study and the specific procedures used to collect the data (see Chapter 32). Although the content of these sections will be similar to the proposal, the author will have to do some serious editing to fit the journal article format. The "forward-looking" statements must be changed to past tense because the project is now completed. The last three sections of the article will be completely new.
TABLE 33.1STRUCTURE OF A JOURNAL ARTICLE FOR REPORTING RESEARCH ||Download (.pdf) TABLE 33.1 STRUCTURE OF A JOURNAL ARTICLE FOR REPORTING RESEARCH
|Section ||Should contain |
|Abstract || |
Overview and purpose of the study
General description of methods
Highlights of results
Statement of significance of results
|Introduction || |
|Methods || |
Criteria for and methods of subject selection
Description and number of subjects
Measurement methods and data collection techniques
Data analysis procedures
|Results || |
Narrative description of statistical outcomes
Tables and figures that summarize findings
Statements to support or reject hypotheses
|Discussion (and Conclusions) || |
Interpretation of statistical outcomes
Discussion of clinical significance of outcomes
Importance of the work
Comparison of results with work of others
How results support or conflict with theory
Critique of the study limitations and strengths
Suggestions for further study
|References || |
Abstract. Most journals require an abstract of the report that the author usually prepares after the manuscript is complete. The abstract summarizes the content of the article including the purpose of the study, the number and type of subjects, the basic procedures used, a summary of the results and the major conclusion. The abstract must be concise. The prescribed limit may be 150 words, occasionally less. Readers will refer to the abstract first to decide whether to read the complete report. Computerized retrieval systems store author-written abstracts. Therefore, they must be able to stand alone, despite their brevity.
Introduction. The introduction can be drawn from the statement of purpose, the background and specific aims included in the research proposal. As in the proposal, the introduction of an article should provide a description of the research question and the context within which the author intended to answer it. After reading the first one or two paragraphs of the introduction, the reader should have a clear understanding of the problem being studied and why it is important. The literature review should reflect the relevant background that is necessary to support the theoretical rationale for the study, and should provide sufficient information for the reader to understand how the research question will be answered. The introduction should end with a statement of the specific purpose of the study, delineating the variables that were studied and the research hypotheses or guiding questions that have been investigated in the study.
Methods. The methods section should begin by describing the subjects, including how many were studied, what criteria were used to recruit them, how they were selected and how they were assigned to groups. Relevant characteristics of subjects, especially age and sex, should be summarized using means, ranges and frequencies. Most journals require a statement documenting that subjects read and signed an informed consent form and that the appropriate committees approved the project.
The methods section continues with a description of equipment and data collection procedures, presented in chronological order so that the reader can follow the procedural flow of the project. If the measurement or treatment procedures are standardized and well known, they can be described briefly and the author can refer the reader to the original sources for a more detailed description. When manufactured instruments are used, the company name and address should be cited. Operational definitions should be provided for all variables, with the intent that someone could replicate the data collection procedures. Many researchers develop a written protocol that they use as a guide during data collection to be sure that all procedures are followed properly. This protocol can easily serve as an outline for this section of the paper. Diagrams, photographs and tables can clarify and simplify the presentation of methods. For example, demographic information and special characteristics of subjects can be summarized in a table, and photographs of a unique procedural setup may make a lengthy verbal description unnecessary.
The methods section ends with a full description of the procedures used to reduce and analyze the data, including specific statistical procedures. If unique or new statistical methods are used, they should be referenced.
Results. The results section contains only a report of results, that is, a narrative description of exactly what happened in order of importance relative to the specific aims or hypotheses of the study. In the course of the study, the researchers may have gained considerable amounts of information, but unless it relates specifically to the stated purpose of the project, such information should not be included in this section. If one simple hypothesis has been proposed, the results section may be stated in a few succinct sentences.
The outcomes of statistical tests must be included to demonstrate or support the statement of results. Although the inclusion of calculated values, degrees of freedom, and the significance level is important, the narrative portion of the results section should emphasize the variables of interest rather than just statistics. For example, in a study of gait comparing elderly and young women, the statement "The differences in step length were significant, t = −3.13 (p < .01)," is not as meaningful to the reader as "The elderly women demonstrated a significantly shorter step length than the younger group (t = −3.13, p < .01)." When detailed statistical or descriptive information related to the study variables is needed in the paper, it is usually easier and often clearer for the reader to refer to tables or graphs that summarize such information.
Two major principles should guide the structure of the results section. One is that tables and figures should not duplicate the narrative; that is, if the author includes values for group means and standard deviations in the body of the text, there is no need to repeat them in a table. The author can refer the reader to the tables and figures for details and should only summarize these details in the text. The reader should be able to understand the results without referring to the tables and should be able to understand the tables without referring to the text. Therefore, the tables and figures should complement but be independent of the text. Second, the author should not discuss results in this section. Statements related to how this information could be applied to practice or interpretation of outcomes should be left to the discussion section.
Discussion. The discussion section is the heart of a research report. It reflects the researcher's interpretation of the results in terms of the purpose of the study and the outside world. This is the part of the paper in which the author can express opinions. The author should comment on the importance of the results, limitations of the study, suggestions for future research and clinical implications.
The commentary about the importance of results should not be a reiteration of the results section, but should focus on alternative explanations of the observed outcomes, emphasizing how they either support or refute previous work or clinical theories. All results should be addressed, including those that were not statistically significant. The author should provide perspectives on the applicability of results to practice or further study.
The limitations of the study, including possible extraneous variables that could have affected the outcomes, should be identified and explained. Some of these factors may have been identified before the study began and others will have become evident during the course of data collection or analysis. These may include small sample size, attrition of subjects or lack of subject adherence to the protocol. The author must consider the relative importance of these limitations to the interpretations of results. It is essential that the author delineate all major extraneous factors so that the reader can examine the results realistically.
Every research endeavor leads to further questions. Sometimes, these questions arise out of the expressed limitations of a study and the need to clarify extraneous factors. In clinical research, alternative methods exist for studying the same or similar research questions and these may need to be examined. Given the results of a study, the author may want to reconsider a particular theory and how it may be applied. Suggestions for future research will develop from these ideas and should be expressed.
Authors should acknowledge the immediate or potential applicability of results to clinical practice. Their perspectives on the clinical relevance of studies are important whether the research focus is primarily on theory, applied science or clinical effectiveness.
Conclusion. The conclusion is a brief restatement of the purpose of the study and its principal findings. It is often written in such a way that the author states the deductions made from the results. Phrases like "the results of this study indicate" and "this study demonstrates" serve to link the summary of results and the meaning of those results.
References. The style of citing references throughout the text and in the listing of references at the end of a manuscript must follow the Instructions to Authors. Many journals use the style suggested by the published "Uniform Requirements."2 Others may follow the American Psychological Association.5
The Internet is an important new source of references. Journals may eventually include instructions for such citations. The Modern Language Association of America (MLA) has published a full array of methods for citing documents that have been obtained from websites.6
Tables and graphs should be used in the results section of an article to facilitate explanation of statistical findings and to provide visual explanations. To be effective, the tables and graphs must follow guidelines for the specific journal and general considerations for developing visual materials. The narrative portion of the paper will often present general descriptions of findings, and the tables and figures will present the details.
The customary table has five components: the title, column headings (horizontally displayed), row headings (vertically displayed), the "field" within which the data are arranged by columns and rows, and footnotes. Journals will format tables according to their style. The well constructed table of research results will present numeric or descriptive data demonstrating the relationships between independent and dependent variables. The title should identify those relationships.
Whether data are oriented vertically or horizontally may depend on the size and format requirement of the journal; however, logically, related numeric data should be presented in the columns, particularly when they will be summarized with totals or means and standard deviations in the last row. The sequence of column headings should progress logically from left to right based on the order of events. In the display of pretest and posttest data, the pretest should come first. The column headings should specify what was measured and the units of measurement, such as "(degrees)." The source of column data must be identified by the row headings, such as subjects by number or code when individual data points are displayed, or with labels, such as "control" and "experimental" when the table represents grouped summaries.
Footnotes may be used to present p-values, to explain abbreviations, or to cite references. A journal's instructions to authors may specify the style for sequentially labeling footnotes. Some may use small italic Arabic letters (a,b,c). Others may use symbolic keys. For example, the Uniform Requirements for Manuscripts Submitted to Biomedical Journals describes a common symbolic sequence: ∗ (asterisk), † (dagger), ‡ (double dagger), § (section mark), || (parallels), ¶ (paragraph symbol), and # (number sign).2
Specialized tables are constructed to present statistical test summaries. Many examples are presented in this text. A tabular presentation of a frequency distribution is shown in Table 17.1. A typical analysis of variance table, called a "source table," is shown in Table 20.1.
Graphs provide a visual demonstration of research results. Trends, relationships and comparisons may be presented more effectively and more concisely by constructing a graph than by writing a detailed text. Graphs can be drawn in a number of ways. Frequency data are commonly represented using a histogram or a frequency polygon. Figure 17.1 in this text shows examples. In a histogram the bars are contiguous, and in a frequency polygon, the data points are connected by lines. A pie chart graphically can display the proportional distribution of selected characteristics of a whole sample where the percentage of each characteristic is drawn to scale as a piece of the whole pie. A bar graph, which is a series of separate bars, may be used to show frequency or magnitude data derived from separate samples, such as control versus experimental group values, or experimental events, such as pretest and posttest values.
In constructing graphs, the author must pay careful attention to the scaling of the units of measurement. Graphs are intended to represent meaningful trends, relationships, or comparisons; therefore scales should be realistic and drawn to illustrate important, true differences in the data. They can, however, be drawn to present a false impression—either exaggerating or diminishing real differences. For example, Figure 33.1 displays shoulder abduction range of motion before and after treatment intervention. The data in both graphs are the same: a mean of 100 degrees for both groups before treatment, 110 degrees for the control group and 120 degrees for the experimental group following treatment. The magnitude of change appears to be greater in Figure 33.1A as compared with Figure 33.IB. Especially considering that normal shoulder abduction range is somewhere between 160 and 180 degrees, Figure 33.1A seems to be an exaggeration of the comparative effectiveness of the treatment intervention.
Two bar graphs showing shoulder abduction range of motion before and after treatment for two groups, illustrating the effect of vertical axis scaling. (A) Scale from 90 to 125 degrees; (B) Scale from 60 to 180 degrees.
In deciding what content to present in narrative form or in tables or graphs, the author should consider a general rule that each element of the results section must stand alone and each must contribute to the complete and accurate presentation of the research findings.
Unique Elements for Reporting Systematic Reviews
The general principles and procedures for preparing papers for publication apply to reporting systematic reviews. The content of these reports, however, should include descriptions of the special methods that are part of the review process, as described in Chapter 16. The reader is encouraged to look for examples in the Cochrane Database of Systematic Reviews.7
Converting Theses and Dissertations for Publication
Most theses and dissertations are written on the basis of a six-chapter format: statement of the problem, review of the literature, methods, results, discussion and conclusion. Journal articles contain the same elements, but are honed carefully by authors to avoid all but the essential content presented succinctly. The full document that is prepared to meet degree requirements is usually too long and over-referenced to be acceptable as a journal article. The challenge for the student is to condense content considerably while retaining substance and meaning.
This is not a simple undertaking; therefore, the student must consider seriously the decision and commitment to proceed. The student should consider whether the outcome of the project will offer journal readers useful information or new insight and perspectives. Only if the answer is affirmative should the process of editing begin.
The review of literature, presentation of methods and results, and discussion will require the most work. The review of the literature in the full document is usually extremely lengthy. The author who writes for publication is obliged to be much more selective, including in the manuscript only those references that provide necessary, contemporary information and explicitly relevant background for the work.
The author of a research paper is also obliged to clearly and explicitly describe the method of study; however, a "blow-by-blow" description is usually unnecessary in a journal article, and can be very tedious for the reader. Often, minute details of the protocol are delineated in appendices of a thesis. For the journal article, the content of appendices must be explained in a few sentences.
The process of scientific writing "is not primarily a 'literary' effort, but is an exercise in organization and clarity of expression."8 The final written report should be strong, reflecting the objectivity and logic of the research project. We have reviewed the customary format and content of the research report to guide the author through the structural elements; however, the readability of the report depends on the author's skill in communicating with precision. Readers often complain that research reports are boring and difficult to follow. These complaints may have nothing to do with the subject but with sentence structure or the flow of ideas—the personal writing style of the author.
We will not present a "style manual" in this text, but will highlight a few common problems and current issues that interfere with clear writing. A selection of useful references for developing and improving writing skills are cited at the end of this chapter. Even the most accomplished author may benefit from consulting such references, especially when the early drafts of a paper seem cumbersome.
People First. Our patients or clients are indeed people first! It is incumbent on all of us to be sensitive to this basic human issue. We can easily reflect this in the description of subjects who have been studied. In choosing words about people with disabilities, the guiding principle is to refer to the person first, not the disability. In place of saying "the disabled," it is preferable to say "people with disabilities." This way, the emphasis is placed on the person, not the disability. As examples, "stroke patients" are patients who have had a stroke; "learning-disabled children" are children who have learning disabilities. Words like invalid, suffering, and victim should be avoided. Attention to people-first language demonstrates an underlying respect that should be reflected in our scientific publications.9
Active versus Passive Voice. A sentence written in active voice is powerful and concrete. Passive voice tends to make sentences ponderous and dull. The writer must create sentences that make the intended point most clearly. At times, the choice between active and passive voice is not so obvious. The following sentences illustrate the choices.
|Passive: ||One hour was spent by the raters to observe the patient's movement patterns so that the number of changes in static posture could be documented. |
|Active: ||The raters spent one hour observing the patient's movement patterns to document the number of changes in static posture. |
The emphasis is different in these two sentences. The amount of time is emphasized in the passive example, whereas the raters and their activity are highlighted in the active example. In addition, the active example is shorter by six words! Consider the following examples as well:
|Passive: ||Increased tension throughout the upper extremity and neck is produced by constant pain in the wrist. |
|Active: ||Constant pain in the wrist increases tension throughout the upper extremity and neck. |
Note the difference in focus of these sentences. If the author is addressing the potential for dysfunction at sites remote from the pathology, then the first (passive) example may be best. If the author is developing a rationale for eliminating wrist pain, the active example is more appropriate.
Passive voice may be appropriate when the subject of the sentence is unimportant or the object or action should be emphasized. For example, "Patients were randomly assigned…" conveys an important action, and the subject (who assigned) may not be important.
Superfluous passive expressions, such as "it has been suggested…" or "it is thought that…" usually distract the reader because they dilute the strength of the message. Who suggested? Who thought? In citing the work of others, authors should acknowledge the "who," for example, "Jones and Brown suggested.…" In discussing present work, the authors might write "Our preliminary results indicate.…" Using active voice in such cases is direct and clear.
One special case should be discussed—the use of first-person active voice. For many years, authors went out of their way to avoid using first-person active voice with the notion that to use it detracts from the "scientific," "objective" nature of research reports. Now, the use of first person is acceptable in selected instances. When authors (researchers) are emphasizing their own actions, experiences, assumptions or opinions, their writing may be more readable and indeed more accurate if they say "We think the logical interpretation of this finding is…" or "We found that this technique is…." On the other hand, overuse of "I" and "we" can be intrusive, calling unnecessary attention to the authors, especially when purely scientific information is being conveyed. Presentation of techniques, procedures, and results requires attention to what was done and how, not who. Therefore, it is stronger to say "The subjects were asked to complete the questionnaire," rather than "I asked the subjects to complete the questionnaire."
Simplicity of Language. In conversation, we tend to use expressions and phrases that are spontaneous, but often superfluous to the point we are making. Many such expressions will be found in the early drafts of written work because of the natural effort to "speak" the text. Authors must, however, remain cognizant of the need to be concise in scientific writing. Many of the elements of creative writing that we learned in school, designed to create metaphors and add color to our words, should be discarded for scientific writing. Many authors try to use different words for the same concept to avoid being repetitious; however, where one word will make the point best, it is better to be repetitious than to be unclear or ambiguous. Adjectives and adverbs are especially useless for describing scientific findings. There is no need to say that an outcome is "very practical" or "extremely useful." It would be sufficient to be practical or useful for clinical care.
The use of expressions should also be tempered for scientific reports. Although it is certainly more interesting to read a paper that is written with variations in sentence structure, the purpose of an article is to communicate findings, not to create poetry. Here are a few examples of complex phrases that can usually be avoided:
|in light of the fact that ||= ||because |
|with the exception of ||= ||except |
|in spite of the fact that ||= ||although |
|is designed to improve ||= ||improves |
|due to the fact that ||= ||because |
|was found to have ||= ||had |
|immediately prior to ||= ||before |
In early drafts, there may be redundant phrases, such as "exactly identical" and "grouped together," and unnecessary qualifiers, such as "blue in color" and "end result." Correcting these kinds of errors is easy, if the author is looking for them.