The two functions of this section are to report the results (past verb tense) of the procedures described in the methods and to present the evidence, that is the data (in the form of text, tables or figures), that supports the results. Some journals combine the results and discussion into one section.
Before sitting down to write the first draft, it is important to plan which results are important in answering the question and which can be left out. Include only results which are relevant to the question(s) posed in the introduction irrespective of whether or not the results support the hypothesis(es). After deciding which results to present, attention should turn to determining whether data are best presented within the text or as tables or figures. Tables and figures (photographs, drawings, graphs, flow diagrams) are often used to present details whereas the narrative section of the results tends to be used to present the general findings. Clear tables and figures provide a very powerful visual means of presenting data and should be used to complement the text, but at the same time must be able to be understood in isolation. Except on rare occasions when emphasis is required data that are given in a table or figure must not be repeated within the text. Sources of help for the preparation of table and figure are Briscoe (1990), Price (in press) and Zeiger (1991).
Tables and figures must be mentioned within the text and should be placed after the related text. Photographs of subjects are often placed within the methods and should be used only if written, informed consent was obtained prior to the taking of the photograph. To preserve anonymity, facial features should be covered. If a manuscript includes a table or figure that has already been published, permission must be obtained from the copyright holder (usually the publisher) and the source acknowledged.
Confusion sometimes arises as to the difference between results and data. Results statements provide the message, that is, they interpret the data. Data rarely stand alone, they are facts, often numbers, which may be presented in their raw form, summarized (e.g. means) or transformed (e.g. percentages, ratios) (Zeiger, 1991). For example, in a hypothetical study comparing vital capacity in supine and standing, the results statement and data respectively might be, “vital capacity was decreased in supine compared to standing” and “mean (SD) vital capacity was 2.95 (0.8) litres and 3.40 (0.6) litres (p<0.05) in supine and standing respectively”. The two statements should be presented together with the results statement given first, i.e. “vital capacity was decreased in supine compared to standing, mean (SD) values were 2.95 (0.8) litres and
3.40 (0.6) litres respectively (p<0.05) in the 20 subjects”. When presenting data, the sample size (n) must be given, any missing data identified and the p-values for data that have been analysed using statistical tests must be included. The significance level (critical alpha () probability value) should be reported but in the event that the null hypothesis (H0) is accepted, the beta () probability value or statistical power should be reported. Studies in which H0 is retained are just as important to report as those in which H0 is rejected, providing such studies have an acceptable level of statistical power. The test statistic and the degrees of freedom, for example t(12) = 3.12 should also be included unless only one statistical test has been used. When citing a p value, always give some idea of the magnitude of the difference (e.g. 20 per cent increase) as a p value in isolation gives no indication of the importance of the finding. It is generally accepted when reporting results that “significant” or “significantly” refer to statistical significance (Zeiger, 1991). Thus it is unnecessary to say “the decrease in pain was statistically significant”.
The order of presentation of the results should be either chronological to correspond with the methods or from the most to the least important. The order of most to least important should be followed within each paragraph. For every result there must be a method in the methods. Careful planning of the tables and figures is important to ensure that the sequencing of these tells a story.
The results must not include a discussion of the findings, methods of data analyses and citations of references, except on rare occasions when a comparison is made of the raw data with the findings of a published study. This applies only when this comparison would not fit well within the discussion.
The discussion should be considered as the heart of the paper and invariably requires several attempts at writing (Portney and Watkins, 1993). It serves to answer the question(s) posed in the introduction, explain how the results support the answers and how the answers fit in with existing knowledge on the topic (Zeiger, 1991). This is the main section in which the author can express his/her interpretations and opinions, for example how important the author thinks the results are, the author’s suggestions for future research and the clinical implications of the findings (Portney and Watkins, 1993). In order to make the message clear, the discussion should be kept as short as possible whilst still clearly and fully stating, supporting, explaining and defending the answers to the questions as well as discussing other important and directly relevant issues. Side issues and unnecessary issues should not be included, as these tend to obscure the message. Care must be taken to provide a commentary and not a reiteration of the results. The recommended content of the discussion is given in Table 4. (Zeiger, 1991)
Table 4: Discussion
|Answers to the question(s) posed in the introduction together with any accompanying support, explanation and defence of the answers (present verb tense) with reference to published literature.
Explanations of any results that do not support the answers.
Indication of the originality/uniqueness of the work
Answering the questions should be done using the same key terms and the same verbs (present tense) which were used when posing the question(s) in the introduction. The answer must be confined to the population studied, for example if the subjects were randomly selected from a population with osteoarthritis (OA) of the knee, generalize to the population with OA knee but not to a population with knee pain from other causes. If more than one question was asked in the introduction, then all questions must be answered in the discussion. All results relating to the question should be addressed, irrespective of whether or not the findings were statistically significant. Answers to the questions that were never asked must not be included.
Support the answer(s) by reference to published work whenever possible. It may be necessary to explain the answer by saying why it is acceptable and how it is consistent or fits in with published ideas on the topic. To defend the answer, explain why it is more satisfactory than other answers and why other answers are unsatisfactory. Where the findings of the study are not in agreement with those of others, this discrepancy should be explained. The sequencing of providing this information is important; discuss the results of the present study before going on to cite the work of others. In the event that unexpected findings occur, decide whether they are of little importance or may be very exciting. Demonstrating a willingness to discuss and evaluate rival explanations for the results highlights a good discussion. To discuss an unexpected finding, begin the sentence by saying it was unexpected and then go on to give the best possible explanation.
Discuss any weakness in study design, for example, extraneous variables that only became apparent during the conduct of the study. Comment on the relative importance of these limitations to the interpretation of the results and how they may affect the validity or the generalisability of the findings. When identifying the limitations, avoid using an apologetic tone and accept the study for what it is. If an author identifies fundamental limitations the reader will question why the study was undertaken (Rudestam and Newton, 1992).
A concise summary of the principal implications of the findings should be provided and regardless of statistical significance, the issue of clinical importance of the findings should be addressed. Where appropriate, make recommendations for clinical practice based on the findings. When discussing the implications, use verbs that suggest some uncertainty such as “suggest”, “imply” or “speculate”. As all research leads to further questions, give recommendations for further research but avoid the temptation to provide a long list and focus instead on one or two major recommendations. When doing so, do not offer suggestions which could have been easily addressed within the study, as this shows there has been inadequate examination and interpretation of the data.
The organisation of the content is important. The discussion should begin by stating answers to the question and supporting the answers with the results. Do not begin with a summary of the results, secondary information (place this after the answer to the question) or indications for further research. At the need, restate the answers to the questions and indicate the importance of the research by stating applications, implications or speculations.
This section should comprise a brief statement of the major findings and implications of the study. It is not the function of this section to summarise the study; this is the purpose of the abstract. New information must not be included in the conclusions.
All important contributors should be acknowledged, for example persons who provided statistical or technical advice and assistance; the subjects; those who helped with recruitment’ and personnel who helped with the preparation of the manuscript. If the research was supported by a grant, then the name of the funding body must be included.
The authors thanks Dr Diana Hopper and Dr Leon Straker for their help in reviewing this manuscript.
The above article was subtracted from:
Jenkins S 1995 How to write a paper for a scientific journal. Australian Journal of Physiotherapy 41(4): 285 – 289