How to write a paper fo r a scientific journal ■ Sue Jenkins Publication in a reputable, peer-review ed journal should be the goal of every researcher, as this provides the m ost effective and perm anent means of dissem inating inform a­ tion to a large audience (Cole 1994, Portney and W atkins 1993). W hen hum an subjects participate in research, it is on the understanding that they are assisting with the creation and dissem ination of know ledge, presenting researchers with the responsibility to com m unicate the outcom e of their research (Cole 1994). The aim of this paper is to provide guidelines to assist with the preparation of a m anu­ script for a scientific journal. Before w riting a first draft, it is im portant to establish that the topic of the m anuscript is likely to be consistent with the focus of the journal. This may be clearly stated within the journal or may be determ ined by examining several recent issues. H aving selected a journal, it is essen­ tial to carefully read and follow the guidelines for authors published within the journal or obtained directly from the editor or publisher. These guidelines are usually very spe­ cific and include rules about word limit, organisation of the m anuscript, m argins, line spacing, preparation of tables and figures and the m ethod used to cite references. Failure to comply with the guidelines may result in rejection or return of the m anuscript for correction, thereby delaying the process of review and publication. Writing the manuscript The art of writing a m anuscript improves with practice and considerable help may be gained by asking others, especially those who have published, to critique and proof­ read drafts. This also provides a m eans of a second check of accuracy and internal consistency. Getting started is often the m ost difficult part and for this reason it is best to begin with the easiest sections. These are usually the m eth­ ods and results, followed by the discussion, conclusion, introduction, references and title, leaving the abstract until last. If possible, try and set aside some time for writing on consecutive days. Long gaps between periods of writing interrupts the continuity of thought. To avoid frustration, ensure all the necessary inform ation, for exam ple all data, references and any drafts of tables or figures, are at hand before starting to write. The task of writing the manuscript may seem easier if each section is viewed as a separate task. Before starting to write, it may help to prepare an outline for each section which includes a num ber of major head- Sue Jenkins GradDipPhys, PhD, MAPA is a Senior Lecturer in the School of Physiotherapy*, Curtin University of Technology, Perth, W estern Aus­ tralia. Correspondence; Dr Sue Jenkins, School of Physiotherapy, Curtin Uni­ versity of Technology, Selby Street, Shenton Park WA 6008. The SA Journal of Physiotherapy gratefully acknowledges the Australian Journal of Physiotherapy for permission to publish this article. ings, su b-headings and paragrap hs covering different points. W hen w riting the first draft, the goal is to get som ething down on paper, so it does not m atter if sentences are incom plete and the gram m ar incorrect, provided that the main points and ideas have been captured on paper. Try to write quickly, to keep the flow going. U se abbrevia­ tions and leave space for w ords that do not com e to mind im mediately. H aving finished the first draft, im m ediately revise it and be prepared to do this several tim es until you feel it is not possible to im prove it further. A cceptance of a m anuscript is invariably conditional on changes being made so be prepared to rewrite and revise the m anuscript extensively. Often a m anuscript has m ore than one author and thus the writing may be shared. H ow ever, the style needs to be consistent throughout so even if sections of the early drafts are written by different authors, the first author m ust go through the entire m anuscript before subm itting, and make any necessary editorial changes. Structure and content of a manuscript A m anuscript is typically com posed of a num ber of sections: abstract; key words; introduction; m ethods; re­ sults; discussion; conclusions and references. In order to maintain continuity betw een the key sections (introduc­ tion, m ethods, results and discussion) it is helpful to con­ sider the m anuscript as telling a story. The strong parts to the story-line are the introduction and the discussion so the link betw een these sections must be clear. The research question which is posed at the end of the introduction must be answered at the beginning of the discussion (Zeiger 1991). H aving invested m any hours in undertaking research, the tem ptation is to try to tell the reader everything you read and learned in the process and to provide all the data gathered. However, in the planning stages, it is essential to rem em ber that a word lim it is usually im posed and there­ fore unim portant or irrelevant inform ation m ust be left out. In the case of a large study, it may be necessary to write several papers which cover different research questions. Title This provides the first im pression to the reader, so select­ ing the m ost appropriate title requires som e thought. The title influences whether a reader is interested in reading the manuscript. It should include all essential words in the right order such that the topic of the m anuscript is accu­ rately and fully conveyed (e.g. clearly related to the pur­ pose of the study) (Rudestam and N ew ton 1992). Avoid long titles (the recom m ended length is 10-12 words) and those which begin with redundant w ords such as "A study of....". Abstract An abstract is a brief sum m ary (of specified w ord limit) of the content of the m anuscript. It should provide the SA J o u r n a l P h y s io th e r a p y , V o l 5 3 N o I F e b r u a r y 1997 P a g e 3 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 3. ) highlights from the introduction, m ethods, results, discus­ sion and conclusions (Table 1). It m ust make sense when read in isolation for those who read only the abstract. This is especially im portant given that m any com puterised searches only retrieve the abstract. The abstract m ust also provide a clear and accurate recapitulation of the m anu­ script for readers who read the entire m anuscript (Zeiger 1991). For exam ple, an abstract m ust not contain data which are not included in the results. The abstract is usually w ritten as one or two paragraphs and it is im portant that the text flows and does not resemble a collection of disjointed sentences. The choice of words should be sim ple, jargon avoided and abbreviations om it­ ted except for standard units of m easurem ent and statistical terms. Citations are not usually included. Excessive detail such as long lists of variables, large am ounts of data or an excessive num ber of probability (p ) values is not acceptable. The trick to producing a clear abstract is to provide just enough detail to dem onstrate that the design of the study was good and that the evidence for the answer to the question is strong. Key words M ost journals require the author to identify three or four key words which represent the m ajor concepts of the paper. These are used for indexing purposes and m ust be selected from the Index M edicus M edical Subject H eadings (MeSH). For exam ple, "p h y siotherap y " is not included in M esh; the equivalent term is "p h ysical therapy". In the rare event that an author does not have access to M eSh, the key words selected should be widely-accepted terms. Lack of access to M eSh should be indicated at the time of m anuscript sub­ mission. Introduction T he pu rpose of the introduction is to stim ulate the reader's interest and to provide background inform ation which is pertinent to the study. The statem ent of the re­ search question is the m ost im portant part of the introduc­ tion. The review of the literature needs to be short and concise. The content of the introduction is outlined in Table 2. References are alm ost exclusively used in the introduc­ tion and the discussion. The references cited should be those which are the m ost valid and the m ost available. Articles in peer-review ed journals satisfy both these crite­ ria. Books, M aster's and PhD theses and some conference proceedings, those for which papers are rigorously re­ viewed, are also valid sources but usually take longer to find. Abstracts do not contain enough inform ation to allow critical evaluation of the work. Journal articles which have been accepted for publication are a valid source but those which have been subm itted (but not yet accepted) are not as they are unavailable. A void citing personal com m unica­ tions and unpublished reports or observations. These are not strong evidence because they are unable to be accessed and evaluated. The num ber of references should be limited to the few est num ber necessary by choosing the m ost im ­ portant, the m ost valid and, where appropriate, the most recent (Zeiger 1991). B la d s y 4 F e b r u a r ie 1997 • S tate m en t of: the q uestion a sked (p re sen t verb ten se) w ha t w as done to a n s w e r the q u e stio n (p ast verb te n s e ) - research design, popula tio n stu d ied , in d e p e n d e n t and d e p e n d e n t variab les • Findings th a t a n s w e r the q ue stio n (past ve rb te n s e ) - the m ost im p orta nt results and e vid e n c e (d ata p re s e n te d in log ical orde r) • T he a n s w e r to the q ue stio n (p re sen t verb te n s e ) If useful, and where word lim it allows, include: • O ne o r tw o s en ten ce s of b ack g ro u n d info rm a tion (placed at the b eginning) • An im p lica tion o r a sp e cu la tio n based on th e a n s w e r (p re sen t verb ten se), placed at the end Table 1: Abstract Table 2: Introduction • B ackgrou nd to the top ic (p ast verb ten se) w h a t is know n o r b elieved a b o u t fhe topic w ha t is still unknow n o r p ro b le m a tic fin d in g s o f relevant s tu d ie s (p ast verb te n s e ) • S tate m en t of the re sea rch q ue stio n sev e ra l w ays can be used to s ig n a l th e re sea rch q ue stio n, e.g. “T o d eterm ine w h e th e r . . . ” “T h e purpose o f the stu d y w as to . . . ” T h e stu d y tested the h y p o th e s is t h a t . ..” T h e stu d y w as u nd erta ken t o . .. ” • A pp ro ach taken to a n s w e r the q u e stio n (p ast verb te n se). Methods This section is descriptive. The m ain consideration is to ensure that enough detail is provided to verify the findings and to enable replication of the study by an appropriately trained person. Inform ation should be presented, using the past verb tense, in ch ro n o lo g ical order. Su b-headin gs should be used, w here appropriate. Reference m ay be m ade to a published paper as an alternative to describing a lengthy procedure. M any journals require m ention of relevant ethics com m ittee(s) approval for the study and that subjects gave inform ed consent. Table 3 provides an outline for the content of the m ethods section. Results The two functions of this section are to report the results (past verb tense)of the procedures described in the m ethods and to present the evidence, that is the data (in the form of text, tables or figures), that supports the results. Som e journals com bine the results and discussion into one sec­ tion. Before sitting down to write the first draft, it is im portant to plan which results are im portant in answ ering the ques­ tion 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 hy­ p o th esises). A fter deciding which results to present, atten­ tion should turn to determ ining w hether data are best presented within the text or as tables or figures. Tables and figures (photographs, draw ings, graphs, flow diagrams) SA T y d s k rif F is io te r a p ie , D e e l 5 3 N o 7 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 3. ) • O utline o f the stu d y design • S ub je cts m ethod o f sam p ling and recruitm ent; n u m b er of sub jects and ju s tifica tio n o f sam p le size inclusion, exclusio n and w ith draw al criteria; m ethod o f a llo c a tio n to stu d y groups • V ariables independent, d ep en de nt, extra ne o us, controls • P ilot stu d ies o u tc o m e o f any p ilot stu d ies w hich led to m odification s to the main study • M aterials e qu ipm e nt, in s tru m e n ts o r m e a su re m e n t tools- (include m odel n u m b er and m a nu fa cture r) • P rocedures detaile d d e scriptio n, in chro n o lo g ica l order, of e xa ctly w ha t w as done and by w hom • M a jor e thical co n sid e ra tio n s • D ata re d u c tio n /s ta tis tic a l analyses m ethod of ca lc u la tin g derived variab les, dea ling w ith outlying values and m issing data m e th od s used to sum m arise d ata (p re sen t verb tense) sta tistical s oftw a re (nam e, versio n o r release num ber); statistical tests (cite a re feren ce for less c om m on ly used tests) and w hat was c o m p ared ; critic a l alpha prob a bility (p) valu e at w hich d iffe r­ enc e s /re la tio n s h ip s w ere con sid e red to be s ta tistically significant. Table 3: Methods 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 pow erful visual m eans of presenting data and should be used to com plem ent the text but at the sam e time m ust 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 w ithin the text. Sources of help for the preparation of tables and figures are Briscoe (1990), Price (in press) and Zeiger (1991). Tables and figures m ust be mentioned within the text and should be placed after the related text. P hotograp hs of subjects are often placed within the m ethods and should be used only if written, informed consent was obtained prior to the taking of the photograph. To preserve anonym ity, facial features should be covered. If a m anuscript includes a table or figure that has already been published, permission m ust be obtained from the copyright holder (usually the publisher) and the source acknowledged. Confusion som etim es 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 num bers, which may be presented in their raw form, sum m arised (e.g. means) or transformed (e.g. percentages, ratios) (Zeiger 1991). For exam ple, in a hypo­ thetical study com paring vital capacity (VC) in supine and standing, the results statem ent and data respectively might be, "vital capacity was decreased in supine compared to standing" and "m ean (SD) vital capacity was 2.95 litres (0.8) SA J o u r n a l P h y s io th e r a p y , V o l 5 3 N o 1 and 3.41 (0.6) (p 0.05) in supine and standing respectively". The two statem ents should be presented together with the results statem ent given first, i.e. "v ital capacity was de­ creased in supine com pared to standing, m ean (SD) values were 2.95 litres (0.8) and 3.4 litres (0.6) respectively (p 0.05) in the 20 subjects". W hen presenting data, the sam ple size (n) m ust be given, any m issing data identified and the p values for data that have been analysed using statistical tests m ust be included. The significance level (critical alpha [a] probability value) should be reported but in the event that the null hypothesis (H0) is accepted the beta (b) prob­ ability value or statistical pow er should be reported. Stud­ ies in which H 0 is retained are ju st as im portant 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 exam ple t^ 2) = 3:12 should also be included unless only one statistical test has been used. W hen citing a p value, alw ays give som e idea of the m ag­ nitude of the difference (e.g. 20 per cent increase) as a p value in isolation gives no indication of the im portance of the finding. It is generally accepted w hen reporting results that "sig nifican t" or "sig nifican tly" refer to statistical sig­ nificance (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 w ith the m ethods or from the m ost to the least im portant. The order of m ost to least im portant should be follow ed w ithin each paragraph. For every result there m ust be a m ethod in the m ethods. Careful planning of the tables and figures is im portant to ensure that the sequencing of these tells a story. The results m ust not include a discussion of the findings, m ethods of data analyses and citations of references, except on rare occasions when a.com parison is m ade of the raw data with the findings of a published study. This applies only when this com parison would not fit well w ithin the discussion. Discussion The discussion should be considered as the heart of the paper and invariably requires several attem pts at writing (Portney and W atkins 1993). It serves to answ er the ques­ tio n ^) posed in the introduction, explain how the results support the answers and how the answ ers fit in w ith exist­ ing knowledge on the topic (Zeiger 1991). This is the m ain section in which the author can express h is /h e r interpreta­ tions and opinions, for exam ple how im portant the author thinks the results are, the author's suggestions for future research and the clinical im plications o f the findings (Port­ ney and W atkins 1993). In order to m ake the m essage 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 discuss­ ing other im portant and directly relevant issues. Side issues and unnecessary issues should not be included, as these tend to obscure the message. Care m ust be taken to provide a com m entary and not a reiteration of the results. The recom m ended content of the discussion is given in Table 4 (Zeiger 1991). C ontinued on p ag e 7 ... February 1997 P a g e 5 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 3. ) 350,450 & 950 Physiotherapists will find this family of Therasonic and Interferential Units ideally suited to their different requirements. TECHNICAL SPECIFICATIONS 350/450 EMS350 Power input Display Size Weight Classification Output Mode Treatment Time Ultrasound Frequency Output intensity Pulse width Output power Transducer Beam type Output Type Output Current Output compliance Carrier Frequency A M F Swing Patterns Programs 10 0 -2 4 0 V 50VA m ax 5 0/60 H z L.E.D. 1 0 0 x 2 4 0 x 2 1 0 m m 1,5 kg IEC 601 Parts 1 & 2.5 (B S 5 72 4 Parts 1 & 2.5) c w and pulsed 1:4 0 -2 0 m in u te s (d ig ita l) T reatm ent linked 1 M H z and 3 M Hz 0 -2 .0 W /c m 3 2 m s 0-10W 5 .0 c m 2 ERA. BI\IR<6 (FDA M e th o d ) C ollim a tin g 950 10 0 -2 4 0 V AC 50/60HZ L.E.D. 1 0 0 x 2 4 0 x 2 1 0 m m 1,5 kg Class 1 Type BF (IEC 6 0 1 .1 ) 2 o r 4 -P ole 3 0 m in u te s m a x im u m C on sta nt c u rre n t 130 m A peak in to 5 00 O hm s ±65V 2 kHz o r 4 kHz / - s 0 -2 5 0 Hz settable in 5 Hz ste p s ; ' 1 / 1 , 6 /6 6 A 6 \ : 8 P reset AM F fre q u e n c ie s EMS 950 Sole Distributors: MEDICAL DISTRIBUTORS (p t y ) l t d P.O. BOX 4019, HALFWAY HOUSE 1685 Phone (011)314-3000 P.O. BOX 195, CAPE TOWN 8000 Phone (021) 47-4440 P.O. BOX 5298, DURBAN 4000 Phone (031)37-1501 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 3. )