|REFLECTIONS ON RESEARCH
|Year : 2020 | Volume
| Issue : 6 | Page : 937-943
Pitfalls in article submissions for publication
Department of Dermatology, Military Hospital Jaipur, Jaipur Cantonment, Jaipur, Rajasthan, India
|Date of Submission||15-Aug-2020|
|Date of Decision||28-Aug-2020|
|Date of Acceptance||27-Oct-2020|
|Date of Web Publication||08-Nov-2020|
Department of of Dermatology, Military Hospital Jaipur Cantonment, Bani Park , Jaipur, Rajasthan
Source of Support: None, Conflict of Interest: None
| Abstract|| |
The aim of every academician and clinical dermatologist is to publish their research in reputed biomedical journals. But from conceptualization to completion, myriad shortcomings creep into the article and by the time it is ready for publication, by default and certainly not by design, the article discourse gets flawed, sometime fatally so. The endeavor of this article is to discuss these pitfalls from conceptualization, statistical machinations, authorial misconcepts, article structuring, and final journal selection. The article can function as a prophylactic checklist, albeit not comprehensive, by any prospective author and is an appreciation of the most oft repeated fallacies usually detected in publication submissions.
Keywords: Article, pitfalls, publication, submissions
|How to cite this article:|
Nair B. Pitfalls in article submissions for publication. Indian Dermatol Online J 2020;11:937-43
The aim of every academician is to convert their innovative and novel ideas of research into a publication. Publication in a reputed indexed journal is the dream of many a researcher. This helps in academic furtherance, further research collaboration with like-minded academicians and also career promotion. But unfortunately, dermatological articles which are submitted having an honest innovative concept is ‘lost in translation'. It is imperative that the viable novel research project get caught up due to certain common pitfalls in the conceptualization, methodology, translation into the final paper and deficiencies in the knowhow regarding publication dynamics. The aim of this article is to basically act as a checklist of sorts for any aspiring researcher. Certain general issues will be covered and specific section wise pitfalls will be highlighted in the subsequent sections.
Pitfalls in selection of research question
Novelty of the research project is the most important facet which the journal assesses. Not conducting an adequate literature review to assess the possible redundancy of the research question is a major deficiency. This is imperative to avoid the phenomenon of research waste. The following points are frowned upon by journals in general
- Rehash of established facts. For example, redox theory of vitiligo is a well appreciated pathogenetic concept. To prove the same again by using another oxidant marker to reinvent the wheel about the redox concept of vitiligo will be redundant.
- Irrelevant research topic selection
- Clinically non-translatable basic research and animal studies
- Selection of a non-generalizable research question
Pitfalls in conceptualization
The most important limiting step is the concept which fructifies into the research project. The following issues crop up in the conceptualization of the project
- The aims and objectives have to be crystal clear and these things cannot be evolved as the study progresses.
- Incorporating the biostatistician during conceptualization and estimation of sample size based on primary outcome measure depending on the primary outcome measure and anticipated effect size.
- Taking on projects without critical evaluation of feasibility and availability of center resources is a common pitfall and needs to be addressed ab initio.
- Not designating a primary outcome measure and ignoring the fact that anticipated sample size is dependent on the primary outcome measure is another common deficiency. Hence primary and the secondary outcome measures have to be clearly delineated
- Non-registration of interventional trials in the Central Trial Registry of India once the protocol is approved by the institutional ethics committee before actual initiation of subject recruitment is a major pitfall which leads to article rejection. The locking of your aims, objectives, methods, and outcome measures in the form of a registered protocol prevents mid-stream switch of research prerogative. Reputed journals are recommending even systematic reviews to be registered with PROSPERO registry to enhance authenticity
- Institutional Ethical Committee approval is mandatory for kinds of studies. Submission of the ethics committee approval document along with the article or on reviewer/editorial request during article processing is desirable.
| Pitfalls in Title|| |
An article title is extremely important facet which initiates the readership process. The following pitfalls may be avoided
- Title not reflective of the nature of study
- Non mention of study design in title. This is important for purpose of systematic review and meta-analysis. If the search parameters is for randomized control trials it will be easier for the systematic reviewers to pick up your research and use the data of your study in the meta-analysis if attempted. This increases the relevance of your data.
- Over-flowery grandiose titles can actually be detrimental and may not reflect what the article contains
- More than 16 words in title feels lengthy and may be undesirable
- The name of the nature of intervention and disability need to be definitely included.
- Whether to reflect the study outcome in the title or be outcome neutral in the title is author's choice. But actually, reflecting clearly the outcome of research may be a good idea but it is not mandatory
- Use of abbreviations, formulas and brand names in title is absolutely not recommended
| Pitfalls in Abstract|| |
The abstract is the bird's eye view that is the most publicly visible and initially evaluated facet of any article. Hence, a well-written abstract is absolutely imperative for consideration by editors.
- CONSORT statement for Abstracts is an essential requirement for interventional trials. Not structuring the abstract is a common pitfall. Even observational analytical papers can structure their abstract based on CONSORT
- The abstracts must be self-explanatory and must be a stand-alone viable synopsis of the article.
- Non reflection of confidence intervals along with P values is a common pitfall. P values alone carry no meaning without estimating confidence intervals.
- Abstracts need to be structured under these heads—Objective with one line on need of research, Design and Setting, Subjects, Measurements (What is being measured, what statistical tests were used, what outcome measures were evaluated), Results, Limitations, Conclusions, and Keywords.
| Pitfalls in Introduction|| |
The introduction is actually a run up into the actual research section of the article. A good introduction needs to be succinct, clear and must encourage the reviewers to approach the actual methods section with a feeling of anticipation and must convince them that this line of research was actually needed and the research question has been chosen with due diligence. Common pitfalls are as below
- Including long passages on nature of disease, clinical presentation of disease, and pathogenesis in an interventional disease. The description of disease must be completed in 1-2 sentences.
- Confusing introduction with discussion
- Not discussing study hypothesis clearly or nebulous hypothesis
- Not reflecting the aims and objectives of research in introduction
- The introduction is best structured like this
- Significance of concept
- Magnitude and significance of the knowledge gap
- Discrepancy between previous studies in literature necessitating the research project (differences in results, conclusions, opinions)
- How does the study set out to clear the knowledge gap?
- Delineation of the research question
- Purpose statement (aims, objectives)
| Pitfalls in Study Design|| |
This is the most important aspect of any paper publication. And hence the generic comments and subsections shall be dwelled into in some length.
- A common pitfall is the wrong naming of design. Calling a cross sectional study as a case-control study is a common observed pitfall. Another common pitfall is calling an interventional study as a case control study. Any kind of intervention whether therapeutic, counselling, etc., constitutes an interventional trial and not a descriptive analytic study (case-control study). Basic awareness of research designs can help avoid this common pitfall.
- Choosing the correct research design is all important. The appropriate design for the appropriate research question is reflected in [Table 1]
- Vague/inadequate method description
- Methods lacking sufficient rigor
- Biased (stacked) protocol—This happens when a newer intervention at adequate dosage with another older intervention at lower than standard dosage (ineffective dosage) and thereby implying greater efficacy for the newer intervention is highly undesirable.
- Small sample size/sample size calculation not done—This is the commonest pitfall in many research submissions.
- The way to circumvent these pitfalls is to discuss with peers, reading of literature, interaction with experienced people in the said line of research and intervention of biostatistician ab initio.
| Pitfalls in Materials and Methods|| |
- Inadequate description of intervention: Not mentioning dose, route of administration, frequency of dosage, duration of intervention, and duration of follow up
- Inadequate mention of study period
- Non clarity regarding study settings
- Non-mention of sampling frame for cases and controls
- How were the cases and controls defined—Clear inclusion and exclusion criteria
- Not mentioning wash out period prior to intervention
- Not confirming informed consent process for patients
- Not describing standard of care if that is what is offered for control
- Inadequate mention of co-interventions in case and control group
- Not mentioning allocation concealment
- Using quasi-randomization methods like alternate recruitments into case and control groups and claiming adequate randomization
- No patient flow chart provided
- Quantum of patients considered for recruitment but not randomized. This needs explicit mention as to why the subjects considered for recruitment were not ultimately randomized. This is an often-forgotten facet of materials/methods and may reflect the acceptability of the intervention by the patients and their apprehensions regarding the interventions offered.
- Reasons for dropouts and crossovers—This needs to be clearly mentioned, preferably in the flowchart
- Who all were blinded—This decreases risk of bias. If the interventions are unlike each other, blinding of investigators or patients may not be feasible. In such scenarios too, outcome assessor must be blinded. If the outcome measured is subjective, it will be important to have multiple outcome assessors and their agreement coefficient must also be assessed
- In case of complex interventions, it is better to include innovative figures to describe the timing of intervention, interventional frequency and follow up
- Using new non-validated scales, scores, instruments, and questionnaires. Adequate references must be provided for previous validation of the scales and instruments used
- Lack of reproducible details. The materials and methods must be provided in exquisite detail without lacunae to facilitate any future researchers or readers of the article to recreate the same processes in case they want to extend the line of research or reproduce the same research with a larger sample size in the future
- Lack of a control group in self-healing conditions like alopecia areata, warts, etc. This is a serious flaw and a control group to factor in the spontaneous healing is mandatory for any viable valid outcomes.
- Not mentioning confounding variables and measures to reduce confounding.
- Mention of institutional ethics committee approval is very important in the manuscript. It is ideal to submit a copy of IEC approval along with manuscript in additional files section
- Any instrument used to assess outcomes must be ideally validated and validation references need to be cited (i.e., MASI, PASI). De novo instruments without adequate validation to measure outcome brings the results into question
- It is ideal to include both objective physician assessment of outcome along with patient generated outcome measures (e.g., DLQI), which improves the patient-centricity of the research endeavor. Ultimately all biomedical research is not for academic furtherance, but for patient benefit and this needs to be remembered at every turn.
- Factoring for confounding variables in observational analytic studies and adjustment for confounding variables in both methods and statistical planning is needed. There is always a risk of confounding as randomization has not been carried out, i.e., some other covariate is associated with both the risk factor or intervention and the outcome. If you know what the covariate is, you can “control” or “adjust” for it using regression analysis.
- Propensity Score (PS) matched analysis is a tool to control for confounding and selection bias and must be resorted to at every probable instance. In the statistical analysis of observational data, propensity score matching (PSM) is a statistical matching technique that attempts to estimate the effect of a treatment, policy, or other intervention by accounting for the covariates that predict receiving the treatment. PSM attempts to reduce the bias due to confounding variables that could be found in an estimate of the treatment effect obtained from simply comparing outcomes among units that received the treatment versus those that did not. It allows us to take many different variables and condense them into a single variable that gives the probability that the intervention will occur in each individual. Thus we can find people or groups with similar propensity scores, and use this score to make the groups being compared more similar. By matching people with similar scores we can see if the results still show that the intervention is associated with the outcome of interest. In this way, PS looks like an attractive tool to simulate a RCT in which all the covariates end up being the same except the treatment. Thus, many biases of an observational study can be obviated by propensity score matching
- Proliferation of submissions involving correlation of inflammatory markers and oxidative markers in dermatological studies. Of late, there has been a proliferation of observational (mostly, retrospective) analytic studies involving such non-specific parameters as Red Cell Width (RDW), Neutrophil: Lymphocyte Ratio (NLR), C-Reactive Protein and myriad markers of oxidant-antioxidant status in various dermatological conditions with the attendant allusions of diagnostic status for positive results gleaned from such studies. This can be patently mis-leading.
- Labelling a single group observational descriptive study as cohort study or cross-sectional study is a common pitfall. A cohort study or cross-sectional study needs to have a control group.
- Labeling a cross-sectional study (comparing two groups at a fixed snapshot in time) as a case-control study just because it has a case and control group is a common avoidable malady.
- Selection bias needs to be avoided in observational studies. Selection bias arises when the study population is not a random selection from the target population which is subject of the observational study. Individuals are then recruited in such a way that they are not representative of the target population. The aim must be a high participation rate, in order to achieve a representative cross-section of the population if possible. Self-selection of participants also takes place when economic, linguistic or health barriers hinder participation. Cultural differences and social status can also influence readiness to participate, for example, in screening programmes. This all tends to reduce the possibility of generalizing the results.
- Measurement error or misclassification may result from lack of care by the investigator or from poor quality of measuring or survey instruments. Collecting blood samples for cases in morning and controls in evening, for example, will bias results. An interviewer treats patients during an interview with more sympathy than the controls, as their status rapidly becomes clear to him during the interview. As a result, he obtains more, and more detailed, information from the patients. These tendencies needs to be eschewed.
| Pitfalls in Results|| |
Results are where you lay out the fruits of your research for reader perusal and has to be simply presented without ostentatious wording in the most ergonomic way possible with judicious combination of text and figures. Common pitfalls in results section noted are as follows:
- Baseline characteristics table clearly lays out the data set for benefit of the interpreter and must always be included.
- Combining results with baseline characteristics in one table is to be strictly avoided.
- Results are where the results of your research endeavor are blandly laid out along with accompanied statistical analysis. Interpreting the results too much in the results section is not recommended.
- Non-use of charts, tables and graphs makes interpretation of results tedious and must be used judiciously. Overuse of graphs and charts also can be counter-productive.
- Repeating what you have said in text in form of tables also is overstating the obvious and needs to be eschewed. Tables must be used to judiciously reduce word count.
- Selective omission of results to match your hypothesis is not justifiable. Sometimes studies omit to mention primary outcome results altogether or try to gloss over the primary outcome in favor of secondary outcome variables and certain post-hoc added variables which have attained the miraculous P value of <0.05. This is deemed avoidable at all costs.
- Presenting results in the same logical order presented in methods makes sense and creates a smooth logical ergonomization of ideas.
- Reduce usages of turn of phrase like “Almost significant. Tends to significance. Nearly significant” to put a spin on results to favor a vested hypothesis. It easily catches a reviewer eye.
- P values not accompanied by confidence intervals in the tables and text are an incomplete presentation of results. Presenting the confidence intervals helps reviewers and readers interpret the data in its totality and can also glean the clinical significance of the statistically significant results.
- Do not include every statistical table generated by biostatistician as is. An article is not a thesis.
- Do not embed tables and graphs in article text. Graphs and charts are submitted in “figures” section and Tables are included after references. Closely follow author instructions of the journal to which you intend to submit your article.
- Another not so uncommon pitfall is that the results in main text and tables fails to match numerical values in abstract of the article. This is a glaring deficiency which raises doubts about the veracity of the results.
- Inclusion of a patient flow chart both pre-randomization and post randomization in an RCT is must. Clear reasons must be stated for drop-outs in study and control arms both pre-randomization and post randomization.
- Presentation of a table depicting results of both univariate and multivariate analysis is important in observational studies.
- Interpreting correlation as causation in observational studies is a common pitfall. Correlation (Pearson's and Spearman's correlation coefficient) with significant values does not indicate causation but only hints at an association.
- Submitting a STROBE checklist along with observational studies will enhance the chances of appreciation from evaluators. STROBE checklist is for observational studies akin to CONSORT checklist for RCTs.
- Missing data has to factored in analysis of results and chances of data missing in observational analysis are much more likely than prospective interventional trials. Retrospective cohort studies and case control studies are more prone to this pitfall as data are retrieved from databases or based on individual memory and recall and are thus prone to bias and can result in chunks of data missing.
| Common Pitfalls in Statistics|| |
The section on statistics can be an article in itself. Hence for detailed discussion, other articles may be referred to. Only common pitfalls are discussed below.
- Not having a statistical plan. Certain articles are submitted wherein authors interpret that treatment A cured 50 patients in Group A and treatment B cured 43 patients in Group B, and hence Drug A is better than Drug B without any formal statistical analysis. This is a cardinal mistake.
- The concept that there is no clinically significant difference if P value magic mark is not attained. It must be remembered that absence of evidence is not evidence of absence.
- Using convoluted statistics with new-fangled statistical contortions in alliance with a permissive bio-statistician to put a positive spin on a negative result is best avoided.
- Non pre-specified (not mentioned in initial study protocol) post hoc data dredging is a red flag for many reviewers and editors.
- Not doing any tests for determining the normality of data distribution is a common fallacy.
- Using parametric tests for non-parametric or small data sets is another pitfall in analysis.
- Multiplicity of analysis without a correction can lead to one comparison which delivers false positive significance. Multiplicity of analysis involves comparing results at multiple time points, comparing results in multiple groups and comparing multiple parameters between study groups. Correction is necessary to avoid multiplicity induced spurious conclusions based on significance achieved by happenstance.
- When there are multiple confounding variables in an observational analytical study, the same has to be factored into the analysis. At times univariate analysis leads to significant results, but once multivariate analysis factoring in the confounding variables is done, results prove to be non-significant. Adequate adjustment at inclusion/exclusion criteria level or statistical analytical level is a must to account for confounding of the results.
- Incorporation of a biostatistician to evolve a statistical plan ab initio at the protocol along with sample size calculation based on outcome variables is essential. Grounds up statistical inputs will help us flesh out a cogent protocol which is the most important stage of any research.
- Uncritical acceptance of statistical outcomes without introspection of clinical value is a pitfall best avoided.
- Both Intention to treat (ITT) and per protocol analysis may be done ideally. Not doing ITT analysis where there are many patient dropouts in either study arms is fallacious.
| Pitfalls in Discussion (Pitfalls in Interpretation of Results)|| |
Discussion involves an interpretation of results derived from the study. Here a comprehensive literature review is a must.
- Repeating results again in toto in discussion section is a common error which is counterproductive and unnecessarily increases word count.
- Springing results not stated in “results section” on the unsuspecting reviewer/reader in discussion is also to be avoided.
- Discussion is about interpreting your results in light of pre-existing literature involving similar research efforts. Tabulation of results of comparable studies helps to contextualize your results.
- Discussion of previous literature while omitting to contrast with your results is a common pitfall made by many authors.
- Including own previous studies and articles irrelevant to the present research effort to increase self-citation is a practice frowned upon by many editors.
- Forgetting to reassess hypothesis in light of results obtained is a common omission.
- Restricting discussion on only variables which deliver significant results at the cost of variables not achieving statistical significance is a common pitfall.
- Limitations. Often authors are reluctant to discuss limitations of the study. Limitations must include inadequacy of sample size, deficiencies in methods detected mid-stream by the authors (which can be avoided by other researchers doing research on similar lines in future), confounding variables omitted erroneously which were detected later, any other issues detected by authors which can affect both internal validity and generalisabililty of results. It is also an appreciable endeavor to suggest the ways to eliminate such pitfalls in any future research endeavors in that subject. A well-fleshed-out limitation section is a boon for any future researchers.
- Selective inclusion of studies that support your hypothesis and omitting contrary results from some studies to bolster your hypothesis will be detected at some stage and must be strictly avoided.
- Inflation of the importance of one's findings in a study is easily detected and discouraged by reviewers and reflects intellectual bias.
- Conclusion. Conclusion must reflect what the results of the study are and cannot include things which have not been proven or disproven by the study in question. It is better to overreach on conclusion. Future research directions can be suggested in conclusions.
- The popular refrain “Further larger studies are required in this area” without a proposed route map for further research on similar lines is a cliché best avoided.
| Miscellaneous Pitfalls|| |
Certain other seemingly minor pitfalls can cause much heartache due to article rejections. Same are mentioned in [Table 1]. The deadly sins which definitely end in rejection are also mentioned in [Table 2].
|Table 2: The cardinal sins of article submission (These result in rejections and might eventuate in banning of authors by particular journals from submitting future research articles)|
Click here to view
| Pitfalls in Journal Selection|| |
Selection of the right journal after the research project/thesis on completion is very important. The aim of every researcher is to submit in an internationally reputed high-impact journal but it may not always be feasible. Certain appreciated pitfalls in journal selection is as follows.
- Not exploring into previous issues of journal to understand the ethos and the areas of interest of a journal is a common pitfall.
- Not reading author instructions in detail of the candidate journal is a common mistake.
- Not scanning previous issues for the journal for similar articles is a common pitfall. This particularly applies to case reports/case series. If a similar case has been published in the journal in recent times, it would be better to avoid submitting your similar case however rare it might be to that journal.
- Submitting basic research to a clinical journal without any translational message and vice versa is sure shot recipe for rejection.
- Understanding the nature of the readership prior to submission is essential.
| Conclusion|| |
The aim of this article is not to cover each and every possible pitfalls in article submission, but to evolve a kind of checklist which includes the most common reasons for rejection, which can be avoided by following this list. The author expects this article to provide a route map to both inexperienced and seasoned researchers to guide their research effort into tangible publications in respected journals.
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Katz, M. J. 2006. From Research to Manuscript. A Guide to Scientific Writing. Dordrecht, The Netherlands: Springer, pp. 152.
Grindstaff TL, Saliba SA. Avoiding manuscript mistakes. Int J Sports Phy Therapy 2012;7:524-30.
Senturia SD. How to avoid the reviewer's axe: One editor's view. IEEE Trans Ultrason Ferroelectr Freq Control
Yan F, Robert M, Li Y. Statistical methods and common problems in medical or biomedical science research. Int J Physiol Pathophysiol Pharmacol
Porcino A, Moraska A. Avoiding Common Writing Mistakes That Make Your Editors and Reviewers Cringe. Int J Ther Massage Bodywork
Azer SA, Dupras DM, Azer S. Writing for publication in medical education in high impact journals. Eur Rev Med Pharmacol Sci
Shakiba B, Salmasian H, Yousefi-Nooraie R, Rohanizadegan M. Factors influencing editors’ decision on acceptance or rejection of manuscripts: The authors’ perspective. Arch Iran Med
Turcotte C, Drolet P, Girard M. Study design, originality, and overall consistency influence acceptance or rejection of manuscripts submitted to the journal. Can J Anaesth
Lee KP, Boyd EA, Holroyd-Leduc JM, Bacchetti P, Bero LA. Predictors of publication: Characteristics of submitted manuscripts associated with acceptance at major biomedical journals. Med J Aust
Radford DR, Smillie L, Wilson RF, Grace AM. The criteria used by editors of scientific dental journals in the assessment of manuscripts submitted for publication. BrDent J
Clark GT, Mulligan R. Fifteen common mistakes encountered in clinical research. J Prosthodont Res
Knight KL, Ingersoll CD. Structure of a scholarly manuscript: 66 tips for what goes where. J Athl Train.
Button, K.S., Ioannidis, J.P.A., Mokrysz, C., Nosek, B.A., Flint, J., Robinson, E. S.J., et al
. Powerfailure: Why small sample size undermines the reliability of neuroscience. Nat. Rev. Neurosci 2013;
Cumming, G. The new statistics: Why and how. Psychol. Sci. 2014;
Kirk, R.E. Practical significance: A concept whose time has come. 1996; Educ. Psychol. Meas.
Altman DG, Schulz KF, Moher D, Egger M, Davidoff F, Elbourne D, et al.
CONSORT GROUP (Consolidated Standards of Reporting Trials). The revised CONSORT statement for reporting randomized trials: explanation and elaboration. Ann Intern Med. 2001 ;134:663-94.
Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM, et al.
Standards for Reporting of Diagnostic Accuracy. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. BMJ. 2003;326:41-4.
Strange K. Authorship: Why not just toss a coin? Am J Physiol Cell Physiol.
Fontanarosa PB, Flanagin A, DeAngelis CD. Reporting conflicts of interest, financial aspects of research, and role of sponsors in funded studies. Journal of American Medical Association 2005;294:110-111.
Leppink J, O'Sullivan P, Winston K. The bridge between design and analysis. Perspect Med Educ. 2017;6:265-269.
Leppink J, Kok EM, Bergman EM, Van Loon MH, De Bruin ABH. Four common pitfalls of quantitative analysis in experimental research. Acad Med. 2016;91:891.
Leppink J, Kulasegaram KM. Avoiding common data analysis pitfalls in health professions education research. Acad Med. 2016;91:11.
Pierson DJ. The Top 10 Reasons Why Manuscripts Are Not Accepted for Publication. Respir Care 2004;1246-52.
Kallet RH. How to write the methods section of a research paper. Respir Care 2004;49:1229-1232.
Bordage G. Reasons reviewers reject and accept manuscripts: The strengths and weaknesses in medical education reports. Acad Med 2001;76:889-896.
Sullivan EJ. Top 10 reasons a manuscript is rejected (editorial). J Prof Nurs 2002;18:1-2.
Ehara S, Takahashi K. Reasons for Rejection of Manuscripts Submitted to AJR by International Authors. AJR 2007;188: W113-W116.
Bordage G. Reasons review ers reject and accept manuscripts: The strengths and weaknesses in medical education reports. Acad. Med. 76:889-896.
Branson RD. Anatomy of a research paper. Respir Care 2004;49:1222-1228.
Rosenbaum, Paul R.; Rubin, Donald B. The Central Role of the Propensity Score in Observational Studies for Causal Effects”. Biometrika 1983;70
[Table 1], [Table 2]