3.2.1. Study Design
The study design section of a scientific paper is the road map of the study method, which leads to a clear understanding of the data obtaining approach and helps the reader to interpret the results properly (
37). The study design should be the first subsection of the methods in a hypothesis-testing paper (
37). It provides an overview of the procedures used to answer the question(s) and is followed by the relevant details in separate subsections (
14). For hypothesis-testing papers, study question(s), intervention(s), variables measured, and the order of the measurements should be explained (
14). Furthermore, it is expected that this section covers the information including dependent and independent variables, controls (e.g., baseline, control group, and placebo), study duration, and sample size (
14).
The authors should present the specific design of the study, for example, randomized controlled trial, prospective/retrospective cohort study, case-control study, cross-sectional survey, and experimental study, or describe its key components (interventional vs. observational study, longitudinal vs. cross-sectional design) (
8). An overview on observational and interventional study designs can be found elsewhere (
38,
39).
For observational studies, study location and relevant dates (i.e., period of recruitment, period of exposure, follow-up, and data collection) should be described (
20). An extension of the STROBE statement (
Table 2) suggests more details for the study design section in molecular epidemiologic studies (
22); these details describe the specific study designs (nested case-control and case/cohort) (
40) and the setting of the biological sample collection (amount of sample, nature of sample collection procedures, participant conditions, time between sample collection and relevant clinical or physiological endpoints), biological sample storage and processing until biomarker analysis (centrifugation, timing, and additives), and biomarker biochemical characteristics (half-life of the biomarker and chemical and physical characteristics).
For human clinical studies, the authors are requested to specify the trial design (e.g., parallel and factorial), phase of clinical trial (phase I, II, III, or IV), and the allocation ratio (ratio of intended numbers of participants in each of the comparison groups) (
41). More information regarding common terms and designs of clinical trials are provided as useful links in
Table 2.
A further subheading entitled procedures or interventions may also be considered for clinical trials. In this section, authors need to provide detailed information for randomization procedures, including the method used to generate the random allocation sequence (computer-generated random numbers) and mechanisms used to implement the random allocation sequence (sequentially numbered containers), stratification, and random block sizes (if applicable) (
41). According to the CONSORT statement (
Table 2), it also should be described who generated the random allocation sequence, who enrolled participants, and who randomly assigned participants to interventions (
41).
If applicable, the authors should state which type of blinding was used (single or double) and who was blinded (participants, care providers, or data analyzer) (
41). Details of interventions, including how and when the interventions were implemented for each group should be specified. Information about the assessment of compliance and adverse events throughout the study should be included (
41). When applicable, it is expected that any interim analysis and cessation of the trial be clarified (
41).
According to ARRIVE (Animal in Research: Reporting In Vivo Experiments) statement (
Table 2), for animal studies, the number of groups, randomization procedure, blinding, and experimental unit (i.e., single animal, group, or cage of animal) should be mentioned (
19); for complex designs, a time-line diagram or flowchart can be useful (
19).
For genetic studies, the authors need to consider nomenclatures of genes and variants (
Table 2) and follow recommendations for the description of sequence variants (
42). For genetic association studies, an extension of the STROBE statement, namely STREGA, advises authors on how to provide further details in the study design section; details on the criteria and methods for the selection of subsets of participants from a larger study should also be described in this section. Furthermore, genetic exposures (genetic variants) and variables associated with population stratification should be clarified (
43).
3.2.2. Methods of Measurements/Assessments
Although describing details in the M&M section depends on the type of study and the target audience, authors need to maintain a balance. As a rule of thumb, the details of the procedures should be included if the study replication would fail without them. All that reader needs to understand is how the key findings in this paper were derived. However, this section should not be like a procedure manual or a cookbook (
4).
The term “condensed” or “extended” has been used to describe levels of details used in the methods section (
44). In the condensed methods, little elaboration or justification is provided, whereas in the extended methods, authors need to provide a rationale of why and how the procedures were performed (
44). In practice, depending on the novelty of the methods used in the study, different levels of details may need to be described (
Table 3). To summarize documented methods, authors may begin with “in brief”; use of “briefly” instead is a common mistake because “briefly” describes the following verb and does not indicate the author’s intention to be brief (
16).
| Method | How to Report |
|---|
| Familiar for everyone in the field | Not to be mentioned |
| Well-established methods, protocols, standards or previously published methods | Should be described in brief with appropriate citation |
| Relatively uncommon methods | Should be described in sufficient details with reference to original description and specific modifications made |
| Newly developed method | Should be described in more details including all reagents, conditions, and equipments |
The rationale for method choices and characteristics of the study design may also be provided in the methods section (
10,
11). From an editor’s point of view, advantages and disadvantages, values and limitations of the techniques and methods, especially new ones, are better to be described using a general background of the field (
45).
In this section, the authors need to clearly describe how study variables (i.e., exposures or independent variables, outcomes or dependent variables, covariates, or potential modifiers) were measured (
8,
15). If applicable, diagnostic criteria need to be clarified for the variables (i.e., exposure, outcome and/or confounder); moreover, sources of data and details of methods of assessments (measurements) should be described for each variable of interest.
In animal studies, details of how, when (time of day), where (home cage and laboratory), and why (rationale for dose and route of administration) for each procedure should be reported (
19).
According to minimum information for publication of quantitative real-time PCR experiments (MIQE), details about sample processing and storage, RNA and DNA extraction and quantification, primer and probe characteristics, reverse transcription details, sample normalization, PCR efficiency, and data analysis should be provided in real-time quantitative PCR (qPCR) experiments (
46).
For genetic association studies, authors need to describe laboratory methods, including source and storage of DNA, genotyping methods and platforms (including the allele calling algorithm used and its version), and error and call rates. The name of the laboratory or center where genotyping was performed and comparability of laboratory methods (if there is more than one group) needs to be clarified. According to the STREGA statement, authors should specify whether genotypes were assigned using all the data from the study simultaneously or separately in smaller batches (
43).
To describe instruments, the manufacturer and model as well as the calibration procedures should be described; in addition, it should be clearly described how measurements were taken (
10,
15). Details of measurement characteristics (i.e., reproducibility, validity, and responsiveness) that influence the interpretation of the main results should also be described (
8); validity and reliability, key indicators of the quality of measurement instruments (e.g., equipment and questionnaires) used for data collection or measurement should be appropriately reported (
18).
3.2.3. Statistical Analysis
The basic requirement of writing the statistical section is providing description and justification for the statistical approaches and selection of statistical tests (
14). General considerations for preliminary, primary, and supplementary analyses derived from statistical reporting guidelines (
47,
48) and the common pitfalls (
49,
50) in writing the statistical section are provided in
Box 1. The Vancouver guideline states “describe statistical methods with enough details to enable a knowledgeable reader with access to the original data to verify the reported results” (
51).
| Items |
|---|
| Useful tips |
| Describe preliminary analyses |
| Identify statistical procedures used to modify raw data or calculate new variables (transformation of data to close to normality, calculation of ratios, calculation of derived variables, categorization of variables) |
| Specify primary analyses |
| Identify included variables in the analysis (dependent variables, independent variables, and potential confounders) |
| Make clear which method was used for analysis (e.g., sample t-test was used to compare the means) |
| Verify that data conforms to the assumptions of the test (e.g., use of non-parametric tests for skewed data, and use of linear regression for linear associations) |
| Describe adjustments were made for multiple comparisons |
| Indicate which approach was used for treating outliers |
| Identify whether test was one- or two-tailed |
| Define within- or between-subject factors |
| Define the statistical significance level (e.g., 0.05) |
| Describe supplementary analyses |
| Describe methods used for ancillary analyses (e.g., sensitivity analysis, imputation of missing data, or testing the assumptions for methods) |
| Describe post-hoc analysis, unplanned subgroup analysis, or exploratory analysis |
| Describe the methods used to determine statistical power (in case of reporting null or negative results) |
| Common pitfalls |
| Inadequate description of methods and analysis |
| Inadequate specification for statistical methods |
| Lack of clarification for categorizing continuous variables |
| Failure to use correct names of statistical methods |
| Lack of appropriate citation or clear explanation for unusual statistical methods |
| Failure to address missing data |
Statistical tests should be discussed in order to be applicable for data analysis (
52). Typically, this section is initiated by preliminary analysis and descriptive statistics, describing the study population, and then it is followed by specific tests describing the association of variables or assessing the effect of experiments (
52).
The exact value of sample size, e.g., the number of human subjects, animals, or cells for each analysis and how the data were presented (mean, median, standard deviation, standard error, or confidence intervals) should be specified. Furthermore, the statistical methods used to determine strategies for randomization/stratification and sample size estimation need to be clarified (
14). Appropriate identification (i.e., name, version, company, city, state, and country) for the statistical package or program used for analysis must be mentioned.