Study manuscript formatting

Many research journals necessitate the prior review of submitted manuscript by an expert with statistical background. It is not so uncommon to find many glaring statistical error in submitted research manuscript which are missed out easily by non statistical experts and scientists. For example, innocuously looking terms like ‘variable’ versus ‘parameter’ indeed have different connotations. In a study sample, researcher deals with variables whereas parameter refers to same outcome in target population. In addition, serially measured data over repeated time points between two groups is mistakenly analyzed using t-tests or Mann Whitney U test at each time point separately. This is indeed incorrect and one needs to compare trend of outcome measured serially at all touch points by repeated measure ANOVA before embarking on posthoc pairwise comparison. Distribution of ‘time to event data’ is not normal or Gaussain and such data are marked by censoring, but researcher may wrongly compare its mean and variance using independent t-test. Ideal technique for analysis of such data is univariate (Kaplan Meier) or multivariate (Cox proportional hazard) survival analysis. There are myriad of statistical fallacies lurking beneath a submitted manuscript which a novice research is not usually aware of. Hence, there is serious need for critical review of manuscript prepared for submission in any reputed journal.