“Good Research Method is Good Ethics”. There is a growing recognition on the importance of sample size calculation and power analysis in research.
When planning a study design, it is very imperative to consider how many participants are needed, therefore, sample size must be premeditated carefully to ensure that the research time, patient effort and support costs invested are not wasted. There are many standard statistical software and stand alone packages available in public domain (internet) for performing sample size/power calculations. However, validity of those software are not authenticated and documented.
On the other hand, the commercial software are unaffordable by the researchers from developing countries. Also, the software packages which are available do not provide the details of the literature behind the technique. Thus, there is a great urge for developing a holistic tool, which will provide the situation where that particular formula needs to be applied, assumptions involved, the relevant bibliography and then the details of the formula. This paved a way for the development of ‘nMaster 2.0’.
Means : Estimating single mean, hypothesis testing of 2 & >2 means (paired and unpaired samples, repeated measurements, ANCOVA)
Proportions: Estimating a single proportion, hypothesis testing of two proportions (equal & unequal), matched paired designs
Agreement :Kappa, Intra-class correlation coefficient (ICC)
Diagnostic Tests : Sensitivity, Specificity, PPV, Positive & Negative likelihood ratios
Regression Methods :Simple correlation & regression, simple & multiple logistic regressions
Survival Analysis :Log-rank test, Cox regression, Exponential survival curves
Cluster Designs :Estimation (continuous & binary), comparison of means, proportions & incidence rates (matched & unmatched)
Equivalence Study : Paired & unpaired tests for means and proportions
Epidemiological Designs : Cohort, unmatched and matched case-control studies
Clinical trials : Comparison of Means and Proportions, Superiority trials, Equivale nce /Bio equivalence, Non inferiority trials from parallel and Crossover designs
non-Parametric : Sign test, will coxon signed rank test, Nann-whitney U test