Missing Data in Research — When to Drop Rows and When to Use Multiple Imputation
A practical guide to missing data: MCAR, MAR, and MNAR patterns, when complete-case analysis holds, and why mean imputation distorts variance.
Practical guides with worked examples. Most articles are in Arabic; titles and summaries are shown here in English.
A practical guide to missing data: MCAR, MAR, and MNAR patterns, when complete-case analysis holds, and why mean imputation distorts variance.
How to choose between parametric and non-parametric tests: Shapiro-Wilk limits in large samples, CLT robustness, matching test pairs, and reporting.
How to calculate sample size and statistical power before data collection: the 4 inputs, a worked two-means formula, dropout adjustment, and post-hoc power.
Learn the exact arithmetic difference between Odds Ratio (OR) and Relative Risk (RR), why OR exaggerates RR when outcomes exceed 10%, and how to report both.
A practical guide to survival analysis: right censoring, why averaging follow-up time fails, a step-by-step Kaplan-Meier table, log-rank test, and Cox models.
Learn to distinguish paired vs. independent designs, see arithmetic proof of how independent tests lose power on paired data, and choose the right test.
How to handle outliers in research: entry errors vs. biological extremes, worked IQR fence and z-score rules, robust methods, and sensitivity analysis.
Understand Cronbach's alpha with a step-by-step worked example, why item count inflates alpha, the reverse-coding trap, and when to report McDonald's omega.
A complete decision tree for choosing the right statistical test. From group comparisons to correlation to regression to repeated measures. 4 questions determine everything.
A confounder makes you see relationships that don't exist. Learn how to identify confounders and control them with multivariable regression.
ANOVA tells you there's a difference — but not where. Learn ANOVA types, Post-Hoc Tests (Tukey, Bonferroni, Dunnett), and non-parametric alternatives.
Using Linear Regression with a binary outcome is the most common thesis mistake. Learn the differences between Linear, Logistic, Multinomial, Ordinal, Poisson, and Cox Regression.
Regression in research is a tool for understanding relationships and controlling confounders, not just prediction. Learn 3 real uses and variable selection.
I² = 89% and don't know what to do? Learn how to measure and handle heterogeneity with subgroup analysis, sensitivity analysis, and random effects.
90% of PhD students can't properly explain a Forest Plot. Learn every element and how to present it to your defense committee in 3 sentences.
Choosing the right statistical test is not memorization. 3 questions determine the correct test: outcome type, number of groups, and distribution.
A researcher spent 4 months on a flawed meta-analysis due to a weak protocol. Learn the 6 essential questions to answer before opening any database.
Over 800 researchers misinterpret the p-value. Discover why the p-value alone is insufficient and learn about effect size and confidence intervals.