Practical guides · Page 15
Search and filter all practical guidesRecover Counts Compatible with Rounded Percentages
Identify integer counts consistent with a reported percentage without claiming that rounding reveals the original data.
Plan Test Length with Spearman–Brown
Forecast reliability after changing questionnaire length while keeping item quality, burden, and assumptions visible.
Preserve Structured Abstract Sections for Screening
Keep section labels and abstract types when building screening records from article XML.
Build a Manifest of Supplementary Files
Match each cited appendix to the actual attachment and identify missing files before extraction.
Carry Table Footnotes Into an Evidence Extraction
Link cells to their table notes, denominator definitions, and units before copying a result.
Transform Means and Spread Under an Affine Conversion
Apply offsets to location and scale factors to variability while preserving interval ordering and covariance meaning.
Define Competing Events Before Estimating Event Probability
Separate the first-event question from censoring and choose a probability summary that matches the endpoint.
Interpret Logistic Interactions Through Predicted Probabilities
Distinguish log-odds interaction from risk-difference contrasts using an explicit four-scenario prediction grid.
Check Logistic Separation Before Trusting Large Coefficients
Recognize perfectly predicted outcomes and investigate unstable likelihood estimates before reporting odds ratios.
Bound a Binary Risk Difference Under Missing Outcomes
Explore extreme event assignments without mistaking logical bounds for a confidence interval or an imputation recommendation.
Report Monte Carlo Precision for Simulated Probabilities
Distinguish simulation repetitions from study sample size and quantify uncertainty in estimated coverage or rejection rates.
Make Regression Reference Categories Explicit
Audit factor levels and coefficient meanings so model tables and predictions describe the intended comparison.