Practical guides · Page 17
Search and filter all practical guidesReview Table Shells Before Filling Them with Results
Resolve populations, denominators, units, time points, and footnotes before generating research tables.
Turn an Allocation Ratio into Transparent Group Counts
Separate group-count arithmetic, sample-size justification, and the randomization procedure when planning allocation.
Check Expected Counts Before a Contingency-Table Test
Audit category counts, independence, and sparse cells before interpreting a categorical association test.
Audit a Research Data Dictionary Before Analysis
Match each dataset column to a clear definition, value labels, missing-value rules, and a traceable derivation.
Audit Two-Rater Agreement Before Reporting Kappa
Build a labeled agreement table and interpret Cohen’s kappa alongside observed agreement and category frequencies.
Summarize an Ordinal Scale Without Hiding Its Categories
Preserve ordered response labels, denominators, and scoring rules when describing ordinal research outcomes.
Use Discordant Pairs in an Exact McNemar Test
Preserve matched binary outcomes and separate directional change from overall agreement before testing.
Document a Reproducible Randomization Sequence Safely
Record generation settings and independent checks while preserving allocation concealment during enrollment.
Use File Checksums to Verify a Research Handoff
Record SHA-256 digests to detect byte-level file changes while keeping content validation and access control separate.
Use Weight-Based Effective Sample Size as a Diagnostic
Inspect unequal weights without confusing a simple effective sample calculation with full survey-design precision.
Check Variable Names Without Silently Changing Their Meaning
Find duplicate, ambiguous, and incompatible column names before renaming research data fields.
Assess Measurement Agreement Beyond Correlation
Plan paired differences, limits of agreement, and acceptable error before comparing two measurement methods.