Practical guides · Page 8
Search and filter all practical guidesForecast recruitment with staggered site activation
Use opening dates, ramp-up and mature site rates to estimate calendar recruitment capacity.
Define recurrent events before choosing a model
Specify episode boundaries, risk periods and terminal events so repeated outcomes represent the intended process.
Audit bandwidth choices in regression discontinuity
Check the assignment threshold, local support and sensitivity before interpreting a discontinuity.
Use Supplied Replicate Weights With the Correct Scaling Rule
Preserve the provider’s replication method instead of treating replicate columns as interchangeable samples.
Export Research Figures with Reproducible Size and Content
Preserve data, code, dimensions, and rendering settings so a publication figure can be regenerated accurately.
Inspect Residuals Against Fitted Values Before Trusting a Regression
Use residual patterns to distinguish mean-shape problems, changing spread, and observations needing investigation.
Pilot questionnaire burden with real task observations
Measure completion effort and comprehension before shortening a survey or launching recruitment.
Choose and interpret a restricted mean survival time
Prespecify the time horizon and explain a survival-curve area difference in time units.
Round Research Results Without Changing Their Interpretation
Keep calculations at full precision and apply a declared display policy consistently across text, tables, and figures.
Write Figure Descriptions That Preserve the Scientific Message
Provide an accessible summary and supporting detail without inventing conclusions absent from the plotted data.
Evaluate Screening Automation Before Letting It Exclude Records
Separate ranking assistance from exclusion decisions and measure missed eligible records on an independently checked sample.
Peer Review a Search Strategy Before Running the Final Search
Check concept coverage, Boolean logic, field choices, and restrictions against the review question.