Monthly Rules of Thumb
Home
Year
Month
Rule
Reference
Bio
2004
May
Sample Sizes When the Variance is Estimated.
Rule 2.1
April
Sample Size for Ratios of Means.
New Rule 2.13
March
Scales of Measurement.
Rule 1.13
Book Contents
February
Dichotomize Continuous Variables for Odds Ratio Analysis.
New Rule 4.13
Monthly Rule
January
Think about SPIN and SNOUT.
New Rule
Errata
2003
December
Don't Summarize Regression Sampling Schemes with Correlation.
Rule 3.2
Links
November
Be Eclectic and Ecumenical in Inference.
Rule 8.11
Book List
October
Web Resources.
Rule 1.18
Presentations
September
Hard and Fast Rule: Always use p=0.05. Not.
New Rule 1.18
Guest Book
August
Use At Least Twelve Observations in Constructing a Confidence Interval
Rule 1.10
Contact
July
Very non-significant P-values are very significant.
New Rule 1.17
Author
June
Correlations need to be substantial to gain advantage in ANCOVA.
New Rule 6.14
Webmaster
Think beyond simple ANOVA when a factor is time or dose--think ANCOVA. Case B: Factorial ANOVA
New Rule 6.13
Think beyond simple ANOVA when a factor is time or dose--think ANCOVA. Case A: One-way ANOVA
In Screening, Ruling Out Disease Requires High Sensitivity, Ruling In Disease Requires High Specificity
The Odds Ratio Approximates the Relative Risk Assuming that the Disease is Rare
Rule 4.2
Begin with a Basic Formula for Sample Size
2002
Listen to, and Heed the Advice of Experts in the Field
Rule 8.12
Number of Events per Variable
Rule 4.6
Be Wary of Surrogates and Accept Substitutes Warily
Rule 4.9
Use Text For a Few Numbers, Tables for Many Numbers, Graphs for Complex Relationships
Rule 7.1
Any Basic Statistical Computing Package Will Do
Rule 8.8
Assess Independence, Equal Variance, and Normality---in That Order
Rule 1.4
Multiple Comparisons
Rule 6.12
Distinguish Between Observational and Experimental Studies
Rule 1.1
© Gerald van Belle
May 3, 2004
