Harvard Biostatistics Multiple Lecture Courses

## Transformation of the data

### Statistics Notes: Transforming data

Source: Bland J Martin, Altman Douglas G. Statistics Notes: Transforming data BMJ 1996; 312 :770

How to access: This is free to access from the above link.

Example(s) used in this: No sample data

Prior math/stat knowledge to fully understand materials: Normal distribution, variance of the data

### Statistics notes: Transformations, means, and confidence intervals

Source: Bland J Martin, Altman Douglas G. Statistics notes: Transformations, means, and confidence intervals BMJ 1996; 312 :1079

How to access: This is free to access from the above link.

Contents: This article illustrate how to construct confidence intervals with the transformed data.

Example(s) used in this: Serum triglyceride data

Prior math/stat knowledge to fully understand matrials: Standard deviation, standard error, confidence intervals

### Statistics Notes: The use of transformation when comparing two means

Source: Bland J Martin, Altman Douglas G. Statistics Notes: The use of transformation when comparing two means BMJ 1996; 312 :1153

How to access: This is free to access from the above link.

Contents: This article discusses a challenge in comparing two means with the transformed data. The conclusion is the logarithmic transformation is strongly preferable to other transformations because of interpretations.

Example(s) used in this: Biceps skinfold thickness (Crohn's disease vs. Coeliac disease)

Prior math/stat knowledge to fully understand materials: Standard deviation, standard error, confidence intervals