Must all of the data be reviewed for either truncation at the extremes (continuous traits) or grouping so that all counts are 20 or greater (categorical traits)?
The guidelines for modifying data are general and therefore not universally applicable. In general, only traits that have features that may be identifiable visually (e.g., height, weight, waist circumference) or could be ascertained through a casual knowledge of a subject (e.g., race, Hispanic status, marital status, employment status, education, income, significant outcomes of low frequencies such as cancers or specific causes of death) should be reviewed for possible modification. In general, multicenter studies have more leeway in performing fewer modifications since the removal of study center (geography) will, by itself, help to resolve many of the data modification issues.
Related Questions
- I am using the SuperTABLE files that contain data as counts for urban areas with populations of 1,500 or greater, but cannot find equivalent files with the data in percentages. Where are they?
- Must all of the data be reviewed for either truncation at the extremes (continuous traits) or grouping so that all counts are 20 or greater (categorical traits)?
- How can a tape data set be REVIEWed?
