For the analysis of combinations of 2 x 2 non-contingency tablea 88 obtuined from density followup studiea (relating a number of evente to a number of person-years of follow-up) an analogue of theMante1-Haenszel test for 2 ~2 contingency tables is widely u d . In this paper the small sample properti
Tests of Hypotheses in Density Follow-Up Studies
β Scribed by B. M. Bennett
- Publisher
- John Wiley and Sons
- Year
- 2007
- Tongue
- English
- Weight
- 217 KB
- Volume
- 32
- Category
- Article
- ISSN
- 0323-3847
No coin nor oath required. For personal study only.
β¦ Synopsis
This paper discusses statistical methods for situations in which two or more ( k + l ) exposure groups, including a reference one, are compared with mpect to rate ratios, or incidence density ratios (IDR) (e.g. ~T T M E N , 1976). Data are assumed availablf on population times (PT) experienced in the exposure groups.
The logit transformation or its generalizations (e.g. Cox, 197C are utilized in order to obtain Fitable tests for the hypotheses involved. Three examples i l l d r a t e these methods, including approximate and exact confidence limits for the IDRs (k = 1). For the case of ( E + 1) groups a linear contrast for the hypothesis of a trend is derived as well as its standard error.
π SIMILAR VOLUMES
Longitudinal studies of dementia are often undertaken to estimate the incidence of dementia and to identify incident cases for the study of risk factors measured on the entire cohort at baseline. The power of these analyses is determined primarily by the number of demented cases identiΓΏed. Increasin
## Abstract Although longitudinal research is essential in understanding the nature and course of posttraumatic mental health problems, high rates of attrition often threaten the internal validity of such studies and make results hard to interpret. C. K. Scott (2004) developed an approach to minimi