A new confidence interval for the difference of two binomial proportions
1990
Abstract A new confidence interval for the difference of two binomial proportions from independent samples is developed. The derivation depends on neither asymptotic nor conditional arguments. The interval is shown to be equivalent to a bootstrap interval constructed using infinite bootstrap sampling, but can also be motivated using first principles. An example is given and numerical comparison with alternative methods is provided.
Keywords:
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
7
References
7
Citations
NaN
KQI