uncond gives a two-sided true profile likelihood confidence interval for the difference of two dependent proportions. It is built by the solution of an inequality. Data are assumed to be of a fourfold table, which contains the number of concordance and the number of discordance of two dependent methods.

1 |

`a` |
first number of concordant paires as described above |

`b` |
first number of discordant paires as described above |

`c` |
second number of discordant paires as described above |

`d` |
second number of concordant paires as described above |

`n` |
number of observed objects |

`alpha` |
type I error; between zero and one |

The true profile likelihood confidence interval has as lower limit the minimum of the solutions for the inequality of the maximum likelihood function and the quantile of the normal distribution. The upper limit is defined as the maximum solution of this inequality.

A list with class '"htest"' containing the following components:

`conf.int ` |
a confidence interval for the difference in proportions |

`estimate ` |
estimated difference in proportions |

Daniela Wenzel, Antonia Zapf

Newcombe, R.G. (1998). Improved confidence intervals for the difference between binomial proportions based on paired data. Statistics in Medicine 17. 2635-2650.

1 2 | ```
# a=10, b=15, c=5, d=20, n=50, type I error is 0.05
conf.int=uncond(10,15,5,20,50,0.05)
``` |

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.

Please suggest features or report bugs with the GitHub issue tracker.

All documentation is copyright its authors; we didn't write any of that.