Description Usage Arguments Details References See Also Examples

`mis`

is a `link-glm`

object that specifies the link function in Neuhaus (1999, expression~(8)) for handling misclassified responses in binomial regression models using maximum likelihood. A prior specification of the sensitivity and specificity is required.

1 | ```
mis(link = "logit", sensitivity = 1, specificity = 1)
``` |

`link` |
the baseline link to be used. |

`sensitivity` |
the probability of observing a success given that a success actually took place given any covariate values. |

`specificity` |
the probability of observing a failure given that a failure actually took place given any covariate values. |

`sensitivity`

+ `specificity`

should be greater or equal
to 1, otherwise it is implied that the procedure producing the
responses performs worse than chance in terms of misclassification.

Neuhaus J M (1999). Bias and efficiency loss due to misclassified responses in binary regression. Biometrika, **86**, 843-855 https://www.jstor.org/stable/2673589

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 | ```
## Define a few links with some misclassification
logit_mis <- mis(link = "logit", sensitivity = 0.9, specificity = 0.9)
lizards_f <- cbind(grahami, opalinus) ~ height + diameter + light + time
lizardsML <- glm(lizards_f, family = binomial(logit), data = lizards)
lizardsML_mis <- update(lizardsML, family = binomial(logit_mis),
start = coef(lizardsML))
## A notable change is coefficients is noted here compared to when
## specificity and sensitity are 1
coef(lizardsML)
coef(lizardsML_mis)
## Bias reduction is also possible
update(lizardsML_mis, method = "brglmFit", type = "AS_mean",
start = coef(lizardsML))
update(lizardsML_mis, method = "brglmFit", type = "AS_median",
start = coef(lizardsML))
``` |

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.