RVAideMemoire (version 0.9-83-7)

wald.ptheo.test: Wald test for comparison of a proportion to a theoretical value

Description

Performs a Wald test for comparison of a proportion to a theoretical value.

Usage

wald.ptheo.test(y, blocks = NULL, p = 0.5)

Value

method

name of the test.

data.name

a character string giving the name(s) of the data.

statistic

test statistics of the test.

p.value

p-value of the test.

estimate

the estimated proportion (calculated without taking into account the blocking factor, if present).

alternative

a character string describing the alternative hypothesis, always "two.sided".

null.value

the value of the proportion under the null hypothesis.

parameter

the degrees of freedom for the t-statistic, only whith overdispersion and no blocks.

Arguments

y

either a binary response (numeric vector or factor, with only two possible values except NA) or a two-column matrix with the columns giving the numbers of successes (left) and failures (right).

blocks

optional blocking (random) factor.

p

hypothesized probability of success.

Author

Maxime HERVE <[e-mail]>

Details

The function builds a logistic (mixed) regression and applies a Wald test to compare the estimated value of the intercept to its theoretical value under H0. Eventual overdispersion is taken into account, by using a quasi-binomial law in case of no blocks or by introducing an individual-level random factor if blocks are present.

If the response is a 0/1 vector, the probability of the '1' group is tested. With other vectors, the response is transformed into a factor and the probability of the second level is tested.

If the response is a two-column matrix, the probability of the left column is tested.

If the reponse is a vector and no blocking factor is present, the exact binomial test performed by binom.test should be preferred since it is an exact test, whereas the Wald test is an approximate test.

See Also

binom.test, glm, glmer

Examples

set.seed(2006)
response <- sample(0:1,60,replace=TRUE)

# Comparison to p=0.5
wald.ptheo.test(response)

# Comparison to p=0.8
wald.ptheo.test(response,p=0.8)

# With a blocking factor

require(lme4)
blocks <- gl(3,20)
wald.ptheo.test(response,blocks)