sediff: differences in computed se

In phoebehlam/michaela: convert statistical metrics for meta-analytic purpose

differences in computed se

Description

the differences in the computed standard error using the upper bound vs. lower bound of 95 confidence interval
computations are based on t-distribution, not z
returns and prints a value/vector of difference(s)

Usage

sediff(est, cil, ciu, n, k, type = c("reg", "exp b", "geo", "percent"))

Arguments

est	the point estimate (e.g., a regression coefficient, a mean)
cil	the lower bound of the 95 confidence interval
ciu	the upper bound of the 95 confidence interval
n	the sample size
k	the total number of predictors. for converting the ci associated with a raw mean, enter k = 0.
type	select one of: "reg" for regular coefficient/mean "exp b" for exponentiated regression coefficients (e.g., estimated odd ratios from exponentiated logistic regression coefficients) and cis "geo" for geometric means (exponentiated arithmetic means of logged raw units) and cis "percent" for percent changes (exponentiated regression coefficients expressed in percent changes) and cis

Examples

# e.g., exponentiated logistic regression coefficient (estimated odd ratios) and cis
se.diff(4.05, 1.56, 10.51, 141, 8, "exp b")

phoebehlam/michaela documentation built on Oct. 23, 2024, 4:10 p.m.