ind.prop.second: A Z test for the equality of two proportions using published...
In rpsychi: Statistics for psychiatric research

Description Usage Arguments Details Value Author(s) References See Also Examples

ind.prop.second conducts a Z test for the equality of two proportions using published work.

1	ind.prop.second(x, n, sig.level = 0.05, digits = 3, ref.ind=1)

`x`	a numeric vector (length(`x`) = 2) contains the counts of successes
`n`	a numeric vector (length(`n`) = 2) contains the sample sizes
`sig.level`	a numeric contains the significance level (default 0.05)
`digits`	the specified number of decimal places (default 3)
`ref.ind`	the reference for the independent variable (default 1)

This function conducts a Z test for the equality of two proportions using published work. Statistical power is calculated using the following specifications:

(a) small (h = 0.20), medium (h = 0.50), and large (h = 0.80) population effect sizes, according to the interpretive guideline for effect sizes by Cohen (1992)

(b) sample size specified by n

The returned object of ind.prop.second contains the following components:

`samp.stat`	returns the proportions, sample sizes, and a effect size index (Cohen's h)
`risk.difference`	returns a risk difference, its' confidence interval, and standard error
`risk.ratio`	returns a risk ratio, its' confidence interval, and a standard error of a log-transformed risk ratio
`odds.ratio`	returns a odds ratio, its' confidence interval, and a standard error of a log-transformed odds ratio
`power`	returns statistical power for detecting small (h = 0.20), medium (h = 0.50), and large (h = 0.80) population effect sizes

Yasuyuki Okumura
Department of Social Psychiatry,
National Institute of Mental Health,
National Center of Neurology and Psychiatry
yokumura@blue.zero.jp

Cohen J (1992) A power primer. Psychological Bulletin, 112, 155-159.

Kline RB (2004) Beyond significance testing: Reforming data analysis methods in behavioral research. Washington: American Psychological Association.

ind.prop, samplesize.h

##Kline (2004) Chapter 5
x1 <- c("relapsed", "not relapsed")
y1 <- c("control", "treatment")

dat <- data.frame(y =         
factor(c(rep(x1, c(60, 40)), rep(x1, c(40, 60))), levels=x1),
x = factor(rep(y1, each=100), levels=y1)
)
tab <- xtabs(~x+y, data=dat)
tab
ind.prop.second(x=tab[,1], n = rowSums(tab))             #Risk for relapse is lower in treatment than control condition.
ind.prop.second(x=tab[,1], n = rowSums(tab), ref.ind=2)  #Risk for relapse is higher in control than treatment condition.