R/ci.sc.ancova.R

In MBESS: The MBESS R Package

ci.sc.ancova <-
function(Psi=NULL, adj.means=NULL, s.anova=NULL, s.ancova=NULL, standardizer="s.ancova", c.weights, n, cov.means, SSwithin.x, conf.level=.95)
{options(warn=-1)
if(standardizer=="s.anova")
{
if(is.null(s.anova)) stop("'s.anova' is needed to standardized the contrast (this is the standard deviation of the errors from the ANOVA model)")
if(missing(s.ancova)) stop("'s.ancova' is needed to standardized the contrast (even when using 's.anova' as the standardizer; this is the standard deviation of the errors from the ANCOVA model).")
}
if(standardizer!="s.ancova" & standardizer!="s.anova") stop("The standardizer must be either 's.anova' or 's.ancova'.")

if(missing(cov.means)) stop("The mean of the covariate within each group (i.e., the vector 'cov.means') is missing.") 
if(is.null(Psi) & is.null(adj.means) ) stop("Input either 'Psi' or 'adj.means'")
if(!is.null(Psi) & !is.null(adj.means) ) stop("Do not input both 'Psi' and 'adj.means'")

if(missing(SSwithin.x)) stop("The sum of squares within from the ANOVA on the covariate is missing (i.e., 'SSwithin.x').") 
  
if(is.null(Psi)) Psi<- sum(adj.means*c.weights)

if(sum(c.weights)!=0) stop("The sum of the coefficients must be zero")
if(sum(c.weights[c.weights>0])>1) stop("Please use fractions to specify the contrast weights")

J<-length(c.weights)
if(length(n)>1)
    {if (length(n)!=J) stop("'c.weights' and 'n' imply different numbers of groups.")}
if(length(n)==1) n<-rep(n, J)

if(length(cov.means)!=J) stop("'c.weights' and 'cov.means' imply different numbers of groups.")

f.x.numerater<- (sum(c.weights*cov.means))^2
f.x.denominator<- SSwithin.x
sample.size.weighted<- sum(c.weights^2 / n )
ratio<- s.ancova/s.anova
alpha<-1-conf.level
nu<-sum(n)-J-1

if(standardizer=="s.ancova")
    {
    psi <- Psi/s.ancova
    lambda.obs<- psi/sqrt(sample.size.weighted+(f.x.numerater/f.x.denominator))
    lambda.limits<-conf.limits.nct(ncp=lambda.obs, df=nu, conf.level=1-alpha)
    
    psi.limit.upper<- lambda.limits$Upper.Limit*sqrt(sample.size.weighted+(f.x.numerater/f.x.denominator))
    psi.limit.lower<- lambda.limits$Lower.Limit*sqrt(sample.size.weighted+(f.x.numerater/f.x.denominator))
    }

if(standardizer=="s.anova")
    {
    psi <- Psi/s.anova
    lambda.obs<-psi/(ratio*sqrt(sample.size.weighted+(f.x.numerater/f.x.denominator)))
    lambda.limits<-conf.limits.nct(ncp=lambda.obs, df=nu, conf.level=1-alpha)
    
    psi.limit.upper<- lambda.limits$Upper.Limit*ratio*sqrt(sample.size.weighted+(f.x.numerater/f.x.denominator))
    psi.limit.lower<- lambda.limits$Lower.Limit*ratio*sqrt(sample.size.weighted+(f.x.numerater/f.x.denominator))
    }

list(standardizer=standardizer, psi.limit.lower=psi.limit.lower, psi=psi, psi.limit.upper=psi.limit.upper)
}