multicon (version 1.6)

tContrast: Constrast T-tests

Description

Computes a t-test for multiple groups using a given set of contrast weights.

Usage

tContrast(IV, ...)
## Default Method
"tContrast"(IV, DV, wgt = c(1, -1), 
	alpha = .05, EQVAR = FALSE, alternative = "unequal", ...)
## Method for class 'formula'
"tContrast"(formula, data = NULL, wgt = c(1, -1),  
	alpha = .05, EQVAR = FALSE, alternative = "unequal", ...)

Arguments

IV

A factor of the same length as DV containing the independent variable codes.

DV

A numeric vector of the same length as IV containing the measured values.

formula

A formula of the form lhs ~ rhs where lhs is a numeric vector containing the data values and rhs is a variable containing the corresponding groups.

data

An optional data frame containing the variables in the formula.

wgt

A numeric vector containing the contrast weights corresponding to each successive level of the IV. Defaults to c(1, -1), implying that the first group is expected to have a higher mean than the second.

alpha

A numeric element > .00 and < 1.00 specifying the Type I error rate.

EQVAR

A logical indicating whether equal variances amongst the groups should be assumed. Defaults to FALSE (Welch's Method).

alternative

A character vector specifying the alternative hypothesis. Must be one of "unequal", "greater", or "less".

...

Further arguments to be passed to or from methods.

Value

Ms

A data.frame with the sample size, mean, and weight for each group.

test

A data.frame with the test statistic (stat), the degrees of freedom (df), the critical value for the test statistic (crit), the p-value, and an r-contrast (effect size).

Details

This function computes a t-contrast for any number of groups based on the specificed constrast weights (Rosenthal, Rosnow, & Rubin, 2000). By setting the EQVAR option to TRUE degrees of freedom are consistent with Student's method. If EQVAR is FALSE (default) then degrees of freedom are calculated using the Welch-Sattertwaite approximation. The wgt option allows one to specify contrast weights to test hypotheses with more than 2 levels of an IV. By default it tests the hypothesis that two means are unequal. If a directional hypothesis is known ahead of time, use "greater" to predict that higher contrast weights have higher means and "less" to predict the opposite. For a robust version of this function see yuenContrast. The entire family of possible T-test equations can be found here:

http://rynesherman.com/T-Family.doc

References

Rosenthal, R., Rosnow, R. L., & Rubin, D. B. (2000). Contrasts and Effect Sizes in Behavioral Research: A Correlational Approach. Cambridge, UK: Cambridge University Press.

See Also

yuenContrast t.test

Examples

dv <- c(rnorm(30, mean=1, sd=2), rnorm(20))
iv <- c(rep(1,30),rep(2,20))

  # Student's t-test (assuming equal variances)
t.test(dv ~ iv, var.equal=TRUE)
  # Welch's t-test (not assuming equal variance)
t.test(dv ~ iv, var.equal=FALSE)
  # tContrast assuming equal variances
tContrast(iv, dv, EQVAR=TRUE)
  # tContrast not assuming equal variances
tContrast(iv, dv, EQVAR=FALSE)

  # Contrast with 3 Groups
dv <- c(rnorm(30), rnorm(20, mean=-.5), rnorm(10, mean=-1))
iv <- c(rep("c",30), rep("b", 20), rep("a", 10))
    # t-contrast with Welch-Sattertwaite DFs
tContrast(iv, dv, wgt=c(1, 0, -1))
    # Compare with yuenContrast with no trimming
yuenContrast(iv, dv, wgt=c(1, 0, -1), tr=0)
    # With the formula method
yuenContrast(dv ~ iv, wgt = c(1, 0, -1))