ri (version 0.9)

omni.ate: Omnibus function for estimation and testing

Description

Estimates the average treatment effect (ATE) and inferential statistics under constant effects hypotheses. Estimation is without covariate adjustment, via weighted least squares.

Usage

omni.ate(Y, Z, perms, invert = FALSE, quantiles = c(0.025, 0.975))

Arguments

Y

numeric vector of length N, outcome variable

Z

binary vector (0 or 1) of length N, treatment indicator

perms

N-by-r permutation matrix, as output by genperms or genperms.custom

invert

logical for generating constant effects confidence intervals through exact test inversion, with the difference-in-means as a test statistic. Default is FALSE.

quantiles

vector of quantiles of the randomization distribution to be returned. Quantiles also used to determine endpoints of confidence intervals. Default is equal-tailed 95% intervals.

Value

ate

estimated average treatment effect

greater.p.value

one-tailed p-value: proportion of randomizations yielding estimated ATE greater than or equal to hypothesized ATE

lesser.p.value

one-tailed p-value: proportion of randomizations yielding estimated ATE less than or equal to hypothesized ATE

p.value

two-tailed p-value: twice the smaller of the two one-tailed p-values, as advocated by Rosenbaum (2002)

p.value.alt

two-tailed p-value: proportion of randomizations yielding absolute estimated ATE greater than or equal to absolute hypothesized ATE

se.null

standard error of the randomization distribution assuming a zero treatment effect

conf.int

confidence interval approximation under a constant effect hypothesis

se

standard error of the randomization distribution assuming a constant treatment effect equal to the estimated ATE

conf.intInv

(Optional, if invert=TRUE) confidence interval under an inverted exact test with the difference-in-means as a test statistic

Details

omni.ate() is a convenience function that implements a number of functions otherwise available in ri. Greater flexibility through use of the individual functions involved.

References

Gerber, Alan S. and Donald P. Green. 2012. Field Experiments: Design, Analysis, and Interpretation. New York: W.W. Norton.

Rosenbaum, Paul R. 2002. Observational Studies. 2nd ed. New York: Springer.

Samii, Cyrus and Peter M. Aronow. 2012. On Equivalencies Between Design-Based and Regression-Based Variance Estimators for Randomized Experiments. Statistics and Probability Letters. 82(2): 365-370. http://dx.doi.org/10.1016/j.spl.2011.10.024

See Also

ri

Examples

y <- c(8,6,2,0,3,1,1,1,2,2,0,1,0,2,2) 
Z <- c(1,1,0,0,1,1,0,0,1,1,1,1,0,0,1)

perms <- genperms(Z) # all possible permutations of assignment

omni.ate(y,Z,perms,FALSE)
# omni.ate(y,Z,perms,TRUE) # may take some time to run