#  This returns data that follow a logit model with the
#  supplied parameters. n is the number of observations.
#  Note how the binary 0/1 variable is generated so that
#  it really follows the logit model
function [y, x] = LogitDGP(n, theta)
	k = rows(theta);
	x = [ones(n,1), randn(n,k-1)];
	y = (1 ./ (1 + exp(-x*theta)) > rand(n,1));
endfunction