“…This suggests that in fact IDEA codes the dependent variable inconsistently and hence noisily. Yet, IDEA offers the only widely used (e.g., Brand, 2010;Gibson & McKenzie, 2012;Iheduru, 2011;Leal, Lee, & McCann, 2012), systematic measure of the concept of interest, so we use it here. 14 In any case, introducing noise into the dependent variable's measurement will most likely increase models' standard errors, thereby imposing a conservative bias on estimates of statistical significance.…”