On 05/04/2016 09:24 AM, Neotarf wrote:
although you would think the Gruniad would not report on something that was obviously flawed.
It's hard to know Gruniad's intention, given the research articles are also relatively old (not just published or forthcoming):
1. Bingham & Scherer (2001) 2. Tinkler, Li, & Mollborn (2007)
Also TL&M (2007) don't cite B&S (2001)...
It goes without saying that a meaningful study should have a random selection process, although it happens all the time that researchers can't always get ideal populations so they study the populations they have. Unfortunately the study is behind a paywall, so you can't see how it was designed,
Here's the appropriate bit for B&S (2001).
! We randomly assigned departments to the experimental (program participation) ! and control (program nonparticipation) conditions. Staff and faculty ! departments were chosen from separate lists, with staff departments organized ! by type of labor (administrative vs. nonadministrative) and faculty ! departments organized by college; this stratification procedure assured that ! different categories of staff and faculty departments in each of the colleges ! would be represented in the two conditions. Random selection of departments ! rather than individuals resulted in a quasi-experimental design in which each ! employee and combination of employees did not have an equal chance of being ! assigned to one of the two conditions of the study. Because our unit of ! analysis was the individual employee, we explored possible sources of bias due ! to differences between employees who were assigned to the experimental and ! control conditions. Specifically, we compared the questionnaires returned by ! experimental and control groups on both demographic and experiential ! variables. Pearson chi-square analyses (p ≤ .05) were performed to test the ! associations between participation condition and each variable. The ! composition of the two respondent groups was not significantly different in ! terms of gender, race/ethnicity, position at the university, prior experience ! as a perpetrator of sexual harassment, experience being accused of sexual ! harassment, or experience as a sexual harassment victim. The phi coefficients ! for this set of variables ranged from .03 to .06. These results suggest that ! employees in the two groups were comparable in important respects prior to the ! intervention.
So it doesn't look like it's suffering from selection bias (wherein the abusers are sent to the training and have worse attitudes from the start).
All behind paywalls.
If someone did want to write a WP article, I'm willing to help on this point.