Grimaldus said:
I spoke with a little birdy.
The 66% greater retention rate (sounds like a big number) turned out to apparently equal 2 troops.
The platoon involved in it was the "warrior platoon". Electronics were taken away from a platoon of newbies who just started basic training. The platoon of warriors were allowed to keep their electronics, some of whom have been kicking around the system for a year +.
Some people when told they were loosing their electronics, quit. (Are those the kinda soldiers we want defending Canada?)
Quoting this again, because it seems to have been forgotten. Go back and read Grimaldus' posts on this. Now, my fancy education doesn't seem to count for a hell of a lot in the job market, but I know when a 'study' has been horribly botched.
If they wanted to do this properly, here's a thought for how to do the study: For a twelve month period (to account for possible variation in 'time of year'... People quitting because they don't like winter; people getting through easier because of a nice Christmas break, whatever) have one or two platoon at any given time given full personal electronic privileges from day 1 in accordance with such regulations as have been the case after the 'indoc' period. Using that sample, figure out over the course of one year what the VR rate from the couple hundred troops in that sample are. Any troops who VR in that time, have an exit interview with a questionnaire that asks about the influence of personal electronics. When the course graduates, same thing- did the troops in question feel better able to keep on for having access to iToys or Snozzberies and talking to family/friends/stock broker?
For a subsequent twelve month period, simply have the no PED rule aplied as has been the case in the past. Same deal with exit interviews and graduation questionnaires.
This gives you two sample of a few hundred troops, each spanning an entire calendar year (eliminates seasonal variables, if they in fact exist), and your samples were not in CLFRS at the same time- you won't have a sample that is aware of the other, and perhaps resentful/envious thereof. At the end of the day, figure out what the correlated VR rate is, whether there's a difference, and what that difference means in hard numbers.
Yup, it would take two years to do- but so what? Why rush things like this when the difference seems anecdotally to be at best a few troops per platoon, based off a horribly flawed sample?