One of the activities that the agency I work for has been doing for a while is tracking career movement using a derivative of the pay-file, and the modest amount of tombstone information it contains. When the pay grade changes, or the location code for the pay changes, the working assumption is that the employee has either moved to another job, or has been reclassified in some manner. The reclassification could be because they were promoted via a competition, because they were reclassified in place resulting from an increment to duties, or because they are temporarily acting for someone who is away for some reason. There are plenty of ambiguous cases, but these are verified via followup. Naturally, you have to have a means to be able to differentiate someone who is now getting paid at a different rate in a different place because they have transferred to another unit of the same agency, and a person who has changed agencies entirely. That’s why I say we work with a derivative of the pay-file, rather than the pay-file itself (which is only intended to make sure your pay and benefits are what they are supposed to be).
As you can imagine, the concern is not with individuals but with overall patterns of movement. It provides a means to measure movement across organizations, and things like how quickly individuals or identifiable groups entering via particular streams move up, as well as pan-governmental “churn”. The result comes in the form of reports like one a colleague published this past autumn ( http://www.psc-cfp.gc.ca/centres/stat-bulletin/2011/pdf/stat-bullet… )
Not all administrative data systems will necessarily permit this sort of data use, whether because of incompleteness, incompatibility, or some very understandable security concerns. And it does require a certain number of FTEs to generate such information; it doesn’t just happen by a few magical clicks. But as long as people draw a paycheque from a common source, there will be a means to see how long they stay put before switching jobs, using data from admin systems.
To the extent that point of entry can be linked to a given recruitment program or commencement of some organizational practice, one could hypothetically examine tenure duration within position, and within-organization, as a function of institutionalization of policies or practices. While such an approach certainly provides some degree of precision and comprehensiveness, it obviously doesn’t really shed much light on what the employee sees as keeping them in place, or the weight they give to multiple factors. After all, simply knowing that these people were hired after that practice came into effect doesn’t really say anything about how the practice was implemented in their case, or how they were exposed to it.
But that said, you certainly CAN ask people about their departure intentions, their departure or stay motivations, and their age and tenure.
On the theme of stuff you can and can’t control. The 2002 Canadian Public Service Employee Survey asked people about their departure intentions and motives. At my urging, they were asked to give simple importance ratings to each of a list of some 13 or more factors, and also asked to give a yes/no/maybe rating for departure intentions. Using age and tenure, I identified those who were eligible for retirement and had indicated intentions to leave within the next 5 years. I further identified those who had indicated retirement as an important reason for leaving, and divided them up into those whose additional reasons that were nominated as important were those the employer could do nothing about (e.g., family commitments, health), and those that were somewhat within the employer’s purview (e.g., workload/stress, making better use of the employee’s training/skills), and was able to identify a cluster of about 10% or so (don’t quote me) whose departure was, in some sense, “remediable”. That is they were intending to retire but it seemed to be for reasons that the employer could conceivably fix (e.g., via job redesign).
Now, I do realize the thread is about retaining those one recruits, but the underlying logic is the same as asking about imminent retirees: you develop survey content that addresses their motivation, identifies their circumstances in some organizationally-relevant way, and ask some sensible questions about their departure intentions and job search behaviour. I might point out that, with unemployment and student debt burden being what it currently is, expect to find a LOT of people coming in to jobs they consider to be inappropriate for themselves, simply to have a foot in the door. In a kind of backhanded way, their quick movement to another more suitable job, and the rabid job-search behaviour that precedes it, is a barometer of their desire to remain within the PS. We shouldn’t necessarily view it as a precursor to them waving goodbye. Consider it like fidgiting around in bed before you finally get comfortable and nod off. But again, that sort of interpretation is only comforting if one views the new recruit as adding value to the overall enterprise of government, rather than just one’s own work unit. That’s why I say, you need to think about what you mean by “retention”.
Is this along the lines you were thinking?