Re: Gen Theory Part 2

David Winsemius wrote:
Rob <user@xxxxxxxxxxx> wrote in
news:zlbpj.10458$421.6712@xxxxxxxxxxxxxxxxxxxxxxxxxx:

I trying to determine if there is a difference between staff
retention times for Baby Boomers vs Gen X

Looking at a list of employees that have joined our company in one particular job over the last ten years.

I have eliminated those people that have joined in the last 10years
but have not left, from the sample( I did not see how I could use
that data)

Which is a gaping flaw in your method. You threw away the "denominator".

snipped output
Can I infer from this that there is a difference in retention rates between the two groups and that it is highly statistically
significant?


No. You can say that the duration of employment for the leavers is different between the two groups but you did not analyze rates.

You are taking an event that is distributed over time and then analyzing it with a cross-sectional methods. Rates should be analyzed by methods that account for the numbers of persons employed at a particular time and then determine the event rates = #leaving/#present_in_prior_period/length_of_period.

Survival analysis incorporates all of those features and avoids throwing away the highly relevant information regarding the persons "that have joined in the last 10years but have not left". You really do want that information to be included in your analysis. Those people form the risk set for the "leaving process".


Survival analysis ! Thank you very much for this. I have a lot of reading to do.
.

Quantcast