A statistical question

[Chang,Victor]

Dear Liters:

I need your advice regarding the best way to calculate the worker turn-over
rate in the following:

Month     Total Termi   Turnover Rate
Sep-2005        395     20      5.1%
Oct-2005        411     10      2.4%
Nov-2005        413     15      3.6%
Dec-2005        403     9       2.2%
Jan-2006        411     11      2.7%
Feb-2006        417     14      3.4%
Mar-2006        416     7       1.7%
Apr-2006        423     11      2.6%
May-2006        425     13      3.1%
Jun-2006        424     17      4.0%
Jul-2006        420     12      2.9%
Aug-2006        410     23      5.6%
AVERAGE 414     14      3.3%

The turnover rate I calculate was to divide the termination by total worker,
on a monthly basis.  The average for the year turnover rate is 3.3% by
average.

Another person, on the other hand, use the sum of each month termination
(176) and then divided by the 414 total workers.  The result is 42.4%.

My point is that the total worker is a constant and non-cumulative variable;
while termination is a cumulative one that is added to the workforce each
month.  It will be a logical flaw to calculate the rate by using
non-cumulative variable against cumulative one, thus inflating the turnover
rate.  In other words, can we add each month's turnover rate to be a year's
one for reporting?

Any advice to this issue will be highly appreciated.

Thanks.

Victor Chang
[hidden email]

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Bob Walker-2]

Hi Victor,

Your weighted approach (e.g., 162/4968=3.3%) correctly reflects the monthly
dynamic: each month, for every 100 workers, about 3 will be terminated. Your
other approach (taking cumulative terminations, but dividing by the average
worker base) makes no sense to me. You could use average workers and average
terminations, but you'll end up in the same place (13.5/414=3.3%).

Bob Walker
Surveys & Forecasts, LLC

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Chang,Victor
Sent: Thursday, March 13, 2008 11:39 AM
To: [hidden email]
Subject: A statistical question

Dear Liters:

I need your advice regarding the best way to calculate the worker turn-over
rate in the following:

Month     Total Termi   Turnover Rate
Sep-2005        395     20      5.1%
Oct-2005        411     10      2.4%
Nov-2005        413     15      3.6%
Dec-2005        403     9       2.2%
Jan-2006        411     11      2.7%
Feb-2006        417     14      3.4%
Mar-2006        416     7       1.7%
Apr-2006        423     11      2.6%
May-2006        425     13      3.1%
Jun-2006        424     17      4.0%
Jul-2006        420     12      2.9%
Aug-2006        410     23      5.6%
AVERAGE 414     14      3.3%

The turnover rate I calculate was to divide the termination by total worker,
on a monthly basis.  The average for the year turnover rate is 3.3% by
average.

Another person, on the other hand, use the sum of each month termination
(176) and then divided by the 414 total workers.  The result is 42.4%.

My point is that the total worker is a constant and non-cumulative variable;
while termination is a cumulative one that is added to the workforce each
month.  It will be a logical flaw to calculate the rate by using
non-cumulative variable against cumulative one, thus inflating the turnover
rate.  In other words, can we add each month's turnover rate to be a year's
one for reporting?

Any advice to this issue will be highly appreciated.

Thanks.

Victor Chang
[hidden email]

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Richard Ristow]

At 11:38 AM 3/13/2008, Chang,Victor wrote:

>I need your advice regarding the best way to calculate the worker
>turn-over rate in the following:
>
>Month     Total Termi   Turnover Rate
>Sep-2005        395     20      5.1%
>Oct-2005        411     10      2.4%
>Nov-2005        413     15      3.6%
>Dec-2005        403     9       2.2%
>Jan-2006        411     11      2.7%
>Feb-2006        417     14      3.4%
>Mar-2006        416     7       1.7%
>Apr-2006        423     11      2.6%
>May-2006        425     13      3.1%
>Jun-2006        424     17      4.0%
>Jul-2006        420     12      2.9%
>Aug-2006        410     23      5.6%
>AVERAGE 414     14      3.3%
>
>The turnover rate I calculate was to divide the termination by total
>worker, on a monthly basis.  The average for the year turnover rate
>is 3.3% by average.
>
>Another person, on the other hand, use the sum of each month
>termination (176) and then divided by the 414 total workers.  The
>result is 42.4%.

42.4/3.3=12.8

Your co-worker got a total rate PER YEAR, you got a mean rate PER
MONTH. Take whichever one is relevant.

Notice that the quotient isn't exactly 12. That's because the total
for the year isn't the sum of the month rates; it's a weighted sum,
weighted by total employment in the month. I'm not doing the
arithmetic to demonstrate that.

-Best wishes,
  Richard

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD