Confidence Interval for Lost Time accidents?

[Cooper, Tim (Australia)]

Hi,

If a Lost Time Accident rate was 6 incidents last year and the target is
4 this year, is this a realistic target, i.e. it possible to show that
these rates are not significantly different?

Common sense says that if you have over a thousand employees and
typically only have 5 to 7 Lost Time Accidents a year, then this is so
close to 4 that it is probably not significantly different.  But how do
I calculate this?  

These rates are not far from zero anyway and there'll always be some one
who makes a mistake, so decreasing the target below some level is
probably not workable.

If I apply a binomial approach, what do I set N as?  Any employee could
have more than one Lost Time Accident, so the N can't simply be the
number of employees.  The Binomial Test's required assumption of
independence may not apply as the same process or even same person could
be responsible.

Really appreciate any help on this as I don't recall this problem on my
stats courses!


Regards,

Tim Cooper


Notice:  This e-mail message, together with any attachments, contains information of Merck & Co., Inc. (One Merck Drive, Whitehouse Station, New Jersey, USA 08889), and/or its affiliates (which may be known outside the United States as Merck Frosst, Merck Sharp & Dohme or MSD and in Japan, as Banyu - direct contact information for affiliates is available at http://www.merck.com/contact/contacts.html) that may be confidential, proprietary copyrighted and/or legally privileged. It is intended solely for the use of the individual or entity named on this message. If you are not the intended recipient, and have received this message in error, please notify us immediately by reply e-mail and then delete it from your system.

====================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

[Marta Garcia-Granero]

Hi

The standard error of a total (total number of incidents in ayer, in
your case) is approximately its square root. Therefore, you can
construct a 95%CI using this formula:

95%CI=n+/-1.96*SQRT(n).

anyway,one condition for the validity of this method is the independence
of the events, and you mention you don't think it will be fulfilled by
your data...

HTH,

Marta García-Granero

> Hi,
>
> If a Lost Time Accident rate was 6 incidents last year and the target is
> 4 this year, is this a realistic target, i.e. it possible to show that
> these rates are not significantly different?
>
> Common sense says that if you have over a thousand employees and
> typically only have 5 to 7 Lost Time Accidents a year, then this is so
> close to 4 that it is probably not significantly different.  But how do
> I calculate this?
>
> These rates are not far from zero anyway and there'll always be some one
> who makes a mistake, so decreasing the target below some level is
> probably not workable.
>
> If I apply a binomial approach, what do I set N as?  Any employee could
> have more than one Lost Time Accident, so the N can't simply be the
> number of employees.  The Binomial Test's required assumption of
> independence may not apply as the same process or even same person could
> be responsible.
>
> Really appreciate any help on this as I don't recall this problem on my
> stats courses!
>
>
> Regards,
>
> Tim Cooper
>
>
> Notice:  This e-mail message, together with any attachments, contains information of Merck & Co., Inc. (One Merck Drive, Whitehouse Station, New Jersey, USA 08889), and/or its affiliates (which may be known outside the United States as Merck Frosst, Merck Sharp & Dohme or MSD and in Japan, as Banyu - direct contact information for affiliates is available at http://www.merck.com/contact/contacts.html) that may be confidential, proprietary copyrighted and/or legally privileged. It is intended solely for the use of the individual or entity named on this message. If you are not the intended recipient, and have received this message in error, please notify us immediately by reply e-mail and then delete it from your system.
>
> ===================
> 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