Employer using Timetracking data to compare averages is not statistically sound?

[zachtaylor008]

I work at a tropical fish wholsaler, on land as an aquarist. My job is to keep the animals alive, disease-free, happy and healthy. We have over 15 independent tank systems that can be at 0% capacity (empty) or 100% capacity on any given week. Given that we work with live animals, the nature of the work and time spent in the day fluctuates, such as length of time feeding, length of time cleaning, etc. My employer uses data from a timetracking database to compare weekly time spent on over 16 activities at work to their historical averages, per employee. However, each activity is affected by a number of variables, as I said above, and there is no way to account for these variables in the database. For instance, on week we may have new animals that have arrived and they won't eat that week because of stress, so our activity "feeding" in the database will be recorded as 0. The next week, they will eat, and will eat alot, so there will be data for that activity field. Given that there are variables and unaccounted data when comparing time spent on activities to their historical averages, is this method of time tracking really statistically accurate?

Can anyone cast their vote on this?

 

[jbwettergreen]

Well, both yes and no unfortunately. I completely get your situation (been there myself), and it is a bit **** of your employer to measure you in that way.

However, my assumption is that this is the best way form him/her to get some sort of information on performance/efficiency/due dilligence/call it whatever across his/hers employees. It may not be accurate, but i may serve as the best option nonetheless. As you mentioned, there are a lot of tanks, and a lot of activites, as long as you're more than five colleagues who do the work you describe, I get that your boss would want some form of measurement. How that information is used however, is a wholly another matter.

Is the variance in your variables (nature of your work, dependent on the situation, animals, and capacity) probably built into the historical averages? Whether the historical averages your are compared against captures the variance, i.e. they are calculated over a time-period where these fluctuations have taken place before would be the key question for me. I'm assuming that they probably are? There should always be some leniency towards using historical averages to measure performance, but as long as the averages are reasonably calculated, they should be guiding at least.

Hope this helps.