Printable View
Hi all,
I want to find the effectiveness of a team. I have two main factors to approach on it,
- Cognitive - Knowledge sharing(try to find the result by giving a simple question paper as a percentage)
- Demographic
Demographic contain three factors again, age, gender and tenure.
What I want to do is, workout a statistic model to find this.
Can someone help me to do this. If you need more clarification I can explain it here.
Thanks
There are a huge number of ways you could go about doing this from the general description you gave. One that I think is reasonable is below, but there are an infinite number of solutions to the problem you've posed (most of them bad, but still they exist).
In general you have to...
A. Determine ways to quantify (i.e. assign a number to) your two factors, congnitive and demographic.
B. Combine those two factors in a fashion of your choosing that fits reality well to get an overall "effectiveness" number.
My specifics are...
1. Make a test for cognitive as you seem to indicate, and assign the score as the cognitive factor, C.
2. For demographic, take the square root of age to keep high values from skewing things too much (a 70 year old isn't necessarily 3.5 times as "effective" as a 20 year old), and multiply that by some constant. Add to that a constant if they have tenure. Do nothing about gender (seems a bit sexist to include; I'm sure there's a statistical effect on gender, but I think it'd be very small. Maybe you meant some other effect from gender that I'm not seeing.)
2a. Your demographic factor D is then D = a1*Sqrt(age) + a2*(tenure?) + 0*(gender)
3. Combine the two factors as a simple linear combination for overall effectiveness E. This assumes that effectiveness scales linearly with both factors which is highly subjective. Note that scaling D by anything is the same as just scaling a1 and a2 by that thing, so we'll drop the factor in front of D. (The next step would have some redundancy if you kept it.)
E = a0*C + D = a0*C + a1*Sqrt(age) + a2*(tenure?)
4. Take a bunch of samples and assign what you believe are good effectiveness scores. Be sure the samples vary as much as possible. Do a fit in Excel or something to find which set of constants {a0, a1, a2} fit your definition of "effectiveness" the best. From then on, use those constants in the above formula for E to rate groups.
Again, there are a huge number of possible solutions to this. It's also incredibly subjective, but oh well.
In terms of gender, I think that you should include a variable that represents the balance of male-female members. The more balanced the team, the likely more effective it will be.