Combination factor
Description of the combination factors for which study results included in the calculation of job effect.
The job effect is calculated based on the study results relevant for the particular search. The algorithm has a built-in combination factor ascribing weight to the study result based on how closely it resembles the search. Below you can see the values taken into account.
Combination factor by intervention
|
Combination factor |
Weights |
|---|---|
|
No combination factor |
1 |
|
Combination factor in either target group or intervention |
0.5 |
If “Interviews” is selected for instance, all study results about interviews will be weighted with 1. The combination factor also plays a role with study results where target group or intervention is composed of several components. This would be the case if you were to select the intervention category “Interviews”, with a study result containing both Interviews and Business-oriented programmes. Here the estimated effect would not be solely attributed to Interviews, and the combination factor on 0.5 would apply.
Multiple results from a study: weighting main results and partial results
Every unique study may contain several study results. The number of study results per study can vary from 1 to 15 study results. To ensure that the studies enter the calculation of the stock of knowledge with approximately equal meanings, the number of study results is assigned a weight, including whether the study result is part of the search as a main or partial result. This weighting is instrumental for placing an upper boundary to how much a single study can contribute to the total stock of knowledge.
|
|
Weights for main target group results |
Weights for partial target group results |
Weights |
|---|---|---|---|
|
Main target group results |
1 |
|
1 |
|
Partial target group results |
|
0.5 |
0.5 |
|
Both main target group results and partial target group results |
0.67 |
0.33 |
1 |
Results aligning with the searched main target group are ascribed a weight of 1. If a unique study contributes with for instance 10 study results, those 10 study results will share the weight 1. This means that each study result will effectively have a weight of 0.1.
If study results are found in the job effect search as partial target group results, the main target group results will take a weight of 0.67, whereas the partial target group results will have a weight of 0.33. Study results containing solely partial target group results in relation to the job effect search take a weight of 0.5.