At our clinic we have access to the WAI-S (Tracey & Kokotovic, 1989). To help us interpret the scores it would be helpful to know if there is some conventional/baseline levels for what can be considered a low / normal / high alliance score using this instrument. For instance, if someone have examples of mean scores from comparable settings (i.e. psychotherapy with adult clients), this would be useful.
Construct these low, medium, high you yourself. Here on RG you find: Working Alliance Inventory-Short Form Revised to measure therapeutic alliance
- February 2017
- Journal of Physiotherapy 63(2)
- DOI: 10.1016/j.jphys.2017.01.001
- License CC BY-NC-ND
- Davy Paap
- Pieter U Dijkstra
You can download their paper.
I can attach a paper so you can see how we did:
Start from your mean= medium group; mean + 1 SD = high, Mean - 1SD = low.
You can also use visual binning and divide the result in 3 equal groups: low, medium and high.
Hi Béatrice,
yes, creating "levels" based on our own gathered data as we go along can be one way to handle this. Thank you for your advice!
-Vidar
The alliance is scored perfectly 75% of the time, cut offs are rather meaningless, in addition, perfect alliance scores are not always associated with the best outcomes. You will find an alliance study on my profile. We categorise alliance as better, same, worse in our study
Strike that, we are in the middle of analysing 1 million alliance items on questionnaires and some of items are answered perfectly 90% of the time
In that case I see that cut off scores based on averages might have limited value. Thanks again!
No problem, the key is any change in direction. Cut off etc are not helpful for alliance
- Badges
- Science topic
- Similar topics
- Philosophy
- Ethics