Eliminate the effect of severity of the Personal Outcomes Scale: Linear regression in persons with intellectual disability

Abstract

Background: The Personal Outcomes Scale (POS) is used to assess quality of life (QoL) in people with intellectual disability (ID) but the results are infl uenced by the severity of the disability. To address this issue, we present the standardization of the Spanish adaptation of the POS. One of the limitations of the Classical Test Theory is the differential effect in some items due to the effect of an external variable. For this reason, we propose the use of multiple linear regressions. Method: We used a sample of 529 people with ID, along with 522 professionals and 462 parents who were administered the POS in order to carry out an in-depth psychometric study. For the analysis, we used the correction strategy by multiple linear regressions to generate centiles corrected for the severity effect of ID. Results: Following this technique, the results show the centiles from the raw POS score corrected for the effect of the severity of ID. Conclusions: This standardization technique is a feasible option to provide a QoL scale without biased results due to the severity of ID.