Psychometric validation

With any new translation or adaptation it is necessary to carry out psychometric analyses and we ask that you carry these out and send us the results of any psychometric analyses conducted. If funding is available and if the sample size is sufficient (>100 or 5x the number of respondents as there are items in the questionnaire, whichever is the greater), we will be glad to conduct the psychometric analyses on your data set.

For users, these analyses are important for justifing summing items into subscales and/or overall scales.  For the original authors of the scales, a record of the psychometric evaluation is important for any future developments, particularly in the unlikely event that the analysis may indicate further linguistic review to be necessary. Psychometric validation is also important for any subsequent users of the questionnaire.

The following are the basic requirements we believe are necessary for those questionnaires that have subscales (e.g. DTSQ, WBQ, ADDQoL):

  1. Confirmatory factor analysis  on items 1,4,5,6,7 and 8 in the questionnaire for DTSQ and all items of the WBQ12, and for all domain-specific items in the ADDQoL (and related measures)
  2. Reliability analysis with Cronbach’s alpha conducted separately for each subscale.

Please let us have hard copy printouts and/or electronic copies of the factor structure and reliability analysis.  We use SPSS software and can read documents in Excel and Word, but do not have SAS software.

Please note that for questionnaires containing items that are scored separately and not combined into scales-subscales (e.g. DCSQ and ADKnowl), there is no necessity for such psychometric evaluation.

Further information relating to scoring, replacing missing scores etc can be found in the User Guidelines for the questionnaire.