Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
350999 | Computers in Human Behavior | 2011 | 9 Pages |
Internet data collection is becoming increasingly popular in all research fields dealing with human perceptions, behaviors and opinions. Advantages of internet data collection, when compared to the traditional paper-and-pencil format, include reduced costs, automatic database creation, and the absence of researcher-related bias effects, such as availability and complete anonymity. However, the validity and reliability of internet gathered data must be established, in comparison to the usual paper-and-pencil accepted formats, before an inferential analysis can be done. In this study, we compared questionnaire data gathered from the internet with that from the traditional paper-and-pencil in a sample of college students. The questionnaires used were the Maslach Burnout Inventory – Student Survey (MBI-SS), the Oldenburg Burnout Inventory (OBI-SS) and the Copenhagen Burnout Inventory (CBI-SS). Data was gathered through a within-subject cross randomized and counterbalanced design, on both internet and paper-and-pencil formats. The results showed no interference in the application order, and a good reliability for both formats. However, concordance between answers was generally higher in the paper-and-pencil format than on the internet. The factorial structure was invariant in the three burnout inventories. Data gathered in this study supports the Internet as a convenient, user-friendly, comfortable and secure data gathering method which does not affect the accepted factorial structures existent in the paper format of the three burnout inventories used.
► Psychometric properties of data on student burnout gathered on the internet vs. paper-and-pencil. ► Questionnaires used were the MBI-SS, OLBI-SS and CBI-SS. ► Data gathered through a within-subject cross randomized and counterbalanced design. ► No interference in application order, good reliability and factorial invariance for both formats. ► It’s feasible to use parallel versions of burnout inventories in both medial representations.