The Collection of Risk Behavioral Data: Applying Palmton (PASI) as a New Data Collection Method

Abstract

The objectives of this paper were to enhance and assess completeness of using PASI (Palmtop-Assisted Self-Interview) for the collection of sensitive self-reports data on behavior risk, compared with ACASI (Audio-Computer-Assisted Self-Interview), SAQ (Self-Administered Questionnaire) and FFI (Face-to-Face Interview) methodologies. Urine samples were collected and tested for chemical markers of recent drug use to assess self-under-reporting in each of the four methods.

Data collection was conducted during November 2002 among 1,283 students (640 males and 643 females) aged 15 to 21 years old at two vocational schools in Chiang Rai province. Student  were recruited at the classroom-based sessions during which trained nurse explained the study. Followed consent, students were randomized equally into one of the four interview methods (ACASI, PAS, FFI, and SAQ) to answer the same questionnaire. The interview consisted of questions regarding socio-demographics characteristics, sexual and drug use behaviors, dietary habits, depressive symptoms and suicidal ideation.

Generally, it was found that computer-based methods were more accurate, complete and had the less deviation when comparing with urine test result. PASI and ACASI were also more effective reduced time and errors in the data entry and data monitoring process. The computer-based interview technology has become available to apply for data collecting regarding sensitive questions. Moreover, studies using computer have shown high participation rates, likely as a result of the perceived privacy of data collection. Overall, PASI and ACASI produce higher rates of negative answers regarding behaviors whereas FFI produced higher rates of positive answer.

PASI was found to be non-inferior to ACASI but provided a cheaper and was more highly mobile. This method, when applied to an appropriate study, becomes an alternative for collecting sensitive behavioral data.