Retest reliability of self-reported daily drinking: form 90.

Area of Science:

• Addiction research
• Psychometrics
• Behavioral science

Background:

• Historically, alcohol consumption was viewed as binary (drinking vs. abstinence).
• Contemporary understanding recognizes drinking as a dynamic process with fluctuations over time.
• Modeling these dynamics requires longitudinal, high-cadence data collection.

Purpose of the Study:

• To evaluate the retest reliability of daily drinking reports from the Form 90.
• To expand reliability estimates beyond averaged data to daily-level assessments.
• To determine if daily self-reports are suitable for capturing drinking oscillations.

Main Methods:

• Secondary analysis of existing data from 70 treatment-seeking adults.
• Data collected using the Form 90, with participants re-interviewed by different interviewers after 2 days.
• A common 88-day recall period was used for both interviews.

Main Results:

• Kappa statistic and intraclass correlation were used to assess agreement.
• Daily kappa values ranged from .481 to .971.
• The average kappa across the recall period was .766, indicating substantial agreement.

Conclusions:

• Self-reported daily drinking demonstrates excellent retest reliability.
• The Form 90 is a reliable tool for collecting daily drinking data.
• These findings support the use of Form 90 data for modeling drinking patterns at a daily cadence.