Supplementary Data for "Digital Classification of Behavioral Stages in mHealth Applications: Validation of a Rule-Based Adaptive Staging Algorithm"

Published: 16 November 2026| Version 1 | DOI: 10.17632/7f7wt2yf2y.1
Contributor:

Description

This dataset accompanies a validation study comparing a rule-based adaptive staging algorithm with a conventional static questionnaire for assessing Transtheoretical Model (TTM) stages of change for physical activity. In TTM-based mHealth apps, the stage determines which content a user receives and must be reassessed over time. The study examined whether presenting the staging items adaptively (one item per screen with rule-based branching) yields the same stage classification as the static questionnaire, whether it reduces completion time and the number of answered items, and how participants rate its usability. Data collection: In a randomized within-subject crossover design, participants completed both staging formats in an anonymous German-language online survey (LimeSurvey) in randomized order (sequence A: adaptive first; sequence B: conventional first). Both formats used the same six yes/no staging items and the same decision rules for stage assignment. After each format, perceived usability was rated with the three-item After Scenario Questionnaire (ASQ; 1 = strongly agree/best to 7 = strongly disagree). Completion times come from the survey page timers. Participation was voluntary, and informed consent was given electronically. Contents: One Excel file with two sheets. "Data" contains one row per participant who completed both formats (n = 54; 33 in sequence A, 21 in sequence B). Incomplete responses and a test entry were excluded; timestamps, randomization seed, language, and open-ended responses were removed. Variables prefixed ada_ refer to the adaptive format and con_ to the conventional format: responses to items Q1–Q6, derived stage, number of answered items, completion time in seconds, the three ASQ items, and the ASQ overall score (mean of the three items). Empty item cells mean the item was not presented on the branching path (adaptive format) or was left unanswered (conventional format). "Codebook" describes every variable, its values, and the English item wording. Main findings: Of 53 participants classifiable in both formats, 45 were assigned the same stage by both formats (84.9%; unweighted Cohen's kappa = 0.688). All 54 participants were classifiable with the adaptive format, compared with 53 with the conventional format. Mean completion time was 20.21 s (adaptive) vs. 47.53 s (conventional), and the mean number of answered items was 2.61 vs. 5.46. Mean ASQ overall scores were 1.25 (adaptive) and 1.60 (conventional). Use: The data allow the analyses reported in the associated article to be reproduced; stages can be re-derived from the item responses using the decision rules described there. Because the sample is small and unevenly distributed across stages, the data are best suited to methodological replication and secondary analyses rather than population estimates.

Files

Categories

Health Psychology, Physical Activity, Psychological Assessment, Human-Computer Interaction, Mobile Health, Digital Health

Licence