Abstract
<p>Objective: This methods and data article operationalizes qualitative multimedia content analysis (QMCA) for short-form video posts on TikTok and Instagram Reels, focusing on the method's execution and the frequency structure of observable grooming-precursor cues masked as relationship advice.Method: A purposive corpus of 100 public short-form videos, evenly split between TikTok and Instagram, was archived, de-identified, and coded using MAXQDA 24 for time-segment coding and Excel for the codebook, audit trail, and thresholds. A five-pass routine separated visual features, audio features, on-screen text, and relational/boundary cues. Category development relied on same-document co-occurrence, with recurrence requiring five or more distinct posts.Results: The final codebook contained 289 codes and produced 1,233 post-level feature hits. Recurrence concentrated in presentation-layer categories: Parasocial Proximity and Platform Normalization appeared in all 100 posts, and Intimate Setting Cues in 80. Covert precursor categories recurred less often: Power Imbalance Normalization in 30 posts, Coercive Reframing in 15, Boundary Erosion in 14, and Isolation and Secrecy in 10.Conclusions: QMCA provides an auditable method for documenting how short-form advice content packages relational scripts through camera, voice, setting, captioning, and platform conventions. The taxonomy supports prevention-oriented screening without inferring creator intent or labeling any post as grooming.</p>