Quality-control dashboard workflow

This article shows how to organise reviewer-facing quality-control checks.

The focus is on transparent reporting rather than automatic exclusion.

Core QC domains

• file structure and import success;
• missing gaze or pupil rows;
• long gaps and interpolation burden;
• baseline coverage and stability;
• calibration or recalibration diagnostics;
• timing resets and signal activity;
• exclusion recommendations and review flags.

Example workflow

gap_qc <- audit_gazepoint_pupil_gaps(all_gaze)
baseline_qc <- audit_gazepoint_pupil_baseline(all_gaze)
reliability_qc <- audit_gazepoint_pupil_reliability(all_gaze)
luminance_qc <- audit_gazepoint_stimulus_luminance(stimulus_data)

signal_qc <- assess_gazepoint_signal_activity(all_gaze)
reset_qc <- assess_gazepoint_time_resets(all_gaze)

exclusions <- recommend_gazepoint_exclusions(
  pupil_qc = gap_qc,
  baseline_qc = baseline_qc
)

dashboard <- plot_gazepoint_biometric_report_dashboard(
  gap_qc = gap_qc,
  baseline_qc = baseline_qc,
  signal_qc = signal_qc
)

Reporting recommendation

Report QC results as descriptive diagnostics and decision support. Avoid presenting automated flags as definitive evidence that a participant or trial is invalid without a documented rule.