Model v5 methodology · revision 5.0.2
What PsychPod measures, in plain language.
PsychPod helps you reflect on changes in your own self-reported wellbeing. It is research-informed, but the combined PsychPod score is not a clinically validated questionnaire, diagnosis, treatment, or crisis prediction.
Your state today
Your answers are converted to a 0 to 100 score for sleep, mood, energy balance, calm, focus, connection, and positive moments. Higher means that area felt more supported in this check-in. Context answers stay separate.
Change from your usual
After enough check-ins, PsychPod compares a gently smoothed recent state with a slow personal baseline. This comparison always uses the same seven questions shared by quick and full check-ins. It uses your own history.
Confidence in the trend
This describes data quality only. It considers distinct check-in days, the time covered, recent coverage, and regularity. It does not express confidence in a diagnosis or prediction.
Quick and full reflections
The quick check-in asks eight questions: one representative item for each scored area plus a direct support preference. The full reflection collects more detail and any habits you chose to track. Missing questions are not treated as zero.
Today's score uses all valid scored answers in that check-in. History charts, period summaries and personal comparisons use only the seven shared questions, so changing check-in mode does not change the ingredients of a trend. This comparison score can differ from a full reflection's score. A missing shared answer stays missing; an overall comparison requires all seven.
The score averages available core areas equally. Sleep and energy use fixed product rules informed by research, but their exact curves and the combined score have not been validated against future behaviour. A score of 70 is an index value, not a 70% chance of an action or an illness.
How smoothing works
A gap-aware moving average reduces the chance that one unusual day dominates the longer view. A slower baseline starts from a robust centre and changes gradually. The main score always shows the latest check-in. The personal comparison uses the longer view and needs at least 14 distinct days across a span of at least 21 days. These are display rules, not evidence that a prediction is reliable. Calculations preserve the recorded local day when available, including after travel, so opening the app later does not create another observation.
What patterns can tell you
History summaries describe the scores you reported and their spread during that period. Next-day association claims are currently withheld: repeated observations and testing several possible patterns can make chance relationships look convincing. Future behaviour prediction will need clearly defined actions, consented outcomes and evaluation on later observations that were unavailable when the prediction was made. No behaviour predictor is currently used to generate your score.
Support choices
You can directly say that support would be useful. PsychPod may also show support resources after a conservative combination of difficult answers. This routing rule can miss distress or show support when it is not needed. PsychPod does not monitor you and cannot contact emergency services.
Open support resources without signing in