Trace each participant's change from first wave to latest on a persistent ID, surface notable trajectories the cohort average hides, flag missing waves, and grade the result green, amber, or red.
In short: Cohort averages hide the people who didn't move. To analyze longitudinal data honestly, trace each participant from their first wave to their latest on a persistent ID, surface the trajectories that stand out, flag anyone with a missing wave, and grade the result green, amber, or red.
Start where every participant carries the same ID across years. This walkthrough runs over DEMO-03 · Workforce Cohort — Vista Workforce Collaborative, multi-year survey waves with persistent contact IDs. Load your Decision Brief first so the assistant traces the metrics and population you actually care about.
You are the Sopact Sense Assistant working over the DEMO-03 · Workforce Cohort dataset (clean data + persistent contact IDs). Load my Decision Brief (decision, audience, outcomes, indicators, evidence standard) first, then wait for my task.
The prompt follows individuals, not the group mean.
For [POPULATION], trace each stakeholder's change on [METRICS] first→latest on persistent ID; per-ID change, distribution, notable trajectories. Grade green/amber/red.
Five elements keep it honest: the dataset with persistent IDs, the per-person on IDs tracing first→latest, the notable trajectories that the average flattens, the instruction to flag missing waves where a participant skipped a round, and the grade G/A/R on the strength of each trace.
Run on the Workforce Cohort dataset (DEMO 03) already loaded in Sopact Sense.
GRADE: green | Strong gain | clear first→latest improvement; amber | Flat East case | real no-change, unexplained; red | Missing-wave ID | latest wave absent
The grade tells you which traces to trust. Green means a clean trajectory — a participant with a clear gain from first to latest wave. Amber means a real but unexplained pattern — a flat East-site case that the cohort average would erase. Red means the trace is broken — an ID missing its latest wave, so no change can be computed for that person.
Fix the broken trace with the smallest realistic change.
Take the lowest-graded element above and fix it using only what the program could realistically measure. Show the before → after grade and the single indicator/edit that moves it to green.
Turn the per-person analysis into a branded "missing & incomplete" report and a shareable link.
Create a 'missing & incomplete' report from this analysis in Sopact branding [or paste your website URL / brand guideline to apply your own]. List every element graded amber or red, what is missing, and the one input that fixes each. Lead with the decision this report informs.
Create a shareable link for this report and open it in a new tab.
Cohort averages lie. A 10-point average gain can hide a third of participants who got worse. Per-stakeholder trajectories on persistent IDs don't.
Persistent IDs are the whole trick. Without a stable ID across waves you can't link a person to their own past self, and per-person change collapses back into an average.
A flat line is data. Don't discard participants who didn't move — they're often the most important story. Flag them and add a barrier question so the flatness is interpretable.
Treat a missing wave as red, not zero. An absent latest wave is unknown change, not no change. Mark the ID and chase the wave or exclude with a reason.
List every ID missing a wave, the wave that's absent, and whether to chase it or exclude with a stated reason.
Link each participant to a persistent ID, then trace their change on your chosen metrics from their first wave to their latest — reporting per-ID change, the distribution of those changes, and the trajectories that stand out. Flagging missing waves and grading each trace turns a multi-year dataset into individual stories instead of a single average.
An average can stay flat or rise while a meaningful subgroup declines, because gains and losses cancel out. Tracing change per person on a persistent ID preserves that variation, so you can see who improved, who stalled, and who regressed instead of one blended number.
Flag them rather than treating the gap as no change. A missing latest wave means change can't be computed for that person, so either follow up to recover the wave or exclude the ID with a stated reason — and report how many you excluded so the analysis stays transparent.
Open Sopact Sense, paste your program description, and put it to work.
Try in Sopact
