Check conditions after your actions
Follow up selected work conditions and outcomes after your actions have had time to take effect. Choose enough questions to assess those changes with workers; the right length depends on what you need to check.
Prepare a pulse
- 1. Factors and questions
Use the model-selected list below, your actions and worker feedback to choose what to measure. A factor can require several questions.
For Prior Work factors, copy their items from the survey question template (.csv). For another instrument, use its original questions and scoring.
- 2. Collect responses
Keep wording and scoring consistent with your baseline. Plan who to invite across the teams and work areas you need to understand.
In a small workplace, consider inviting everyone. A shorter questionnaire reduces time to respond; it does not by itself reduce the number of responses needed.
- 3. Upload and compare
Code one row per respondent using the loaded model's factor names and state codes. Keep missing answers blank.
Review the column preview before running the comparison. Distinguish measured answers from estimates for unmeasured factors. Check the coding guidance.
Choose a useful time
Run a pulse when your actions could reasonably have changed conditions. A routine check at 6–12 months is one option. Review sooner if risks or work change; do not wait for a survey date.
Check what the baseline represents
A previous workplace survey can provide a measured baseline. A sector model is a reference, not evidence of change since a survey of your workplace. For an imported model, check its source.
Plan invitations and interpret differences
A narrow sample can make results look better or worse than the wider workforce. Plan recruitment for the comparisons you need and check whose experience is missing. Keep confidentiality arrangements clear.
Use worker feedback and other evidence alongside the pulse. Differences may reflect who responded or other workplace changes. Check each factor's desired direction: stronger support can help, while more overload can harm.
Updating model probabilities needs valid answers for a factor and its required inputs in the same response. Measuring a factor alone may leave its starting probabilities unchanged; the result shows the usable counts.
No model loaded
Pulse analysis requires a fitted model. Load a sector model in Data Analytics, or upload and fit your survey data, or import an existing fitted model.
Open Data Analytics