The Importance of Proper Analysis
The best survey is useless if results aren't properly analyzed. Analysis is the critical translation step from raw data to concrete action recommendations.
Avoiding Common Mistakes
- Interpreting results without context
- Only looking at average values
- Overvaluing individual outliers
- Not using comparison values
- Not discussing results with those affected
Understanding Key Metrics
Averages
The average shows typical response behavior:
- Good for a first overview
- But: Can hide extreme values
- Example: Average of 3.0 can consist of all 3s or of 1s and 5s
Standard Deviation / Spread
The spread shows how uniform responses are:
- Low spread: Uniform perception in the team
- High spread: Different experiences – look closer!
Distributions
How are responses distributed across the scale?
- What percentage respond critically (e.g., "disagree")?
- Are there clusters at the extremes?
- Is the distribution symmetric or skewed?
Response Rate
The participation rate affects how much the results carry, but there is no normative cut-off. What matters more than a fixed percentage is how the responses are distributed and what is usual in your setting:
- Context rather than cut-off: what counts as a good response rate depends on the setting. In public administration, staff surveys typically reach somewhere between a quarter and a third of the workforce; clearly more than half already counts as a very good result there. In small, well-informed teams, higher rates are normal.
- Check the distribution: a moderate overall response is fine as long as every activity group returns enough answers. A high overall response does not help if one department is almost entirely missing.
- Name the bias: if a group is systematically missing, record that in the documentation and evaluate that area separately rather than discarding the figures.
Benchmarking: Contextualizing Results
Your results alone don't tell you if they're good or bad. Comparison values are crucial:
Types of Benchmarks
- External benchmarks: Comparison with other companies, industries, nationwide
- Internal benchmarks: Comparison between departments, locations
- Time-based benchmarks: Comparison with previous surveys
Interpretation
- Better than benchmark: Strength – but don't get complacent
- In benchmark range: Average – check improvement potential
- Worse than benchmark: Action needed – develop measures
Identifying Critical Areas
Traffic Light Logic
Many tools use a traffic light system:
- Green: Uncritical, better than benchmark
- Yellow: Attention required, in benchmark range
- Red: Critical, significantly worse than benchmark
Prioritization
Don't tackle everything at once. Prioritize by:
- Severity: How strong is the deviation?
- Impact: How many employees are affected?
- Feasibility: What can realistically be changed?
- Leverage: What has the biggest positive effect?
Deepening Results
Quantitative survey results show the "what" but not always the "why." To deepen understanding:
Results Workshops
- Discuss results with affected teams
- Ask about backgrounds
- Analyze causes together
- Collect measure ideas
Focus Groups
- In-depth conversations with selected employees
- Gain qualitative insights
- Test hypotheses
Manager Conversations
- Discuss results with supervisors
- Include context information
- Gain commitment for measures
Communicating Results
Create Transparency
- Inform all employees about overall results
- Name positive and critical areas
- Announce next steps
Area-Specific Feedback
- Managers receive results for their area
- Observe minimum group size for anonymity
- Offer support for interpretation
Conclusion
Analysis is the key to your risk assessment success. Take time to interpret results properly and involve those affected. SafeMind provides automatic benchmark comparisons and action recommendations – making analysis a breeze.
