Method
Response rate: why low participation ruins the conclusion
22 January 2026 · 5 min read
Short answer
A low response rate is not a smaller dataset but a skewed one: the most dissatisfied and the most pressured are least likely to answer. Below roughly 60 percent, results should be read as indications, and units with fewer than five responses should not be reported at all.
The skew is systematic
Two groups do not answer: those with no time, and those who do not believe it makes a difference. Both are overrepresented where the problems are. So a survey with a low response rate typically comes out too flattering, not merely too small.
What actually raises the response rate
01
Short format
Under ten minutes, no free text, no long scale batteries. Length is the single factor that costs the most answers.
02
Credible anonymity
Promising anonymity is not enough. It must be technically explainable why the link does not exist — and the threshold must be visible in the report.
03
Access without email
Employees without regular screen access will answer if they can redeem a printed one-time code. Otherwise whole groups drop out.
04
Reminders with cadence
Day 3, 7 and 10 typically capture the most. More than three reminders move little and annoy much.
05
The last round led to something
The strongest driver is that people remember a concrete change from last time. Response rate is in practice a measure of trust in the process.
What not to do
- Produce person-level lists of who has not answered. It destroys anonymity and with it data quality.
- Lower the threshold in order to report small units. The threshold is what makes the answers honest.
- Compare response rates between units as a management KPI. It turns the survey into a competition in participation rather than honesty.
Read on
Want to know which value actually wins in your organisation?