You've carefully designed your survey, distributed it to a large sample, and checked your email anxiously for two weeks. The response rate is 12%. This is a crisis for your research timeline and a very common one.
Why Response Rates Have Declined
Online survey fatigue is real. People are asked to complete surveys constantly — by companies, governments, universities, apps. Academic surveys compete for attention against all of these. The average response rate for unsolicited online surveys has been declining steadily for a decade. A rate of 15 to 30% is now considered typical for email-distributed surveys to unknown populations.
Design: Shorter Always Wins
Every additional question reduces your response rate. Ruthlessly remove every question that doesn't directly serve your research question. If you can reduce your survey from 30 to 20 minutes, you will significantly improve both your response rate and your data quality — fatigued respondents answer carelessly toward the end of long surveys.
The Subject Line Matters More Than You Think
For email distribution, the subject line determines whether your survey gets opened. "Research Study Participation" is invisible in a crowded inbox. Something specific and honest — "5-minute survey: burnout in nursing staff, [University Name] study" — performs better. Tell people immediately what they're agreeing to and why it matters.
Two Reminders, Not Three
A single follow-up reminder at four to seven days after initial contact typically generates 30 to 40% of your total responses. A second reminder at ten to fourteen days adds a smaller but meaningful increment. A third reminder rapidly approaches diminishing returns while risking annoyance. Two reminders is usually the optimum.
Incentives: When and How Much
Small token incentives (lottery entries, gift card draws) increase response rates in general population samples. They're less effective with professional samples who may interpret them as inappropriate. Charitable donations made in respondents' names are increasingly popular and have good evidence for effectiveness. Direct cash payments are most effective but expensive at scale.
When Your Response Rate Is Lower Than Target
If you end up with fewer responses than your power analysis required, you have options before concluding you've failed: conduct a sensitivity analysis to determine what effect size your actual sample can detect, report the response rate transparently as a limitation, and consider whether a non-response bias analysis is appropriate. Boss Statistics can help you assess whether your achieved sample is statistically sufficient for your analysis.