Why this matters in MENA
In the markets this platform serves, mobile is not a segment - it is the majority. Roughly two-thirds of responses arrive from phones, and the share is higher for younger demographics. If mobile responses were systematically worse, most regional online research would be built on sand. Some researchers still down-weight or exclude mobile completes on exactly that suspicion. We tested it.
Setup
210,000 responses across 640 studies were segmented by user-agent into mobile and desktop. We compared completion time, open-text length, straight-lining rates, attention-check pass rates, and final composite quality scores, controlling for survey length and question mix. Studies with fewer than 50 responses per device class were excluded.
Findings
| Metric | Mobile | Desktop | Significant? |
| Median completion time | 4m 12s | 5m 23s | Yes |
| Median open-text length | 38 chars | 43 chars | Yes |
| Straight-lining rate | 4.1% | 3.8% | No |
| Attention-check pass rate | 94.6% | 95.1% | No |
| Mean quality score (0-100) | 78.2 | 79.0 | No |
Interpretation
Mobile responses look different on the surface - faster, terser - but the differences live in style, not substance. Speed on mobile reflects the interface (tapping beats mousing), not carelessness: mobile respondents pass attention checks and avoid straight-lining at the same rates as desktop respondents. Shorter open-text reflects typing cost; when we scored text for informativeness rather than length, the gap narrowed to insignificance.
“Mobile is not worse - it is just shorter and faster.”
- Author's note
Recommendations
- Do not down-weight, exclude, or separately quota mobile responses on quality grounds - the data does not support it
- Do design for thumbs: shorter prompts, larger tap targets, and 5-point rather than 7-point scales that fit a phone screen without scrolling
- Set speed thresholds relative to device-specific baselines, or mobile respondents will be unfairly flagged by desktop-calibrated timing
- Preview every survey on a phone before launch - if a grid question requires horizontal scrolling, rebuild it as separate items
Scale-length guidance for small screens is covered in our Likert scale guide.