Prompt: report a survey percentage without hiding the blanks
2 yes out of 3 answers is 66.7%; 2 yes out of 5 people is 40%. Get both figures with their fractions, and never let a blank answer quietly become a no.
Works with: ChatGPT, Claude, Copilot, Gemini
The task
Summarize a small internal survey honestly, with blanks and “unknown” answers counted separately and both percentages shown.
Summarize this fictional five-person survey without inventing responses. Question: Would you use the proposed shared FAQ? Valid answers are exactly Yes or No; blank and Unknown are not valid answers. Show Yes count, No count, blank count and Unknown count separately. Calculate Yes as a percentage of valid answers and as a percentage of all five records, showing each fraction and rounding to one decimal place. Explain why these are different measures. Make no inference about staff outside this fixture. Return a short brief plus the calculation table. S1: Yes S2: No S3: [blank] S4: Yes S5: Unknown
How to use it
- Run it with the five fictional records.
- Replace them with your own answers (no names), and keep the line about blanks and Unknown.
- When you report the result, give both percentages with their fractions.
What you get
Separate counts for Yes, No, blank and Unknown, both percentages with their fraction (2/3 and 2/5), and a short explanation of why they differ.
(excerpts) Yes: 2 (S1, S4) · No: 1 (S2) · Blank: 1 (S3) · Unknown: 1 (S5) · Valid answers: 3 | Yes as % of valid answers | 2 / 3 | 66.666…% | 66.7% | | Yes as % of all five records | 2 / 5 | 40% | 40.0% | The two figures share the same top number (2 Yes) but divide by different totals. The 26.7-point gap between them comes entirely from the two records with no valid answer. We can't tell whether S3 or S5 would have said Yes or No, so neither figure is the "true" rate. Report both, with their fractions, so the reader can see what each one counts.
Where should you start with AI?
The free assessment gives you, in 4 minutes, the 3 tasks to automate first in your business.