Google describes its 1- or 2-question surveys as “microsurveys” and RIWI calls its 3- to 7-question surveys “Nano-Surveys™”. When Google Consumer Surveys recently rolled out a new 10-question format, it decided to call them “Full surveys”.
This seems like hyperbole, since a full survey should have 60 pages, 20 skip patterns, 50 items with conditional visibility, require an offshore quality assurance team to verify that it is working properly, and leave half its respondents in induced comas.
Which led to this Twitter exchange:
While poll and survey are pretty synonymous, I do believe that the frequent use of poll to refer to one-question intercepts on websites has led to the word typically being used to describe shorter questionnaires.
But I also believe that poll can be used to refer to any length questionnaire.
To prove or disprove these two hypotheses, I recorded the number of pages that Google has indexed for the exact phrases “1 question survey” and “1 question poll”, “2 question survey” and “2 question poll”, etc.
The only formulation that occurred more frequently for poll than survey was “1 question poll”, which occurred on 202,000 pages, as opposed to “1 question survey”, which occurred on 90,900 pages. The average length of a poll across these 30 observations is 5 questions, and the average length of a survey is 15 questions. So polls are shorter than surveys.
Since poll is used on 107 million pages, and survey is used on 154 million pages (69% as often), you might expect these counts of poll to show a similar relationship. But, in fact, the formulation “x question poll” occurs only 13% as often as “x question survey”, indicating that it is more typical to describe surveys in terms of their length than polls.
Now as for the other hypothesis, that poll can be used to describe long questionnaires, that is observed as well, though to a much lesser extent than similar length surveys: 18,200 web pages use the phrase “50 question poll”, compared to 456,000 pages using the phrase “50 question survey”.
There are fewer than 100 web pages each mentioning “40 question poll”, “60 question poll”, “70 question poll”, “80 question poll”, “90 question poll”, but an average of 98,000 pages for each corresponding formulation using the word survey.
The average across the 10 observations from 10 questions to 100 questions is 42 questions for survey, 19 questions for poll.
Would a survey by any other name smell as sweet?
No, I’m pretty sure that any excessively long survey or poll stinks.
Data & Methodological Note
The following data was recorded on February 6, 2014, from Google. These numbers change over time.
When doing analyses such as this, always enclose the phrase in quotes, otherwise you get every page that has all three words anywhere, as opposed to together: for instance, there are 718 million pages with the words “1”, “question” and “survey”, but only 90,900 pages for “1 question survey”.
Neither word microsurvey nor nanosurvey occurred frequently enough to be analyzed in similar fashion.
1 to 30 Questions
| survey | poll | |
| 1 question | 90,900 | 202,000 |
| 2 question | 89,700 | 32,400 |
| 3 question | 153,000 | 44,300 |
| 4 question | 219,000 | 44,900 |
| 5 question | 271,000 | 76,900 |
| 6 question | 89,400 | 482 |
| 7 question | 143,000 | 1,880 |
| 8 question | 73,400 | 18,100 |
| 9 question | 23,600 | 11,800 |
| 10 question | 164,000 | 81,700 |
| 11 question | 239,000 | 6,390 |
| 12 question | 73,000 | 5,030 |
| 13 question | 265,000 | 1,370 |
| 14 question | 290,000 | 320 |
| 15 question | 67,800 | 4,250 |
| 16 question | 314,000 | 1,060 |
| 17 question | 164,000 | 115 |
| 18 question | 165,000 | 154 |
| 19 question | 168,000 | 6,490 |
| 20 question | 122,000 | 7,380 |
| 21 question | 177,000 | 3,180 |
| 22 question | 174,000 | 88 |
| 23 question | 85,100 | 115 |
| 24 question | 110,000 | 204 |
| 25 question | 160,000 | 17,900 |
| 26 question | 90,700 | 6 |
| 27 question | 10,500 | 9 |
| 28 question | 73,000 | 62 |
| 29 question | 35,300 | 30 |
| 30 question | 267,000 | 933 |
10 to 100 questions
| survey | poll | |
| 10 question | 164,000 | 81,700 |
| 20 question | 122,000 | 7,380 |
| 30 question | 267,000 | 933 |
| 40 question | 272,000 | 88 |
| 50 question | 456,000 | 18,200 |
| 60 question | 43,700 | 45 |
| 70 question | 45,000 | 10 |
| 80 question | 108,000 | 7 |
| 90 question | 19,900 | 4 |
| 100 question | 4,330 | 2,380 |
Author Notes:
Jeffrey Henning
