Just wanna say that this is a really cool survey, one I haven't seen before. I'm really curious to see the results! Two questions: when does the survey and and results are posted? And I know you said you'd be posting aggregate data, but does that apply to all wisdoms? (For example, would you post "X number of people said they multiaccount", or "X% of survey takers"?)
Hello! Thank you for the kind words. We haven't seen a survey like this either, and it seemed like a glaring gap in the community's knowledge.
When does the survey end?
The short answer is: The survey ends when very few new responses are coming in, or in two weeks, whichever is longest.
We expect "two weeks" to be longer, but we want to be prepared if it isn't.
When will results be posted?
I don't want to promise an exact date, but results will be posted as soon as we are able after the survey ends. It will take time for us to aggregate results and present them in a neat, coherent manner.
We may post aggregated results across several days, and then one final post summarising our findings and linking back to the previous ones. This is not a firm plan, but it is probably the most workable solution.
Will you be posting exact numbers in aggregate data?
Yes. In the interest of data integrity, and being as clear as we can, we intend to reveal the following information:
The number of respondents who answered the relevent question.
The percentage of those respondents who gave specific answers.
The value of that percentage - the exact number of respondents who gave the answer.
If any two of these data are visible, it will be easy to guess the third. Thus, to keep the value private, we would also have to hide the sample size of the question itself.
Failing to reveal your sample size is a bad idea, and it helps no-one. A sample size of 20 is significantly less representative of the wider population than a sample size of 2,000. If we do not reveal our sample size, we render all of our data unusable and leave its validity subject to question.
To make it clear to everyone, we mocked up some quick example data:
Fruit Preference recieved 79 responses.
Of those, 40 (50.6%) preferred apples; 13 (16.5%) preferred oranges; and 26 (32.9%) did not like either fruit.
We do not expect giving values in addition to percentages to harm anyone's privacy, as there is no way to know who answered the survey or who answered a given question. They are just numbers.
We will also be dealing with much larger numbers than in the example, which is beneficial for everyone.