I want to launch surveys on very specific binary questions with low "Yes" answer rate, for example Do you have a Ferrari ?, Did you buy a Chanel bag ?
I'm able to do that on at least 20 different websites from different categories, like a finance forum or gaming platform.
+---------+-------+------+
| Website | % Yes | % No |
+---------+-------+------+
| A | 15% | 85% |
| B | 10% | 90% |
| C | 7% | 93% |
+---------+-------+------+
For each website i have a set of keywords describing the website with a score.
+---------+----------+-------+
| Website | Keyword | Score |
+---------+----------+-------+
| A | Finance | 0,9 |
| A | Business | 0,6 |
| C | Gaming | 0,8 |
+---------+----------+-------+
My end goal is to be able to score the keywords for the people which answered yes. On the example tables above, i would like to output a sort of ranking which will give the keyword finance above the keyword gaming.
My questions are :
Is there a sort of ranking methodology or formula i can use ?
How can i estimate the minimum sample size ? i think this question can be split in two questions
- How can i determine the minimum number of answer for each website to have a relevant percentage of yes and no
- How can i determine the minimum number of different websites to diffuse the survey to have relevant keyword output
Thank you
