Timeline for Figure out the number of samples needed to increase the precision of an estimate
Current License: CC BY-SA 4.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 3, 2022 at 15:45 | comment | added | user371955 | Could you potentially explain how I could go about doing this assuming a 'worst-case' distribution. As mentioned assuming I have only 4 samples per review and I want to infer whether 4 is enough (has low variance) or I need more samples. Maybe when I fit a line through the function I find that 10 samples is actually what I need. | |
| Nov 3, 2022 at 15:15 | comment | added | user371955 | I am sampling from actual data. So for example in my actual dataset I have 3 people max for each review. I sample from these 3 with replacement for N = 1, 2 and 3. I then fit a line to the function (variance vs. N). I read this plot and infer that I need an N of ~10. Does this make sense? | |
| S Nov 3, 2022 at 9:51 | review | First answers | |||
| Nov 3, 2022 at 9:57 | |||||
| S Nov 3, 2022 at 9:51 | history | answered | Jannis | CC BY-SA 4.0 |