Javai
Fledgling Freddie
- Joined
- Jan 8, 2004
- Messages
- 1,531
Comments specifically on the current poll and statistical validity. Polls are only as good as their question wording and their response rate. In this case the poll fails to deliver anything you could act on as it fails on both these tasks.
1. It lumps all moderators into one category when several people have commented in the thread that they are happy with some moderators but find some overstep the mark. It would be wrong to have a poll on each moderator named, if you feel a moderator oversteps the mark PM'ing Jupitus is th ebest way to go then he can see if he gets a string of complaints about a particular mod.
2. The poll asks if the moderators are up to the task - if doesn't ask if the task itself is properly defined. So you could be unhappy with the rules but think the moderators enforce them perfectly fairly. The poll as it is currently worded doesn't invite constructive discussion of a ruleset to make the forum a better place it invites attacks on the mods.
3. Response rates. You can only infer from a sample if it is unbiased, in this case there is strong reason to suggest that those who do not reply are different in their opinions than those who do - any inference from a minority response in such a situation is potentially misleading
1. It lumps all moderators into one category when several people have commented in the thread that they are happy with some moderators but find some overstep the mark. It would be wrong to have a poll on each moderator named, if you feel a moderator oversteps the mark PM'ing Jupitus is th ebest way to go then he can see if he gets a string of complaints about a particular mod.
2. The poll asks if the moderators are up to the task - if doesn't ask if the task itself is properly defined. So you could be unhappy with the rules but think the moderators enforce them perfectly fairly. The poll as it is currently worded doesn't invite constructive discussion of a ruleset to make the forum a better place it invites attacks on the mods.
3. Response rates. You can only infer from a sample if it is unbiased, in this case there is strong reason to suggest that those who do not reply are different in their opinions than those who do - any inference from a minority response in such a situation is potentially misleading