# Guidance on creating and organizing tags

Prompted by the "new tag suggestions" question, I wondered about

"guidance on tagging principles: should we aim for more generality, thereby associating each tag with more threads and enriching their semantic web, or aim for more specificity, which risks "Balkanizing" otherwise related ideas?"

I suspect general tags are better, because in appropriate combinations they can characterize any question precisely while linking it quite broadly with other questions. What is the community's opinion?

• What would be an example of "general" vs "specific"?
Sep 7 '10 at 19:37
• General: statistics, probability, modeling. Specific: computational-statistics, conditional-probability, random-effects-model. (All are actual examples.) Somewhat contrary to what I wrote in the question, it seems possible for some tags to be too general to be of much use, such as "statistics."
– whuber Mod
Sep 7 '10 at 22:57
• It seems there are several threads roughly related to this (and this is one) but I can't find a point blank answer to the question - when should you (or not) create a tag? A few minutes ago on this thread: stats.stackexchange.com/questions/27550/…, I created the z-test tag, because it seemed odd to me that it didn't exist. But, I didn't know if it was bad form to just create it out of nowhere since it's so similar to the t-test tag (which I could've used instead). Is there a link to a discussion on this or should I create a new question? Jun 8 '12 at 14:23
• @Macro The SE FAQ on tags may have what you are looking for. I found the emphasis on using broad tags to be especially interesting.
– whuber Mod
Jun 8 '12 at 14:28