8
$\begingroup$

Our tag (220 questions) is ambiguous and seems to be used inconsistently: it can refer either to Pearson's correlation or to Pearson's chi-squared goodness of fit test.

I think it should be split into [pearson-r] and [pearson-chi-squared].

Afterwards we should map [pearson-chi-squared] (and [pearson-goodness-of-fit]?) to (1100).

$\endgroup$
15
  • $\begingroup$ I didn't get what the difference between pearson-chi-squared & [pearson-goodness-of-fit] is supposed to be? Is the first the test for independence/homogeneity? And then why create two tags only to map them both to one? $\endgroup$ Commented Feb 19, 2016 at 12:22
  • 1
    $\begingroup$ I had a cursory look at the [pearson] tag to see what it is used for apart from the obvious correlation usage. I see that sometimes people write smth like "I am doing Pearson's goodness of fit test" and sometimes "I have problems with Pearson's chi squared test". Of course it is the same thing (that's why I suggest to map both to [chi-squared]) but I thought that perhaps some people would not know that and so might be confused about which tag to use when they start typing "pearson..." in the tag window. But if we think that this is uncommon then it's enough to have [pearson-chi-squared]. $\endgroup$
    – amoeba
    Commented Feb 19, 2016 at 12:57
  • 2
    $\begingroup$ We could re-tag existing questions with just one of the two, say pearson-chi-squared, but still map both to the synonym chi-squared. $\endgroup$ Commented Feb 19, 2016 at 13:35
  • 2
    $\begingroup$ Yes, this makes sense. In fact, if one is re-tagging anyway then one can directly retag into [chi-squared]. But we would need synonyms for future convenience anyway. And what do you think about [pearson-r] vs. [correlation]? I think most of the [correlation] questions refer to Pearson's correlation, so it is a bit strange to have 10-times-less-used [pearson-r] separately. On the other hand, "correlation" as a term can certainly refer to more things than Pearson's correlation. $\endgroup$
    – amoeba
    Commented Feb 19, 2016 at 13:48
  • $\begingroup$ Meanwhile, I edited the Q based on your suggestion. $\endgroup$
    – amoeba
    Commented Feb 19, 2016 at 14:33
  • $\begingroup$ I just tried to create a synonym suggestion for [pcs] & [pgof], but it won't let me. It says the tags have to exist first. Do you think we should add them onto a Q so that we can make the synonym suggestions, or just drop the issue? $\endgroup$ Commented Feb 12, 2017 at 14:33
  • $\begingroup$ @gung Yes, one needs to create them first (put on any suitable Q). I still think it's worth doing, at least for [pcs]. $\endgroup$
    – amoeba
    Commented Feb 12, 2017 at 15:26
  • 1
    $\begingroup$ OK, I just created the 2 tags (here, & here), & made the synonym suggestions here. $\endgroup$ Commented Feb 14, 2017 at 14:09
  • $\begingroup$ @gung Cool, I voted myself and have pinged a couple of people to vote. $\endgroup$
    – amoeba
    Commented Feb 15, 2017 at 9:47
  • $\begingroup$ @Scortchi It's been some time, but I followed up on that. Take a look at the update to my answer here. I think [pearson] tag is ready to be renamed [pearson-r]. Can you help here? $\endgroup$
    – amoeba
    Commented Feb 22, 2017 at 9:36
  • 1
    $\begingroup$ @amoeba: Thanks! Done. $\endgroup$ Commented Feb 22, 2017 at 12:54
  • $\begingroup$ @Scortchi Great, thanks! Could you also take a look at the similar situation with [permutation]: meta.stats.stackexchange.com/a/4461 $\endgroup$
    – amoeba
    Commented Feb 22, 2017 at 13:00
  • 1
    $\begingroup$ @amoeba: Done too. $\endgroup$ Commented Feb 22, 2017 at 13:32
  • $\begingroup$ @Scortchi Thanks! I noticed that there was a shuffle->permutation synonym that is now pointing permutation-test. I am not sure it's helpful (?). Perhaps we should rather delete this synonym and see if [shuffle] reappears?.. $\endgroup$
    – amoeba
    Commented Feb 22, 2017 at 13:34
  • 1
    $\begingroup$ @amoeba: I've removed it. $\endgroup$ Commented Feb 22, 2017 at 13:36

1 Answer 1

4
$\begingroup$

Here is my specific plan about what to do with the tag, based on the above discussion with @Scortchi a year ago:

  1. Create tags and tag synonyms $\to$ and $\to$. Update: DONE.
  2. Go through all [pearson] threads and retag those about chi-squared into . This will be a minority of [pearson] threads, so not very difficult. Update: DONE.

    Specifically, I looked at all questions that:

    • had "chi" in them; or
    • had "goodness" in them; or
    • had "residual" or "residuals" in them; or
    • did not have "correlation" in them.


    I retagged a couple of dozens, the remaining 263 must be all about Pearson's $r$.

  3. Rename into via a mod merge hammer: make a synonym, merge, delete the synonym mapping. (This is needed to remove the ambiguity of the [pearson] tag name). Update: DONE by @Scortchi.

  4. Profit.
$\endgroup$
2
  • $\begingroup$ Good job with the tag work. After you started participating in Meta, discussions/issues related to tags evolved a lot! $\endgroup$ Commented Feb 22, 2017 at 17:57
  • 1
    $\begingroup$ @AndreSilva Thanks. To tell the truth, this takes a lot of time and effort. I don't think I will be doing this indefinitely, but I would like to try to push forward all the tag issues that have been pending for some time. By the way, I am happy to see that some people joined our efforts in tidying up the tags, in particular mdewey. $\endgroup$
    – amoeba
    Commented Feb 22, 2017 at 19:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .