Talk:FDR (disambiguation)

Requested move 29 June 2019

FDR (disambiguation) → FDR – I'd like to see if there would be consensus in the community for moving this disambiguation page to the base name FDR, which at the moment is just a redirect to Franklin D. Roosevelt. I feel that there are other uses that are no less important than Roosevelt's initials. Whereas Roosevelt was no doubt a very important American president from a hundred years ago, False discovery rate (FDR) is a key concept in modern science. And FDR is also an acronym for several other important things - just go through the list in the disambiguation page and you'll see. This is why I believe the disambiguation page should be at FDR. I was wondering what you guys think and if there'd be consensus to make this move. Thanks, Dr. Vogel (talk) 20:30, 29 June 2019 (UTC)

• Oppose various historians refer to Roosevelt more than anything else by this initialism. Searches on Google and Google Books are full of results on him. SNUGGUMS (talk / edits) 21:07, 29 June 2019 (UTC)

Absolutely. Definitely many results about Mr. Roosevelt. But a quick search on Web of Science gives 38,269 papers about false discovery rate. And these are scientific, peer-reviewed papers, published in scientific journals. Dr. Vogel (talk) 21:19, 29 June 2019 (UTC)

* Support per False Discovery Rate. --Comment by Selfie City (talk about my contributions) 21:58, 29 June 2019 (UTC)

• Hah, if you were citing the contents of the article with those initials as your reason for your vote, that's pretty clever. Paintspot Infez (talk) 23:50, 29 June 2019 (UTC)

• Oppose: The Franklin D. Roosevelt article is viewed about 11,000 times per day, about 25 times as much as False discovery rate. That's too lopsided. It is clear that Wikipedia readers are much more interested in Roosevelt, and he also has high historical importance. —BarrelProof (talk) 23:24, 29 June 2019 (UTC)

But how much of that traffic arrives via the redirect? Dr. Vogel (talk) 23:29, 29 June 2019 (UTC)

• Despite my cleverness, per Google Ngram I must oppose this nomination. --Comment by Selfie City (talk about my contributions) 00:09, 30 June 2019 (UTC)
• Comment: the question is not whether Roosevelt is a more popular topic than all the other meanings of FDR combined (which is probably true) but whether nearly 100% of people who type FDR are looking for Roosevelt. Obviously more people know about and search about Roosevelt than about false discovery rate, simply because there are more people who have heard of world events than people who know science.

To put this in terms of conditional probability, the question is not whether:

$P(Roosevelt)>>P(otherthingsthathavethesameabbreviation)$

but whether:

$P(Roosevelt\mid FDR)>>P(notRoosevelt\mid FDR)$,

which would be a very strong claim, and according to the Sagan standard, extraordinary claims require extraordinary evidence.

If there isn't sufficient evidence to support that claim,

there isn't sufficient evidence to keep the disambiguation page not at FDR. Dr. Vogel (talk) 00:45, 30 June 2019 (UTC)

Per Bayes Rule these probabilities are related to the prior distribution (how much interest is there in Roosevelt in general? How much interest is there in other things that can be referred to as FDR?) and the conditionals $P(FDR\mid Roosevelt)$, $P(FDR\mid FalseDiscoveryRate)$, etc. We've established that the prior on Roosevelt is high. I think a reader is reasonably likely to use the search "FDR" when looking for information about Roosevelt, given that the full name is pretty long to type, the surname alone is ambiguous, and the initialism is well-known. What probability does that translate to? My handwavey order-of-magnitude guess would be 10-20%. Do you think the initials "FDR" are so commonly used as to refer to False Discovery Rate that someone looking for information about FDRate is much more likely to use the initials than someone looking for FDRoosevelt? Colin M (talk) 03:22, 4 July 2019 (UTC)

• Support – terms this ambiguous should be disambiguated, not redirected to one topic. See Dr. Vogel's analysis, which I agree correctly describes the current situation as extraordinary. Dicklyon (talk) 05:04, 30 June 2019 (UTC)
• Oppose – When I type "FDR" into the Search Wikipedia box, the drop-down menu gives no pointers to anything related to False Discovery Rate. Instead, I'm given FDR Drive, FDR (video game), FDR Skatepark, and FDR in Trinidad, all of which refer to the president. The FDR moniker in that context is alive and well. Dhtwiki (talk) 13:28, 30 June 2019 (UTC)

That's because those articles have titles that begin with FDR. Dr. Vogel (talk) 13:41, 30 June 2019 (UTC)

That those titles only have to say "FDR" and be understood as to what they're referring is the point that is made. Dhtwiki (talk) 15:04, 30 June 2019 (UTC)

Yes, that is true, but it doesn't constitute sufficient unbiased evidence to support the very strong claim that $P(Roosevelt\mid FDR)>>P(notRoosevelt\mid FDR)$

• Oppose. The redirect has been stable for quite some time and nothing in the above discussion indicates the president is not the primary topic for FDR. Calidum 22:38, 30 June 2019 (UTC)
• Oppose – the American president is primary topic for the base name FDR. A hatnote can suffice to redirect to the scientific term. CookieMonster755✉ 06:36, 1 July 2019 (UTC)
• Thanks and question - Thanks everybody for participating in this, even if we disagree :)

Could any of the people voting oppose please address the fact that no unbiased and sufficient evidence has been given to support the very strong claim that $P(Roosevelt\mid FDR)>>P(notRoosevelt\mid FDR)$ ? Extraordinary claims require extraordinary evidence.

The reason I'm asking this is because, ultimately, that is the rationale for displacing the dab page with lots of terms, some of which are important, from its natural location. Thanks! Dr. Vogel (talk) 08:38, 1 July 2019 (UTC)

The WP:BURDEN should be on you to prove that your proposition is not true, rather than ours. Plenty of evidence has been presented to indicate that having the president considered the primary topic is not an extraordinary one. The propositional notation you use seems in error as a summary of the text at WP:PRIMARYTOPIC, which says A topic is primary for a term with respect to usage if it is highly likely—much more likely than any other single topic, and more likely than all the other topics combined—to be the topic sought when a reader searches for that term. Your "much-greater-than" notation (i.e. >>) should be merely ">" for the combined alternatives, which is what I take $P(notRoosevelt\mid FDR)$ to mean. I'll not attempt a similar proposition for the "much more likely than any other single topic" part of the PRIMARYTOPIC text, as I'm not sure the propositional calculus (for some value of the term) is a good choice for clarifying arguments here. Dhtwiki (talk) 17:08, 1 July 2019 (UTC)

I feel that it's the exact opposite. Displacing a disambiguation page to a secondary location just because of 1 meaning is more extreme than allowing the disambiguation page to sit at the primary location. I think the WP:BURDEN should be on whoever wants to favour a more extreme option over a less extreme option. Dr. Vogel (talk) 18:39, 1 July 2019 (UTC)

Take a look at the nGram I linked to, earlier in this thread. --Comment by Selfie City (talk about my contributions) 19:14, 1 July 2019 (UTC)

I did. But that just shows that the president is a more popular topic than the statistical method. It says nothing about the relative probabilities of what is meant by FDR. And it also doesn't address the fact that this is about the president (1 meaning) vs. all the other meanings (many). Dr. Vogel (talk) 19:24, 1 July 2019 (UTC)

If you mention another topic you think gets a lot of hits, I can add it to the Ngram. --Comment by Selfie City (talk about my contributions) 03:27, 4 July 2019 (UTC)

• Oppose, FDR is one of those letter combinations which have a clear primary topic, the former American president. Randy Kryn (talk) 20:08, 1 July 2019 (UTC)

I agree that's the general perception. But do we actually have any evidence for that claim? Dr. Vogel (talk) 22:36, 1 July 2019 (UTC)

Pageviews for the past 90 days show the FDR redirect averaging 47 hits per day and FDR (disambiguation) averaging 21 hits per day. Even assuming all hits on the dab page arrive via the hatnote on Franklin D. Roosevelt (unlikely), fewer than half of readers using FDR apparently want something other than the president. If we further assume all readers of the dab page want False discovery rate (also unlikely), fewer than 5% of readers of False discovery rate arrive at that article via FDR. Now that False discovery rate has been added directly to the president's hatnote, even those relative few can get there with one fewer click. Station1 (talk) 23:05, 1 July 2019 (UTC)

It's not clear what available data would be appropriate here. The ideal thing would be to know how many people search for "FDR", and of those people, what proportion click the hatnote to the dab page, and go to a different topic like False Discovery Rate. But afaik that data is not publicly available. In cases like this, books, ngrams, and page view stats give us some circumstantial evidence, but in large part we have to rely on our intuition to close the gap. If the first thing I think of when I see "FDR" is Roosevelt, and a bunch of other people here say the same thing, then it seems pretty safe to say the average user will probably not be astonished to land at the Roosevelt article if they search FDR. (Yes, there are some issues with treating RM participants as representative of Wiki readers as a whole, but we work with what we have.) Colin M (talk) 03:33, 4 July 2019 (UTC)

• Oppose Per WP:PRIMARYREDIRECT._{ZXCVBNM (TALK)} 14:15, 3 July 2019 (UTC)