Talk:Linguistic Analysis of Candidates

This is an interesting concept. I see a potential difficulty, though. There are certain concepts which I really don't think you can omit entirely when evaluating conservatism (the right to life, for example, and religious freedom.) However, the trick is in determining which key phrases are most likely to be employed, and thus turn up in searches. After all, a mention of "abortion" in a speech isn't a reliable indicator of conservatism--a rabidly pro-abortion candidate might use the word, as well.

How best to evaluate conservative leanings on these very important issues? --Benp 21:22, 12 January 2009 (EST)

I don't fully follow your criticism. "Pro-life" is part of the test, for example. Perhaps you'd like to look at the first analysis and reconsider or rephrase your criticism. Thanks.--Andy Schlafly 21:48, 12 January 2009 (EST)

Not so much a criticism as musing on the most efficient way to go about it, Andy. I don't think anything you've already included needs to change; I just wonder if other terms are also warranted. Pro-life is there (and I did, indeed, overlook that the first time; thank you for pointing it out!) Pro-life is certainly a good term, and one that's only likely to be used by conservatives, but is it sufficient? Would it be worthwhile to include terms like "right to life" as well? What about terms that accurately describe the abortion industry, but which liberals avoid like the plague: "pro-abortion," for instance, or "abortion on demand?"

I suppose what I'm trying to figure out here is where the optimal balance falls between being thorough and getting bogged down in too many redundant terms. Rather than add terms wholesale, I thought I'd seek guidance here first. --Benp 22:20, 12 January 2009 (EST)

If time were unlimited, the list could grow in an unlimited manner. But then priority is lost. Twenty (20) seems to balance breadth with priority, and I'd be reluctant to expand the list. But feel free to make specific suggestions here. So far, I don't see how to improve on the term "pro-life". The terms "pro-abortion" and "abortion on demand" seem redundant at best. Overlap between terms is to be avoided.--Andy Schlafly 22:55, 12 January 2009 (EST)

I do see your point. I guess the only question I still have, then, is about religious freedom and school prayer; I don't see any terms on the list that directly touch on these issues, which I would certainly consider important to most conservatives. --Benp 16:35, 13 January 2009 (EST)

Shouldn't this be an essay? KevinS 21:50, 12 January 2009 (EST)

No, I think the concept is encyclopedic and susceptible to objective analysis. Encyclopedias include analysis of novels, so why not linguistic analysis of candidates?--Andy Schlafly 22:55, 12 January 2009 (EST)

As I understand it, though, such analysis comes from other sources. It's not a huge issue, anyways. KevinS 16:45, 13 January 2009 (EST)

To test the effectiveness of this approach we should compare the results for known conservatives such as Palin and McCain with known Liberals such as Obama. If this test is effective it should give very different results. If I have time at the weekend I'll put something together. AndyJM 11:17, 16 January 2009 (EST)

Please do, but note that no one ever felt McCain was a "known conservative," and Palin, though to the right of McCain, remains largely unknown in her political views.--Andy Schlafly 11:26, 16 January 2009 (EST)

Excellent. Do you have any people I could test this on or should I go ahead with McCain, Palin and Obama? Anyone whose speeches are easily available on the internet would do. AndyJM 11:29, 16 January 2009 (EST)

Jeb Bush is considered more conservative than McCain, and many congressmen (e.g., Steve King) are more conservative than both.--Andy Schlafly 11:30, 16 January 2009 (EST)

I'd like to suggest a supplement to this collection that might also produce interesting results. There exist uncomplicated techniques in the field of document analysis for determining similarity ratings between two bodies of text. Some work much like the work being done here, but more comprehensively, and controlling for the total number of words a speaker uses. (So a speaker that says twice as many words won't look twice as conservative.) This is very close to my personal area of expertise, so it would be very quick and easy for me to perform this sort of analysis. All it would take is full transcripts of the speeches to be analyzed, and some sort of "ideal" conservative speech, which could be as simple as a list of key words, which can be any length without increasing my workload, and can even be tailored (for example, there could be a "strong on defense" list, a "social conservative" list, and an "economic conservative" list. If someone wants to prepare such lists of key words as well as links to relevant speeches or groups of speeches to be compared, I would be happy to help. (Note that I'm basically obligated to say that the numbers I spit out are not true measures of conservatism, just a measure the extent to which the speeches analyzed use similar words to an ideal.) (Also, "phonics"?) - DaveB7 17:01, 4 July 2009 (EDT)

Dave, I'd love to work with you on this. This entry has some test words, but they are based on much earlier and more primitive form of Essay:Best New Conservatives Words. I can come up with a stronger list now. How many key words would you like, and in what form, and how many speeches?--Andy Schlafly 18:15, 4 July 2009 (EDT)