Remove this ad

#21 [url]

Hey there, Alpha Dog! Always a pleasure to see a friendly riposte! I hope you're enjoying this discussion as much as I am. 

Now, you stated, "Even though the scope of this discussion morphed over the past few days and we delved into some highly-subjective gray areas, my original thesis remains true. In case you missed it, I used a three-part form of deductive reasoning called a syllogism to prove my point." I actually didn't miss it, although as I am not a trained logician by any stretch of the imagination I will admit to not jumping up and down and saying, "That's a syllogism! Look, there are his major and minor premises! Look at his conclusion! EXPERTLY DONE, good sir!" However, it did make me wonder, because I didn't recall you making the initial point quite so bluntly.

So... I went through and re-read everything that you've said so far -- your initial board question and your posts here on the Board Comment Board (completely off-topic, THANK YOU to the Board for creating this forum, since such a discussion would be hard to do via corrections). Your initial question was statistical in nature. As the discussion has progressed, you have increasingly been trying to argue this in a logical manner -- or at least, that is the impression I got from reading your work. Both statistics and logic are generally unemotional, so let's call this approach the Logical Plane (although, personally, I call it Vulcan). The Board writers and most other commenters on the BCB tend to argue either based off of emotions and experiences (which we can call the Emotional Plane) or via a two-pronged approach trying to straddle both planes at once.

Let's look at your initial question, shall we? I apologize in advance for potentially using the incorrect terminology; as I said before, I am not a trained logician. I have taken some statistics classes and generally scored well in them, but I am not a statistics major and it has been a while, so my stats vocab might also be a bit off. As I analyze what you said, please feel free to correct me if you believe I have in any way, shape, or form mischaracterized your words.

Also, I know you didn't present it as a statistical claim, but since that is essentially what it is, it makes sense to look at it in that way. So I'll refer to things as data points and evidence and variables, even though you and the writers do not explicitly do so.

Also, I apologize to women everywhere for referring to them as "data points" in this discussion.

Alpha Dog's Initial Question

You made the following observations:

  • You noticed an exponential trend between a woman's dates and her physical attractiveness (claim). Because this is an observation about a trend, you are entering the realm of statistics.
    • Data point 1: You have a friend who is physically attractive (cute face, hourglass figure) who averaged three dates a week and had a horde of men chasing after her. You mention that this was over her "whole time at BYU", but don't give a specific timeframe. I will assume four years, as that is generally the expected duration of an undergraduate program, but please correct me if that is not accurate. The variables here are physical attractiveness, number of dates, number of men, and timeframe.
    • Evidence 2:  You have another friend who is not quite as attractive as the aforementioned friend but who averaged one date per month and is now married. You don't say how many men were interested in her, but "not a ton of action" suggests to me (given the location) that you're trying to say "not many men." No timeframe is given, but your wording gives me the impression that she is married while at BYU. Again, please correct me if that is not accurate. The variables here are physical attractiveness, number of dates, number of men, and timeframe​.
    • Evidence 3: You have another friend who, in your own words, "isn't that good-looking at all." (Alta was referring to this when she said, "...and then you go behind their backs and call them ugly?" The connection between "ugly" and "isn't that good-looking at all" seems linguistically sound to me, but later you dispute this.) She was at BYU for five years and didn't get asked out once (which would also imply not many men). The variables here are physical attractiveness, number of dates, number of men, and timeframe​.
    • Hypotheses:
      • Alternate hypothesis (and the one you actually imply here): The more attractive a woman is, the more likely she is to go on a date. 
      • Null hypothesis: Physical attractiveness does not influence a woman's number of dates.
      • The dependent variable would be "number of dates", with the remaining three being the independent variables. I think that you could probably find a strong correlation between the number of men and the number of dates, so you might want to dismiss the "number of men" variable as describing the same phenomenon as "number of dates" to make your results more accurate. (In fact, after your first data point you seem to do so.)
      • Your sample size is three.
    • You also mentioned some variables that could potentially interfere with your results (extraversion and flirtatiousness). You dismiss these as outliers, while Alta suggests that these could be control variables instead.
While you never explicitly state it in statistics jargon, you reject the null hypothesis that physical attractiveness has no influence on the number of dates that a woman will be asked on. The idea in statistics is that for something to be statistically significant, you have to be able to reject the null with 95% confidence (or even 90% in some cases). Here you can reject the null three out of three times, so with 100% confidence.

I think you also realized (consciously or not) the limitations of a sample size of three, so you asked a number of other female YSAs, including a fellow BYU alumna (excellent use of Latin grammatical gender, by the way, my figurative hat is off to you!). They all introduced another variable -- Man is LDS (y/n). You don't explicitly control for this in your initial data points, but one could assume that the vast majority of the men doing the asking at BYU are LDS. Still, best practice would be to explicitly control for this dummy variable.

You also make the claim (unsupported anywhere else in your submission, which makes me think you didn't expect it to be questioned) that "to a great extent, physical beauty is a choice."

You then ask two questions: "Have any of you noticed this phenomenon?" (Asking for writers' experiences and anecdotal evidence.) "What do you think causes it?" (Asking for opinions, which may or may not be backed up by evidence.)

Writer Responses on the Board

To recap, you posted a statistical argument and then asked for experiences/anecdotal evidence and opinions on the cause.

Sheebs responded with her experience (anecdotal evidence) that when she was "conventionally very attractive", she "dated much less". However, she does point out that there was a confounding variable of mental health. She then rebuts your second claim that physical beauty is a choice by saying that "I hope you realize that you are talking about many things you don't understand." In conclusion, she submits a data point (herself) that doesn't agree with your statistical argument and questions your unsupported claim. Because her data point does not agree with your conclusion, she does not proffer an opinion to support your conclusion.

Of importance here is that this data point means you have only rejected the null three out of four times (75%), so you can no longer reject the null with 95% confidence. The relationship is thus no longer statistically significant.

Alta challenged the validity of your variables by saying that being outgoing and flirtatious are not actually outliers but key variables. You should have controlled for them. She does think that the LDS variable is important but points out that the sample size is too small to make an accurate generalization. She does not provide another data point, but does provide opinions on the cause for the LDS variable in particular, which is what you asked for. She also challenged the validity of your research methods by asking how you were able to gather the information and suggested another potentially confounding variable (your own attitudes).

Of importance here is that she challenged the methods used to calculate your statistical relationship. In order to continue, you should control for extraversion, flirtatiousness, and your own attitudes while increasing your sample size in a transparent manner.

Auto Surf challenges the representativeness of your sample, and explicitly says, "I'm not saying you're wrong, just that your research methods are lacking, so I would be careful to make conclusions." She mentions a few other potentially confounding variables: age, location, and male confidence. None of this is controlled for in your experiment.

In order to continue, you should try to control for these variables.

Luciana challenges the validity of the variable "attractiveness" by questioning both the scale and how it is calculated. She rejects your unsupported claim that beauty is a choice and also the implication that it is a choice that women should be making. She also disputes using women as data points (I'm sorry, Luciana, in this discussion I am guilty of that as well).

While she doesn't provide any additional data points, she does agree that there might very well be a statistical relationship between attractiveness and the number of dates. She questions the limitation to BYU and the implication that men are shallow, but instead suggests that confidence should be the real dependent variable and that attractiveness is a contributing variable to confidence.

In order to continue, you should try running the experiment with confidence as the dependent variable and see if there is a stronger relationship, as well as analyzing whether or not your attractiveness variable is correctly measured.

Did you get what you asked for?

You asked for experiences/anecdotal evidence ("Have any of you noticed this phenomenon?") and opinions about the cause ("What do you think causes it?"). While you did not get both from each writer, you did get both. Sheebs provided experiences and Luciana and Alta provided partial opinions on the cause. You also got a healthy dose of feedback regarding your statistical model.

Now... is this what you really wanted? I don't know. I sent my mind-reader back for a refund because it kept jumping to faulty conclusions.

Where could you have gone from here? If this were a stats class or a research project, you would have been well advised to incorporate the feedback from the writers and make a stronger project. However, since this is (most likely) neither, you would also be well within your rights to just let it go. You would not, however, be justified in assuming that your model is still indicative of a statistically significant relationship.

According to "The Skeptic's Guide to the Universe":
Also it is important to note that an argument may use wrong information, or faulty logic to reach a conclusion that happens to be true. An invalid or unsound argument does not necessarily prove the conclusion false. Demonstrating that an argument is not valid or not sound, however, removes it as support for the truth of the conclusion – it means that the conclusion is not necessarily true.
...
The premise may or may not be true, but it has not been established sufficiently to serve as a premise for an argument.

I submit to you, sir, with all due respect, that because you failed to address the concerns of the writers in their initial response to your question, we cannot know that your arguments are valid, even if the conclusion may be accurate. 
 

Quote    Reply   
Remove this ad
Add Reply

Quick Reply

bbcode help