Is there a God?

[cheesehead]

Originally Posted by Zeta-Flux

Is it more or less charitable to suppose, a priori, that you were being flippant?

You didn't have to suppose I was being flippant, but doing so would have been considerably more charitable that supposing that I'd committed a straw man fallacy.

There were alternative suppositions that also would have been more charitable, such as supposing that you needed to ask for clarification rather than make an accusation of fallacy.

As I hadn't read your comment as being flippant, how might I have interpreted your meaning? etc...

How? As something other than a logical fallacy, or as something not clear enough to react to without clarification.

[cheesehead]

Originally Posted by AES

In the county I currently call home, churches are the only establishments allowed by law to serve alcohol (wine) before noon on Sundays. It's a good thing... that the Methodists practice open Communion.

?? When I attended a Methodist church (1950s, '60s) in the U.S., the communion liquid was always nonalcoholic grape juice. Did you mean some other denomination, are Methodists in your country "off the wagon", or is it just that they keep the grape juice unrefrigerated a bit too long before serving?

[Zeta-Flux]

Originally Posted by cheesehead

You didn't have to suppose I was being flippant, but doing so would have been considerably more charitable that supposing that I'd committed a straw man fallacy.

Then we differ in opinion. I believe that supposing someone is committing a fallacy is much more charitable than supposing something (negative) about their internal emotions. Fallacies are *easy* to commit, and one shouldn't be looked down on for falling into one of their traps.

There were alternative suppositions that also would have been more charitable, such as supposing that you needed to ask for clarification rather than make an accusation of fallacy.

At what point is it appropriate, in your view, to state that you believe one is committing a fallacy, rather than to ask for clarification?

How? As something other than a logical fallacy, or as something not clear enough to react to without clarification.

Suppose that I mistakenly believed I understood the intent. What are some possible readings in that case?

[cheesehead]

Originally Posted by Zeta-Flux

Then we differ in opinion. I believe that supposing someone is committing a fallacy is much more charitable than supposing something (negative) about their internal emotions. Fallacies are *easy* to commit, and one shouldn't be looked down on for falling into one of their traps.

But let's look closer at what the "straw man" fallacy is.

It's an exaggeration -- a purported portrayal of someone else's idea, but distorted to seem unreasonable.

Even if that's a subconscious deception, I consider it non-innocent, because it's motivated by the desire to distort, not a desire to be honest. The portrayer of a "straw man" is not quoting or paraphrasing someone else's words as accurately, or even as neutrally, as possible. I don't view that charitably.

Perhaps it is easy to fall into the subconscious habit of committing "straw men", especially if one grows up among folks who habitually use it, but it's still, at heart, no more excusable than habitual easy lying or habitual easy stealing, is it? I can view the reasons for development of a habit charitably in such a case, but not the continuation of that habit once it is challenged for what it is.

That it's easy to be dishonest shouldn't cause us to excuse dishonesty, or consider it charitably, should it? (There may be other reasons, such as its intent, to consider dishonesty charitably in certain circumstances -- but not its ease.)

At what point is it appropriate, in your view, to state that you believe one is committing a fallacy, rather than to ask for clarification?

When you believe it, of course. (But we were previously concerned with charitableness, not appropriateness.)

One can learn to modulate ones threshold of believability in this regard, so as to learn to more often ask oneself whether one correctly interprets a suspected fallacy.

Suppose that I mistakenly believed I understood the intent. What are some possible readings in that case?

Whatever your possible readings are, of course. (But we were previously concerned with whether a reading was charitable, not whether it was mistaken, understood or possible.)

[Zeta-Flux]

Originally Posted by cheesehead

But let's look closer at what the "straw man" fallacy is.

It's an exaggeration -- a purported portrayal of someone else's idea, but distorted to seem unreasonable.

I disagree with this characterization of the "straw man" fallacy. One only needs to misrepresent the position. No exaggeration is necessary.

Even if that's a subconscious deception, I consider it non-innocent, because it's motivated by the desire to distort, not a desire to be honest. The portrayer of a "straw man" is not quoting or paraphrasing someone else's words as accurately, or even as neutrally, as possible. I don't view that charitably.

Perhaps it is easy to fall into the subconscious habit of committing "straw men", especially if one grows up among folks who habitually use it, but it's still, at heart, no more excusable than habitual easy lying or habitual easy stealing, is it? I can view the reasons for development of a habit charitably in such a case, but not the continuation of that habit once it is challenged for what it is.

That it's easy to be dishonest shouldn't cause us to excuse dishonesty, or consider it charitably, should it? (There may be other reasons, such as its intent, to consider dishonesty charitably in certain circumstances -- but not its ease.)

What if the misrepresentation was not conscious nor subconscious, but merely mistaken-ness? Would you still consider that non-innocent and dishonest? If so our opinions differ.

If you believe one must consciously, or subconsciously, distort the other's statement to commit the "straw man" fallacy, then we differ on the definition.

When you believe it, of course. (But we were previously concerned with charitableness, not appropriateness.)

One can learn to modulate ones threshold of believability in this regard, so as to learn to more often ask oneself whether one correctly interprets a suspected fallacy.

To rephrase my question: at what point is it charitable (rather than just appropriate) to state that you believe one is committing a fallacy, rather than to ask for clarification?

Whatever your possible readings are, of course. (But we were previously concerned with whether a reading was charitable, not whether it was mistaken, understood or possible.)

Let me clarify then. What are the possible charitable readings, given than I misunderstood the intent of your statement, but believed I completely understood the intent? Or is that impossible in your view, and a charitable reading would tell me that I'm probably wrong, and I should always ask for clarification?

P.S. I'd appreciate if you could answer the questions in the quiz, to get us a little back on topic. Thanks.

[AES]

Originally Posted by cheesehead

?? When I attended a Methodist church (1950s, '60s) in the U.S., the communion liquid was always nonalcoholic grape juice. Did you mean some other denomination, are Methodists in your country "off the wagon", or is it just that they keep the grape juice unrefrigerated a bit too long before serving?

Well, they do recognize one another when they meet in a liquor store.

[cheesehead]

Originally Posted by Zeta-Flux

Okay. I reconsidered. Here is a re-written quiz.

Part C:

Q C1: What is the total sample size of prosperous nations? (I.e. what is the total number of nations under consideration)

Answer from last time: 17 or 18.

In paragraph [10] of the article, 17 are described as "developed democracies", and Portugal is described as a "second world democracy", with no mention of "prosperous" in either case. So I conclude that all 18 are included in the title's characterization: "prosperous democracies".

Q C2: What is the cut-off line for prosperity?

Answer: Unknown.

The way the phrase "prosperous democracies" is used in the article suggests to me that there is a specific definition for that term in the social science fields for which the article is intended. That is, to a professional in that area, the term has a well-defined meaning. Likewise, the use of phrases "developed democracy" and "second world democracy" suggest to me that those also have well-defined meanings to those in that field. But I could be wrong, and I don't know the formal definition anyway (a modest Google search didn't help).

Q C3: How does the sample size change if we vary the cut-off line (i.e. how robust is the choice)? For example, if we change prosperity by $5,000 does the number of nations change dramatically?

Due to your answer to C2, I think you will have to answer "unknown." But then again, maybe it is still a poorly worded question from your point-of-view.

Yes, unknown.

It's not so much that the question is poorly worded as that there may be a formal definition of "prosperous democracy" that does not involve a particular dollar figure.

Your $5,000 figure seems to implicitly assume that the prosperity in question is a per-capita or per-household average rather than a national GNP figure. However, one could replace the $5,000 by an equivalent figure proportional to GNP in that case, so that's of no consequence.

But why should we change the prosperity cutoff at all? Your "robust" reason doesn't really apply here. You'd just be restricting the data to a different range of prosperity, not discarding truly anomalous data points. Perhaps it would show that the correlation coefficients differed significantly between prosperity ranges, but it wouldn't test the robustness of the coefficients over any particular range.

Q C4: How many of the nations in the sample are over 50% believers? How many are under?

There are five measures of religiosity. I presume that by "believers", you are referring to the "absolutely believe in God" measure. From the upper left chart in Figure 1, it looks like two are right at the 50% mark, and one is definitely above it. The others are definitely under.

However, there is a problem with eyeballing the article's charts without having the corresponding actual figures: Not all letters are shown on all graphs. Perhaps when two or more nations would be plotted at the same position on a graph, only one of their letters can be shown. For example, the graph I just mentioned shows only 11 letters, not 18. So we don't know where seven of the nations would be on the graph. That makes it 2 at 50%, 1 above 50%, 8 under 50%, and 7 unknown.

The upper left graph of Figure 2 seems to show 17 letters. Two are at 50%, two are above 50%, 13 are under 50%, and one is unknown -- better than in Figure 1. Figure 3's upper left shows 17 also, with 1 at 50%, 1 very closely under 50%, 1 above 50%, 14 under 50%, and 1 unknown. Figures 4 and 5 may show all 18, with the counts same as Figure 3 except 2 above 50% and none unknown.

Q C5: If we drop the two nations which are the most extreme in terms of being believers/non-believers, does it change the answers to C4 much?

I presume that by "the two nations which are the most extreme in terms of being believers/non-believers", you are referring to Japan and United States, judging by the same charts I used in answering C4.

I question the motive for dropping the "most extreme" nations at all. The reason for dropping outliers in experimental data is that they are likely to be caused by errors in measurement that did not affect the other data points - statistical error. Furthermore, these dropping-the-outlier cases are cases in which all the measurements are of the same phenomenon or event or item, so there's no inherent reason to think they'd differ if all the measurements had been infinitely accurate and precise. That's not the case here. There really are differences between the populaces of different nations.

You've introduced no evidence that the data points you want to drop are the result of measurement error. In this case, the differing numbers for different nations correspond to different realities, not different measurements of the same thing. Sure, there's some error range for each nation's number, but what you're proposing is to drop outliers with no regard to their distance from the other data points relative to error-of-measurement range, but just because of their values.

It still seems to me that the only reasons you have for proposing to drop outliers in this case are that they don't fit your preferred result, and that you're mimicking the dropping of statistical outliers without having the reason for dropping that that practice requires. (Again, you've presented no evidence at all that the measurements for those nations are any less accurate than the measurements for other nations, so it can't be because of measurement error.) Sorry for the harsh words, but I've seen no reason why they are wrong. Your proposals have never mentioned any reason except for their outlying status, and that's not a good enough reason to drop them. You've never mentioned any reason to think that the numbers you want to drop are inaccurate.

There's a famous case in science that illustrates bias toward an expected result. I'll quote from Richard Feynman's famous "Cargo Cult Science" speech at http://www.lhup.edu/~DSIMANEK/cargocul.htm

We have learned a lot from experience about how to handle some of the ways we fool ourselves. One example: Millikan measured the charge on an electron by an experiment with falling oil drops, and got an answer which we now know not to be quite right. It's a little bit off, because he had the incorrect value for the viscosity of air. It's interesting to look at the history of measurements of the charge of the electron, after Millikan. If you plot them as a function of time, you find that one is a little bigger than Millikan's, and the next one's a little bit bigger than that, and the next one's a little bit bigger than that, until finally they settle down to a number which is higher.

Why didn't they discover that the new number was higher right away? It's a thing that scientists are ashamed of--this history--because it's apparent that people did things like this: When they got a number that was too high above Millikan's, they thought something must be wrong--and they would look for and find a reason why something might be wrong. When they got a number closer to Millikan's value they didn't look so hard. And so they eliminated the numbers that were too far off, and did other things like that. We've learned those tricks nowadays, and now we don't have that kind of a disease.

BTW, there's a separate controversy about whether Millikan's own results showed bias toward an expected result in his own discarding of many measurements -- especially since he claimed not to have discarded any measurements (but his lab notebooks show that he did). Some have explained that what he did had no significant effect on his published result for the amount of an electron's charge (values he dropped were both below and above the average), but only resulted in claiming a smaller experimental error range than was warranted.

The later discovery that Millikan had a systematic error in his results because, as noted by Feynman, he used an incorrect value for air viscosity is a separate matter from bias-toward-expected-result.

Q C6: What is the correlation coefficient between homicide rate and religiosity?

I presume that would be calculated from the data points in the five graphs of Figure 2.

I could try measuring the positions of the plotted data points, but there'd still be the question of which method of correlation to use.

Just by eyeballing, it looks like the correlation coefficient in the upper left graph of Figure 2 would be positive; for the upper right graph, positive but less than for the upper left; for the lower right graph, positive; for the lower middle graph, negative; for the lower right graph, negative.

Q C7: Is this a strong positive correlation?

After noting that the sign of the coefficient in the lower middle and right graphs differs for the same relative sense of religiosity, I'd say ... define "strong" and give me the actual data numbers (not literally -- I know you don't have them) so I can calculate whether they meet your definition of "strong".

Q C8: How robust is this correlation? In other words, if we throw away the two extremes (say, with respect to religiosity), does the correlation change much?

See C5 for my objection to throwing away the two extremes, in the absence of any reason to think that they're any less accurate than the other data points, or of any reason to think that they'd actually be related to robustness.

[cheesehead]

Originally Posted by cheesehead

It still seems to me that the only reasons you have for proposing to drop outliers in this case are that they don't fit your preferred result, and that you're mimicking the dropping of statistical outliers without having the reason for dropping that that practice requires. (Again, you've presented no evidence at all that the measurements for those nations are any less accurate than the measurements for other nations, so it can't be because of measurement error.) Sorry for the harsh words, but I've seen no reason why they are wrong.

Zeta-Flux,

I invite you to present good reasons for dropping the outliers that you want to drop, to replace my harsh speculations. (You've always been free to do so, but I want to make it explicit.)

Your repeated proposals to drop certain data points, but without giving any reason why their quality differs from any other data point, look ... odd. Why are the data points you want to drop inferior to any other data points in the article?

You've criticized this article for unscientific use of statistics, yet you repeatedly propose your own unscientific use of data points. Why this seeming contradiction? What is the reason?

Perhaps I've been uncharitable in my interpretation of your proposals to drop data points, but unless/until you present good reasons for carrying out your proposal, it's difficult to arrive at any more-charitable interpretation than that you want to shape the data to fit your own preconceptions.

You've accused the author of bias ... but exactly what bias do you suspect he had? Say it straight out, please.

You gave a list of four reasons for saying that the article was unscientific. I've refuted three of them, and am still waiting for detail on the Simpson's law claim.

[Zeta-Flux]

cheesehead,

I found your last two posts very informative. Thank your for taking the time to answer the questions on the quiz. As you (repeatedly) have asked me to provide some sense to the questions, let me do so now.

Point #1: There was absolutely no reason at all for dropping the two "outliers". You are entirely correct that that would be a bad thing to do, in many ways and contexts. Even if we changed the outliers from the extremal points to those with the most variance from the mean (or some other statistical measure), it would still be a bad idea on many fronts. One doesn't simply drop data points, without a good reason to believe that the data points were due to faulty measuring. [And even then, one should repeat the measurement, after fixing the fault. etc...]

So, why did I ask you to do it? To see your response, of course. I hinted at this, when I first created the quiz. In one sense, the questions had nothing to do with the paper, per se.

Now, what was your response? You, yourself, characterized them as "harsh speculations." You felt justified in forming opinions about the intent of the questions, due to your understanding of certain principles of statistics.

I similarly understood one simple fact: correlation does not imply causation. Yet, when I read some of the article in question, he seemed to be saying exactly that. For example, the author writes "The view of the U.S. as a ‘shining city on the hill’ to the rest of the world is falsified." For more such quotes see section 12 in http://moses.creighton.edu/JRS/2006/2006-1.html In fact, they write "[Paul's] methodological problems do not allow for any conclusive statement to be advanced regarding the various hypotheses Paul seeks to demonstrate or falsify."

Point #2: You approached the quiz with a critical eye. You looked for problems with the questions, from the outset. Yet, you seemingly passed over some of the glaring problems you uncovered in the article. For example, how does the author define prosperity? You give him the most charitable interpretation: prosperity must mean something generally understood by the specific social scientists he is talking to; even though Paul is a freelance paleontologist.

But this isn't the case. This is, in fact, one of the points brought up in http://moses.creighton.edu/JRS/2006/2006-1.html, that G. Paul is conceptually ambiguous.

You similarly give him a free pass when even basic statistics are not present in the article. See http://moses.creighton.edu/JRS/pdf/2006-7.pdf for another response on that front; revealing further complexities.

You gave a list of four reasons for saying that the article was unscientific. I've refuted three of them, and am still waiting for detail on the Simpson's law claim.

If you say so. Although Gary Jensen (the author of the second article I linked to) says "Paul’s analysis generates the 'desired results' by selectively choosing the set of social problems to include to highlight the negative consequences of religion". Sound familiar? [BTW, I already gave details on the Simpson's law claim; I even linked to a paragraph on wiki.]

[Zeta-Flux]

You've accused the author of bias ... but exactly what bias do you suspect he had? Say it straight out, please.

I think the following quote from Jensen sums it up: "His conclusions were based on an examination of scatter-plots for a small set of nations with no attempt to consider alternative explanations nor to encompass the research in the larger body of sociological theory and research on the topic."

[cheesehead]

Twice I've had a post here that I composed off-line, as per good suggestion by Zeta-Flux. In each case, it looked so different on-line that I kept revising and revising and revising. So I've just replaced it with this for now.