Religion Damages Society?
Yes, says UK Timesonline:
Societies worse off 'when they have God on their side'
By Ruth Gledhill, Religion Correspondent
http://www.timesonline.co.uk/article/0,,2-1798944,00.html
The research on which these assertions are supposedly based is reported in "Cross-National Correlations of Quantifiable Societal Health with Popular Religiosity and Secularism in the Prosperous Democracies: A First Look," by Gregory S. Paul of Baltimore, Maryland. The article appeared at Journal of Religion and Society, an online journal out of Creighton University, and is available at http://moses.creighton.edu/JRS/pdf/2005-11.pdf. The author does not list any academic affiliation.
The article compares the US with a group of European countries, plus Canada, Australia, New Zealand, and Japan, on a selected list of social welfare indicators, finding that the US generally ranks poorly. The US is also by far the most religious of the countries examined, so we automatically have a correlation between religiosity and low ranking on some social indicators. Does that mean religion is causing those low rankings, as the Times article claims?
A causal connection could only be taken seriously under two conditions: (1) The large number of other factors that could explain the differences in crime rates, etc. between the US and these other countries have been introduced as control variables; (2) Some plausible description is provided of the causal paths by which the author thinks religion might produce these negative impacts, such as exactly how an increase in church attendance produes an increase in the rate of syphilis. If the relationship between religion and low social indicator ranking holds up when other possible explanations are accounted for, and if there is some plausible account of how religion operates to create those deficits, then we have something to talk about. But there is no mention of either of those requirements in the Times article, which suggests that their Religion Correspondent didn't know enough to ask.
In the article itself, we find the following:
The first paragraph quoted above explicitly disavows any causal inferences. It simply describes the two things the study tells us. First, that European social democracies generally score better than the US on some social welfare indicators, which is something everybody already knows. And, second, that the US is more religious than those other countries, something else everybody already knows. Something new might have to do with causation, but about that we learn nothing. For all we can tell, religion is the effect, with low rank on social indicators causing people to become more religious and higher rank causing them to become less so. Much more likely, of course, is that the entire correlation between religion and social indicators is is not causal in either direction, but a spurious product of the myriad uncontrolled factors that distinguish the US from these other countries, like racial-ethnic composition, socioeconomic inequality, organization of social services, family structure, and availability of guns, among many others. If those differences were properly adjusted for, the original correlation of religion with social indicators might well have disappeared completely.
All of which, of course, raises the obvious question: Why was this published?
The second paragraph quoted above tells us that the author did not try to assess any causal link between any variables in his study (1) Because of variability in the size of correlations, which must mean something, but I can't guess what; (2) Because the possible paths of causation are too complex, which is bizarre since that's the precise reason you do regression analysis, not a reason to avoid doing it; (3) Because he didn't want to, by which point one begins to wonder about the level of statistical expertise in play here; (4) Because regression analysis risks producing misleading results, which is certainly true if you don't know what you're doing; and (5) No statistical controls are necessary because the countries sampled are so similar, which is preposterous.
There appears in this list not a single good excuse for failing to carry out the multivariate analyses necessary to find out, at minimum, whether the zero-order correlations are spurious. So we are left with a trivially predictable pattern of simple correlations on which no light has been shed, but which lends itself to misinterpretation as so amply demonstrated by the Timesonline Religion Correspondent.
But, never mind, the Timesonline's misleading and sensationalist headline went around the world at the speed of light, Mr. Paul is appearing on Fox this evening, and thus is advanced the debate over the role of religion in society. Bad show all around.
Societies worse off 'when they have God on their side'
By Ruth Gledhill, Religion Correspondent
http://www.timesonline.co.uk/article/0,,2-1798944,00.html
RELIGIOUS belief can cause damage [note that wording] to a society, contributing towards high murder rates, abortion, sexual promiscuity and suicide, according to research published today.
According to the study, belief in and worship of God are not only unnecessary for a healthy society but may actually contribute [again] to social problems.
The research on which these assertions are supposedly based is reported in "Cross-National Correlations of Quantifiable Societal Health with Popular Religiosity and Secularism in the Prosperous Democracies: A First Look," by Gregory S. Paul of Baltimore, Maryland. The article appeared at Journal of Religion and Society, an online journal out of Creighton University, and is available at http://moses.creighton.edu/JRS/pdf/2005-11.pdf. The author does not list any academic affiliation.
The article compares the US with a group of European countries, plus Canada, Australia, New Zealand, and Japan, on a selected list of social welfare indicators, finding that the US generally ranks poorly. The US is also by far the most religious of the countries examined, so we automatically have a correlation between religiosity and low ranking on some social indicators. Does that mean religion is causing those low rankings, as the Times article claims?
A causal connection could only be taken seriously under two conditions: (1) The large number of other factors that could explain the differences in crime rates, etc. between the US and these other countries have been introduced as control variables; (2) Some plausible description is provided of the causal paths by which the author thinks religion might produce these negative impacts, such as exactly how an increase in church attendance produes an increase in the rate of syphilis. If the relationship between religion and low social indicator ranking holds up when other possible explanations are accounted for, and if there is some plausible account of how religion operates to create those deficits, then we have something to talk about. But there is no mention of either of those requirements in the Times article, which suggests that their Religion Correspondent didn't know enough to ask.
In the article itself, we find the following:
This study is a first, brief look at an important subject that has been almost entirely neglected by social scientists. The primary intent is to present basic correlations of the elemental data. Some conclusions that can be gleaned from the plots are outlined. This is not an attempt to present a definitive study that establishes cause versus effect [somebody please tell Ms. Gledhill, the Religion Correspondent] between religiosity, secularism and societal health. It is hoped that these original correlations and results will spark future research and debate on the issue.
Regression analyses were not executed because of the high variability of degree of correlation, because potential causal factors for rates of societal function are complex, and because it is not the purpose of this initial study to definitively demonstrate a causal link between religion and social conditions [Ms. Gledhill?]. Nor were multivariate analyses used because they risk manipulating the data to produce errant or desired results, and because the fairly consistent characteristics of the sample automatically minimizes the need to correct for external multiple factors. Therefore correlations of raw data are used for this initial examination.
The first paragraph quoted above explicitly disavows any causal inferences. It simply describes the two things the study tells us. First, that European social democracies generally score better than the US on some social welfare indicators, which is something everybody already knows. And, second, that the US is more religious than those other countries, something else everybody already knows. Something new might have to do with causation, but about that we learn nothing. For all we can tell, religion is the effect, with low rank on social indicators causing people to become more religious and higher rank causing them to become less so. Much more likely, of course, is that the entire correlation between religion and social indicators is is not causal in either direction, but a spurious product of the myriad uncontrolled factors that distinguish the US from these other countries, like racial-ethnic composition, socioeconomic inequality, organization of social services, family structure, and availability of guns, among many others. If those differences were properly adjusted for, the original correlation of religion with social indicators might well have disappeared completely.
All of which, of course, raises the obvious question: Why was this published?
The second paragraph quoted above tells us that the author did not try to assess any causal link between any variables in his study (1) Because of variability in the size of correlations, which must mean something, but I can't guess what; (2) Because the possible paths of causation are too complex, which is bizarre since that's the precise reason you do regression analysis, not a reason to avoid doing it; (3) Because he didn't want to, by which point one begins to wonder about the level of statistical expertise in play here; (4) Because regression analysis risks producing misleading results, which is certainly true if you don't know what you're doing; and (5) No statistical controls are necessary because the countries sampled are so similar, which is preposterous.
There appears in this list not a single good excuse for failing to carry out the multivariate analyses necessary to find out, at minimum, whether the zero-order correlations are spurious. So we are left with a trivially predictable pattern of simple correlations on which no light has been shed, but which lends itself to misinterpretation as so amply demonstrated by the Timesonline Religion Correspondent.
But, never mind, the Timesonline's misleading and sensationalist headline went around the world at the speed of light, Mr. Paul is appearing on Fox this evening, and thus is advanced the debate over the role of religion in society. Bad show all around.
posted by Byron | Wednesday, October 05, 2005
|
0 comments
