The PhilPapers Fallacy (Part 4 of 9)

Table of contents

Series announcement

Part 1: Introduction

Part 2: Relevant expertise matters

Part 3: Causality matters

Part 4: Selection effects matter

Part 5: How PhilPapers respondents interpreted the survey questions matters

Part 6: Independence matters

Part 7: Philosophical fashions matter

Part 8: Demographic, social, and cultural forces matter

Part 9: Conclusion

---

2.3 Selection effects matter

There may be substantial selection effects that cause people who are disposed towards a view to study philosophy, or to specialize in a given topic.

For instance, specialists in philosophy of religion are disproportionately likely to be theists, compared to philosophers in general. 

Philosophers in general:

Theism 18.9%

Atheism 70%

Philosophers of religion”
Theism 69.5%
Atheism 19.9%

Note how massive of a reversal this is. And who are the specialists, exactly? How much more do the opinions of philosophers of religion count on the matter of theism than of analytic philosophers in general? 

And what do you think is behind this effect? It seems far more plausible to me that most of the reason for the high rate of theism among philosophers of religion is that theists and people disposed towards theism are more likely to study philosophy of religion than that the arguments in philosophy of religion are just so good that they cause those who study them to become theists. 

Should we take the fact that most philosophers who specialize in philosophy of religion are theists as good evidence of theism? I don’t think so. It’s plausible that those who choose to study philosophy of religion are more likely to be theists from the outset. These results are consistent with studying philosophy of religion having a causal impact on reducing theism. Indeed, this may be the case, since there is at least some evidence that studying philosophy of religion tends to cause people to become atheists, which you can read about here. One study exploring why philosophers began studying philosophy of religion indicates that overall, more philosophers became atheists or agnostics as a result of studying philosophy of religion than the other way around. As Mannino observes:

These numbers show that there was an overall shift toward atheism/agnosticism of 3.7% if we compare both directions of belief-revision: the direction of belief-revision was most frequently in the direction of atheism/agnosticism.

These findings illustrate how selection effects can obscure the fact that a field can cause a reduction in belief from an initially high baseline. Under these circumstances, it’s unclear why we should think that the fact that most philosophers of religion are theists is good evidence of theism.

Selection effects could have huge magnitudes within specializations. But it’s also possible that analytic philosophy as a whole tends to attract people with certain dispositions and mindsets that differ from other people. The number of people who pursue PhDs in the sciences is far greater than the proportion who pursue PhDs in philosophy. Would you guess that there is zero difference in a disposition towards empiricism between the two populations? If so, I’d be happy to take bets on the matter with you: my bet is that those who prefer STEM fields are much more likely, on reflection, to favor an empiricist over non-empirical approach to philosophical issues, if they could be persuaded to consider the matter carefully. And I’d bet they did so before acquiring PhDs in the sciences.

What we don’t know, then, is if the 62% of philosophers who endorse moral realism endorse it as a result of the quality of the arguments for realism, or because they adjusted downwards from an initially high baseline, or because people who study philosophy are more likely to be realists to begin with (including those who specialize in metaethics, since the rate of realism is about 65% for specialists), or (as is most likely) it’s some combination of these, and it’s really a matter of how much these and other factors contribute. It’s certainly possible that studying analytic philosophy increases support for moral realism. If it does, that would serve as better evidence than the mere 62% itself.

Given the possibility of substantial selection effects and the lack of clarity about the relevant causal factors, it’s altogether unclear why we should take a slight majority in favor of a view to be much evidence for moral realism. Note that evidence that realism increases when studying philosophy would serve as evidence for realism, but it would still not be decisive. One would still need to demonstrate that this increase was a result of reliable methods.