Accused of a fallacy? Suspect a fallacy? Ask Dr. Bo and the community!

Quickly register to comment, ask and respond to questions, and get FREE access to our passive online course on cognitive biases!

# Ecological Fallacy

(also known as: ecological inference fallacy)

Description: The interpretation of statistical data where inferences about the nature of individuals are deduced from inference for the group to which those individuals belong.

Logical Form:

Group X has characteristic Y.

Person 1 is in group X.

Therefore, person 1 has characteristic Y.

Example #1:

Men score better on math than women do. Jerry is a man. Therefore, Jerry is better at math than Sylvia, who is a woman.

Explanation: The fact that men score better on math than women is a group generalization. This does not mean that any individual man will score better than any individual woman on math. Educated guesses could be made if we knew more about the statistics. For example, if we just knew that men scored an average of 8% higher than women, we could not even say that any given man is likely to be better at math than any given woman. This is because there could be what is referred to as an uneven distribution, that is, there could be a small group of women who are really bad at math or a small group of men who are really good at math that throws off the curve.

Example #2:

A study was done recently showing that church attendance was positively correlated with marriage longevity, that is, those couples who attended church together more often were more likely to stay married. This really should not be a surprise considering the general view of divorce within religion. What this does not mean is that any given couple who does not attend church is more likely to get divorced than any given couple that does attend church.

Explanation: To make this claim, we would need more information on the raw data used. Perhaps just religious fanatics who go to church daily have a practically non-existent divorce rate of say 2%. Then it is possible that those who never go to church have a lower divorce rate than those who do go to church every Sunday, but because of the fanatics, the distribution is not evenly distributed.

Tip: Although this is a statistical fallacy, it is commonly extended to everyday situations. If you don't mind getting into statistics a bit, understanding statistical fallacies could improve one's overall reasoning ability.

References:

Babbie, E. R. (2016). The Basics of Social Research. Cengage Learning.

Nitonise
Friday, April 26, 2019 - 02:13:11 PM
I feel like this entry confuses three different fallacies. Namely in its short description it says that it's fallacious to reason about an individual based on his/her/its population. I both agree and disagree. It's a fallacy from the point of view of deduction, but it's basically as induction works. And for induction it's completely reasonable to conclude that "Therefore, person 1 has characteristic Y.", just keep in mind that it's not 100% certainty.

I name the second fallacy "baseless mean=mode" (I made up the name because I don't know the name for this fallacy). It's what was dicussed in example #1, situations when we conclude that something is in majority based on the mean without good reason. When we have normal distribution it's true that the mean and the mode are equal, but in other cases can be false. For example, if distribution looks like the normal distribution, but is left-skewed (example:outstanding idiots bring the whole average IQ down) or right-skewed (example:geniuses are so smart that they bring the whole average IQ up).

As for the third fallacy, I can't quite put my finger on it, but assuming that induction is OK it's different from "baseless mean=mode". In the second example we deal with correlation instead of average. I haven't learned how to calculate correlation yet, so maybe I will post latter more technical description about what exactly goes wrong here.

2 replies

working...

Bo Bennett, PhD
Friday, April 26, 2019 - 02:23:01 PM
I feel like this entry confuses three different fallacies.

Many fallacies contain other fallacies within or are very similar to other fallacies.

Namely in its short description it says that it's fallacious to reason about an individual based on his/her/its population.

Actually, it says "The interpretation of statistical data where inferences about the nature of individuals are deduced from inference for the group to which those individuals belong." The key here is "deducing from inference." We can make deductions, and we can make inferences, we just can make a deduction from an inference. Deduction (in logic, not casual usage) is a process that yields a certain conclusion. This cannot, or should not. Thus, the fallacy.

working...

Nitonise
Friday, April 26, 2019 - 02:47:26 PM
@Bo Bennett, PhD: Thanks for answering!

Many fallacies contain other fallacies within or are very similar to other fallacies.

I don't know about other fallacies (I inspected this particular fallacy because I wanted to clear up some things in my understanding of statistics), but personally for me it's hard to understand what this particular fallacy is about at all if we assume that induction is acceptable (and examples seem to be based on assumption that induction is acceptable). This was the reason why I made up "baseless mean=mode" fallacy, in order to make sense from example #1.

The key here is "deducing from inference." ... Deduction (in logic, not casual usage) is a process that yields a certain conclusion. This cannot, or should not. Thus, the fallacy.

Ah, I see. But then it's strange that your examples involve inferential statistics. Inferential statistics is based on induction, not deduction. Your explanations of examples are strange too. If it was your prime goal, then instead of saying things like "there could be a small group of women who are really bad at math or a small group of men who are really good at math " you could just said something like "There is still non-zero chance that Jerry is bad of math, even if 99.999% of men are good at math, thus fallacy". Both examples and their corresponding comments gave me impression that induction is acceptable as long as it's done correctly.

working...

Jonathan Thomas
Friday, December 21, 2018 - 11:44:01 AM
Just saw an example of this today:

"A new study says people who drink a glass of wine a day are hospitalized less often than those who don’t. Do you drink wine?"

The inference is that if you drink wine you will be hospitalized less. When you think about it it's rather a ridiculous claim since it doesn't make any reference to the cause of the hospitalization, not to mention engaging in the fallacy itself.

0 replies

working...

Monday, April 17, 2017 - 01:43:54 PM
This looks to be the same as Fallacy of Division. Is there a difference?

working...

Bo Bennett, PhD
Monday, April 17, 2017 - 04:27:36 PM
This is a more specific form that refers to groups and individuals, and mostly within statistics.

working...

### Become a Logical Fallacy Master. Choose Your Poison.

Logically Fallacious is one of the most comprehensive collections of logical fallacies with all original examples and easy to understand descriptions; perfect for educators, debaters, or anyone who wants to improve his or her reasoning skills.

Get the book, Logically Fallacious by Bo Bennett, PhD by selecting one of the following options:

### Not Much of a Reader? No Problem!

Enroll in the Mastering Logical Fallacies Online Course. Over 10 hours of video and interactive learning. Go beyond the book!

Enroll in the Fallacy-A-Day Passive Course. Sit back and learn fallacies the easy way—in just a few minutes per day, via e-mail delivery.

Have a podcast or know someone who does? Putting on a conference? Dr. Bennett is available for interviews and public speaking events. Contact him directly here.

Website Software Copyright 2019, Archieboy Holdings, LLC.