| Both sides previous revisionPrevious revision | |
| knowledge:ecological_fallacy [21.07.26, 08:55:30] – sascha | knowledge:ecological_fallacy [21.07.26, 09:29:17] (current) – [IQ differences between population groups] sascha |
|---|
| ==== IQ differences between population groups ==== | ==== IQ differences between population groups ==== |
| |
| FIXME This article is still in preparation and thus incomplete. | There are a whole heap of //misconceptions// regarding the differences between [[wp>Normal distribution|normally distributed]] properties. This will be illustrated here using the so-called "[[wp>Intelligence quotient|intelligence quotient]]" (<abbr>IQ</abbr>) as an example: |
| | |
| | Comparisons of IQs between different population groups (e.g. broken down by occupational groups, countries of origin, political party preferences, media consumption, etc.) are published frequently. Such comparisons often contain, at the very least, certain elements of [[wp>rhetoric/appeal_to_emotion/index|appeal to emption]] or //self-aggrandizement//, which is why they are so readily shared on social media. After all, who wouldn’t like to believe that a group to which they feel they belong is more intelligent than others … |
| | |
| | It would take us too far out of the scope of this web site if we tried to list all the problems that such comparisons have. However, drawing a conclusion about differences in the intelligence of //group members// based on differences in //average// scores – a scenario specific to the topic of this article – can be seen as a //perfect// example of the //ecological fallacy//. In short: even in the "most intelligent" group, there are always individuals of varying degrees of intelligence. |
| | |
| | <aside float-right> |
| | <figure box> |
| | {{:knowledge:comparing_normal_distributions.svg?256|Comparison of two normally distributed characteristics}} |
| | <figcaption>Two //normal distributions// with a large area of overlap</figcaption> |
| | </figure> |
| | </aside> |
| | |
| | Since the IQ is, by definition, //normally distributed,// it would in principle even be possible to calculate the overlap between the distributions of this trait in the two groups with great precision. However, it is not really worth the effort, as the figures are still rather imprecise and, in most cases, a distinction can only be made if one states them with a [[mathematics:statistics:misrepresentations:spurious_accuracy|misleading degree of precision]]. In most cases, however, the area of overlap between the two populations is so large (see figure) that it is hardly possible to draw any meaningful conclusions. |
| | |
| | Under no circumstances can any conclusions be drawn from such //group averages// with regard to the IQ of //individuals//. Even less so, by the way, with regard to their //intelligence//, which is not necessarily the same thing (<span maniculus "More information under:">[[abstraction/semiotic_fallacy|Semiotic fallacy]]</span>). |
| |
| ===== See also ===== | ===== See also ===== |
