IOW, the Standard Deviation that is used on the curve has nothing to do with the accuracy of the IQ test, it’s simply a way of scoring the test so most people score closer to a score of 100 and fewer score far away from 100. That is, it has nothing to do with the actual distribution of the scores (if they are actually normal).
What you’re really interested in is how accurately the test measures where you belong in this distribution. That is, if the Ground Truth of your IQ is 125, then you should expect that the IQ test should score you with a mean of 125, not 100. Thus, what you really need to compare is the samples taken for the different aged children and compare them using the Standard Deviation of the test that was used. I imagine they have tests with error FAR less than 15 points.
Further, we can use a t-test to determine how different two populations are, and the differences can be made more clear by either having smaller standard deviations OR larger populations. Thus, a difference of 2.3 points could be legit, but I’d have to see the paper to make sure.
Note that the main point is determining if the effect is gestational or social. The birth order effect itself was first claimed by Sir Francis Galton, Darwin’s cousin and one of the founders of modern statistics, in 1874. The effect has since been confirmed in numerous studies, but I do not know the extent of research on the subject in different cultures and social systems globally.
I noticed that children who were not born first, but who became the eldest child in the family due to the first child dying also had the higher IQ:
The implication here is that the IQ difference is due to nurture, not nature. That’s interesting, because the upshot then is that parents treat kids differently and somehow foster greater intelligence in their first children. Is it because the first child has time alone with the parents before the next child comes along? I would think a study of how only children compare to eldest siblings might show if this is true.
Another question comes to mind. At what point is IQ fixed? How old does the child have to be when he becomes eldest, if his older sibling dies, in order to reap the benefits of his changed social rank within the family?
I have read various ideas on this subject. One hypothesis is basically the one you suggest - the eldest child spends some time in an environment in which the average age of the entire family is higher than for any subsequent children. Another is that the eldest child learns more in the process of teaching his younger siblings.
You think that the finding that differences in IQ is in part due to familial social factors, rather than genetic or gestational, is almost entirely irrelevant? Even if females have an entirely different pattern, certainly a finding of such a pattern for half the population is of some interest with regards to the development of IQ.
The OP completely misunderstood the entire point of the study, as well as the statistics behind it. What is your reasoning for why it is “useless”? I trust it is not the same as the OP’s.
Don’t forget to check out their supplementary material, including a full description of the sampling methods and statistical analysis. They even give the raw output of the multiple linear regression analysis (I’ve never seen that before).
Very cool study on a fairly esoteric question. The oft documented very minor but real superior performance of eldests on standardized IQ tests is indeed due to their social ranking and clearly not to some biological effect of the mother being multiparous or the father’s advanced age. As to the significance of those IQ points, one might read the accompanying Perspective
The Perspective goes on to note some other intersting facts including that younger sibs actually do better when young but that apparently the opportunity to tutor the younger sibs drives the eldest up as time goes on. Only childrren have no such benefit and do not do as well.
I sense that part of your ire is directed toward the Media giving this study a distribution and trumpeting beyond its significance, given the relatively small difference that was measured.
My gut is that the major traction for this story comes from the opportunity to promote nurture over nature as a cause for IQ differences. This is the spin I have seen put on the study by most of the stories. Were it a study that supported nature (genes) over nurture, with the same near-trivial (albeit statistically significant) difference, I think it would have been left alone as Too Hot to Handle.
It’s not politically incorrect to announce the inferiority of us younger siblings as long as it is in the name of promoting Nurture over Nature, which is a larger Correctness.
As mentioned above I started a thread in GD over this story because I was struck at how small the difference was and therefore am wondering if the study will backfire on the Nurture advocates. If the difference is that small, is the only remaining explanation for differences among siblings Nature?
I don’t see any problem talking about difference in intelligence due to genetics. People often freak out if you throw race into it, but there can be no doubt that some people are just born smarter than other people. Sheesh, if intelligence weren’t hereditary, we’d be romping with our cousins, the chimps.
I’ve responded to this in the GD thread. The 3 point difference is only that attributable to birth order and social rank. The study controlled for other nurture factors that could be important even within families, among them maternal age at birth, sibship size, birth weight, and year of conscription. Of course, there are many other nurture factors that can account for IQ differences between families and between other social groups, including socioeconomic status, education, culture, health, and nutrition.
A few years back I read Steven Pinker’s The Blank Slate: The Modern Denial of Human Nature (having previously been impressed by his other books) and was amazed that he wrote a whole book stating what to me seemed like a non-issue. He must have read more of these threads than I had.
His fairly well documented upshot was briefly summarized on page 380-381:
Shared Environment refers to family and Unique Environment is everything else that happens to you differently than to your siblings, from birth order to infections to bumping into the bully at the wrong time in school. He of course gives all the caveats: these studies exclude extreme cases like abuse, neglect, and differences between cultural groups. His numbers may or may not be spot on, but most studies bear it up.
IQ may indeed be a “mismeasure of man” but the concept that we have strong biological predispositions to our intellectual potentialities and that how they are manifest depends on culture, our unique experiences, and our families (and likely in that order) seems hard to dispute other than as a matter of political manifesto.
Just one more thing to add to this - the confidence interval is not the same as measurement error. Whatever the measurement error is for IQ tests, unless there is a reason to believe that the measurement error varies depending on birth order, over a big enough sample it will wash out when comparing two populations.
I’d love to find out if the results change in different societies. The Times article I read said nothing about this.
And an observation: when a random poster blasts the authors of a paper in Science for being statistically naive, the random poster almost certainly doesn’t know what the hell he’s talking about.