Friday 9 January 2015

Educated parents more important than rich parents

Although academics should be above such things, sometimes a particular paper becomes cherished because it seems to prove a particular point, and becomes a totem for a general world view. For many years it has been taken as axiomatic that family environments have a very strong impact on child development. How could they not? Parents are there every day to inculcate values, teach skills and attitudes, nurture or neglect, and shape their young charges into the paths of righteousness, or sling them down the slippery slope of dissolution. Furthermore, rich parents are able to buy their children educational toys, books, tutors, quiet rooms for study, good schools, proper accents and habits, and the shiny baubles of wealth indicators.

Social class is often seen as the engine room of social success, and most psychological and sociological research measures socio-economic status as a matter of course, sometimes neglecting other measures as a consequence. However, social class is far from being a pure sociological measure, most notably because most parents are bringing up their own children, so genetics and social forces are confounded. Absent a genomic analysis of parents and children (which may become standard one day) it would be at least one step forward to measure the intelligence of parents and their children. Useful, but somewhat time consuming and expensive compared with a quick assessment of social class.

As a consequence, most data sets use cheap and cheerful measures: socio-economic status of family, and educational level of parents, either as highest level achieved, or total years of schooling. Both of these last measures are pretty crude, and will under-estimate the finer detail of scholastic attainments.

The general picture which emerges from worldwide PISA results is that parental education is more influential than parental status or wealth. In the German PISA sample for Germany there were very weak correlations between wealth indicators and academic achievement: PISA sum and availability of one's own study place:
r=.09(N=32947); own desk: r=.11 (N=32983);mobile: r=−.25 (N = 32113); TV: r = −.06 (N = 32968); game computer: r = −.20 (N = 31990); and video: r = −.22 (N = 31861). By contrast, another indicator of students' learning environment, which requires relatively less expenditure than electronics, is much more and stably and
positively related to their achievement: the number of books: r = .43 (N=32568).

Having a mobile phone, a video recorder and a game computer are associated with lower scholastic ability, and the only substantial positive correlation is with the number of books, and of course the cause may not be the books themselves, but the intellect and character of the families who choose to buy books.

The frequently supposed positive effects of wealth-driven private school attendance are hardly supported by research results. According to studies done in 22 Western countries (Dronkers & Robert, 2008a,b), the “effect of private schools” on PISA scores through parental attributes is much larger than the small positive or even negative effects of private school attendance itself: “The main explanation of their gross differences in scholastic achievement is the better social composition of private
schools.” (Dronkers & Robert, 2008a, p. 541). Thus, these schools do not make students smarter, but smarter kids improve these schools, at least their measured achievement test results. Similar results were found by Bond and Saunders (1999) for Great Britain: Effects of private school attendance varied between a small positive impact on abilities (β = .07) and negative ones on later occupational status (sum
β = −.12).

Educational level is an indicator of parental cognitive ability, which can influence their educational behavior and their shaping of a benign developmental environment. This can include wealth (e.g. via books) having an environmental impact on children's cognitive development. Further, parental educational level
can be an indicator of genes underlying parents' and children's behavior.

Now Rindermann and Baumeister have taken a close look at Hart and Risely (1995) and find that as regards cognitive development, parental education is more important than socio-economic status.

Rindermann, H. & Baumeister, A. E. E. (2015). Parents’ SES vs. parental
educational behavior and children’s development: A reanalysis of the
Hart and Risley study. Learning and Individual Differences, 37, 133-138.

In their seminal study “Meaningful differences in the everyday  experience of young American children”, Hart and Risley (1995) showed a  close relationship between differences in family environment and  children’s development. The conventional conclusion was that children’s  cognitive development (measured by psychometric intelligence and verbal  ability tests) depends on parental socioeconomic status (SES), especially on wealth. We reanalyzed their data and show that the quality of education given by parents is crucial (beta-PEB=.58) and not the diffuse aggregate measure of SES (beta-SES=.11). Additionally, we  compare their sample with a similar but larger sample (Hoff, 2003) showing the same pattern of results. Possible causal factors (associated environmental and genetic factors) are discussed.

As you will see, the original study had a sample size of 42, of whom 17 were African American. So, a very small sample on which to make any statements about large issues, and furthermore, a sample which conflates social and ethnic effects. Is there any hope for psychology?


  1. Is there any hope for psychology? I doubt it.

    I studied psychology at a lower tier university, where, due to the nature of the job market, the professors all had impressive degrees from Harvard, Yale, and McGill.

    The level of scholarship wasn't very impressive.

  2. "the better social composition of private schools": better? How very dare they?

    The freedom-of-speech lovers at the Grauniad will be after them for that infelicity.

  3. "the original study had a sample size of 42, of whom 17 were African American": that's a bit scathing. Is it a micro-aggression?

  4. A micro-observation, since whatever the interpretation of racial differences in intelligence the environmental component would be highly relevant and could bias the results.

    1. there is no "environmental component".

      the P(G,E) surface does NOT equal G + E.

      are you even capable of understanding that?

    2. it's clear at this point that you either:

      1. have no interest in the truth or

      2. you're too stupid to understand it.

  5. My understanding is that micro-aggression includes any failure to praise to the skies. On American campuses anyway. And if now there, then soon here.

  6. ...and becomes a totem for a general world could it not?...

    just more blah blah blah from this chav.

    a false and simplistic weltanschauung is your problem.

    what's "it"?

    apparently psychologists are incapable of understanding that genes have no independent effect on any psychological traits. (that is so far as psychological traits themselves exist independently of the particular time and place.) and that's not "truthiness" on my part. it's a fact.

    Jimmy was raised on a council estate so he did what every gay and craven man does, he bent over for the Establishment and asked, "Please sir. I want some more."

    one might suppose that there is some sort of nature vs nurture argument which may be engaged in. no.

    hereditism = lower class.

    the problem isn't facts or eve their interpretation. the problem is that hereditists are all psychologists and have low IQs. they can't even grasp that h^2 is only a measure of how well the P(G,E) surface can be approximated by a plane within a very restricted range of G and E.

    if only there were any psychologists in the BGI study.

  7. It's the same the whole world over
    It's the dim wot gets the blame
    It's the bright wot gets the pleasure
    Innit all a bleedin' shame?

  8. But to be more solemn: is this really the first study to have a look at this effect? And is a heterogeneous sample of 42 the best available? Golly.

    1. As far as I know this is an often quoted study. Rindermann will be publishing on larger samples shortly