Loading ...
Sorry, an error occurred while loading the content.
 

[evol-psych] Re: Iq" genetic or environmental?

Expand Messages
  • Gordon M. Burghardt
    A cute example that misses the point. Certainly the width and length can vary themselves and thus the area can vary if, as we know, IQs can vary. If length is
    Message 1 of 20 , Nov 30, 1999
      A cute example that misses the point. Certainly the width and length can
      vary themselves and thus the area can vary if, as we know, IQs can vary.
      If length is genetics and width environment then the contribution of each
      to the variance in IQ or any trait can, theoretically at least, be
      determined. All this example illustrates is that both are necessary, not a
      profound point, but one used to "prove" that instincts did not exist (and
      why Hebb developed this argument in his 1953 paper to begin with).

      Gordon Burghardt

      At 01:36 PM 11/30/99 -0500, Fredric Weizmann wrote:
      >Donald Hebb made essentially this same argument years ago. He said to
      >argue whether genetics or environments are more important in determining
      >IQ is like arguing whether the length or width is more important in
      >determining the area of a rectangle.
      >


      ----------------------------------------------------------------------------
      ------------
      Gordon M. Burghardt
      Department of Psychology
      University of Tennessee
      Knoxville, TN 37996-0900

      Ph: 423-974-3300
      Fax: 423-974-3330
      E-mail: gburghar@...
      Web site: http://web.utk.edu/~gburghar/

      Departmental websites: PSYCHOLOGY http://web.utk.edu/~jlawler
      ECOLOGY & EVOLUTIONARY BIOLOGY
      http://www.bio.utk.edu/ecology/index.html
      __________________________________________________

      Obligatory Quotation: And he gave her a look
      you could have poured over a waffle (Ring Lardner)
      __________________________________________________
    • Fredric Weizmann
      Actually, your reading of what I said is why I included the second part of my message; you may not agree with my interpretation, but I did not miss the point.
      Message 2 of 20 , Nov 30, 1999
        Actually, your reading of what I said is why I included the second part of
        my message; you may not agree with my interpretation, but I did not miss
        the point. The kinds of mistakes people make about intelligence lead to
        the implication that Hebb was writing about; when it is stated in
        its bald form people presumeably would not accept it, but what they do
        argue does represent the same kind of thinking. Arguments about
        the importance of environments and genes based on relative contributions
        to variance frequently lead to that kind of mistake. Yes of course one
        can compute relative variances, but statements about variance (and
        heritabilities, which are also based on variances) are quite limited and
        are descriptive in nature; variances do not constrain change.

        Fredric Weizmann

        On Tue, 30 Nov 1999, Gordon M. Burghardt wrote:

        > A cute example that misses the point. Certainly the width and length can
        > vary themselves and thus the area can vary if, as we know, IQs can vary.
        > If length is genetics and width environment then the contribution of each
        > to the variance in IQ or any trait can, theoretically at least, be
        > determined. All this example illustrates is that both are necessary, not a
        > profound point, but one used to "prove" that instincts did not exist (and
        > why Hebb developed this argument in his 1953 paper to begin with).
        >
        > Gordon Burghardt
        >
        > At 01:36 PM 11/30/99 -0500, Fredric Weizmann wrote:
        > >Donald Hebb made essentially this same argument years ago. He said to
        > >argue whether genetics or environments are more important in determining
        > >IQ is like arguing whether the length or width is more important in
        > >determining the area of a rectangle.
        > >
        >
        >
        > ----------------------------------------------------------------------------
        > ------------
        > Gordon M. Burghardt
        > Department of Psychology
        > University of Tennessee
        > Knoxville, TN 37996-0900
        >
        > Ph: 423-974-3300
        > Fax: 423-974-3330
        > E-mail: gburghar@...
        > Web site: http://web.utk.edu/~gburghar/
        >
        > Departmental websites: PSYCHOLOGY http://web.utk.edu/~jlawler
        > ECOLOGY & EVOLUTIONARY BIOLOGY
        > http://www.bio.utk.edu/ecology/index.html
        > __________________________________________________
        >
        > Obligatory Quotation: And he gave her a look
        > you could have poured over a waffle (Ring Lardner)
        > __________________________________________________
        >
      • H. M. Hubey
        ... Both perimeter P= 2*X + 2*Y and area A=X*Y are symmetric functions of width and length. Since both addition and multiplication are commutative (the order
        Message 3 of 20 , Nov 30, 1999
          Fredric Weizmann wrote:
          >
          > Donald Hebb made essentially this same argument years ago. He said to
          > argue whether genetics or environments are more important in determining
          > IQ is like arguing whether the length or width is more important in
          > determining the area of a rectangle.

          Both perimeter P= 2*X + 2*Y and area A=X*Y are symmetric
          functions of
          width and length. Since both addition and multiplication are
          commutative
          (the order does not matter) he could have also said that arguing
          about
          whether the width or length is more important in determining the
          perimeter
          of a rectangle.

          I suspect many people are upset at having to multiply for IQ

          i.e. Iq = G*E

          and find that something seems to be missing. Of course, it is
          missing.
          Can simple multiplication be the best model for something so
          complex
          when it does not suffice even for freshman physics which
          describes
          point masses, acceleration along straight lines, massless springs
          etc
          of nonvolitional dead bodies?


          > I suspect that even many of those who acknowledge that the environment is
          > important in performance still believe that qualitites like intelligence
          > are somehow preformed in the genes (possessing some sort of "dotted
          > line" or ghostly existence). The use of words such as "potential" or
          > "predisposition"--essentially meaningless in this context--often disguise
          > this belief.

          I think they want a more complex mathematical model which is
          better at
          taking into account much of the evidence from real life. For
          example,
          they might want an equation that does something like "if the
          genetic
          disposition is low...etc etc."

          As far as "performed in the genes" that does not register in my
          brain.
          What makes humans have the brain-size and character have its
          peculiar
          composition is in the genes. That is not deniable.


          > Fredric Weizmann
          >
          >
          --
          M. Hubey
          Email: hubeyh@... Backup:hubeyh@...
          WWW Page: http://www.csam.montclair.edu/Faculty/Hubey.html
        • Mark Macnair
          ... I know of no quantitative genetic model that would say IQ=G*E. The standard model is additive: P=A+D+E+G*E+I etc where P is the phenotype (IQ in this case)
          Message 4 of 20 , Dec 2, 1999
            >
            > I suspect many people are upset at having to multiply for IQ
            >
            > i.e. Iq = G*E
            >
            > and find that something seems to be missing. Of course, it is
            > missing.
            > Can simple multiplication be the best model for something so
            > complex
            > when it does not suffice even for freshman physics which
            > describes
            > point masses, acceleration along straight lines, massless springs
            > etc
            > of nonvolitional dead bodies?
            I know of no quantitative genetic model that would say
            IQ=G*E. The standard model is additive:
            P=A+D+E+G*E+I etc
            where P is the phenotype (IQ in this case) , A is the
            Additive genetic effect, and D the non-additive effect (ie
            dominance); E is the effect of the environment. G*E is not
            a multiplicative function, but a statistical inteaction, as
            are gene x gene interactions (I).

            These terms are all pretty meaningless when applied to
            individuals, but are easily interpreted as VARIANCES (when
            poplation diffrences in these terms can be determined).

            If you have suitable data (difficult to get for humans and
            IQ, but easy for plants and other phenotypes) you can get
            estimates of all these terms and show that this additive
            model of gene action is pretty good.

            Why should IQ be any difefrent to any other character in
            humans, animals or plants?

            >
            > I think they want a more complex mathematical model which is
            > better at
            > taking into account much of the evidence from real life. For
            > example,
            > they might want an equation that does something like "if the
            > genetic
            > disposition is low...etc etc."
            This contingent type model will come out in G*E term of the
            standard model.

            ----------------------
            Mark Macnair
            Professor of Evolutionary Genetics
            Director of Postgraduate Studies
            Hatherly Laboratories
            Department of Biological Sciences
            University of Exeter
            Prince of Wales Rd
            Exeter EX4 4PS
            UK

            Tel: 44 (0) 1392 263791
            Fax: 44 (0) 1392 263700
          • H. M. Hubey
            ... Great. I know that too. Isn t that a shame? ... Yes, that is also the model seen in the literature. That is also shameful. ... I guess explanation of why
            Message 5 of 20 , Dec 2, 1999
              Mark Macnair wrote:

              > I know of no quantitative genetic model that would say
              > IQ=G*E.

              Great. I know that too. Isn't that a shame?

              The standard model is additive:
              > P=A+D+E+G*E+I etc

              Yes, that is also the model seen in the literature. That
              is also shameful.

              > where P is the phenotype (IQ in this case) , A is the
              > Additive genetic effect, and D the non-additive effect (ie
              > dominance); E is the effect of the environment. G*E is not
              > a multiplicative function, but a statistical inteaction, as
              > are gene x gene interactions (I).

              I guess explanation of why it has to be multiplicative is
              not sufficient. Perhaps someone can explain why it has to
              be additive.

              Or is the reasoning like this: it is additive because X
              did it this way N years ago and we have all been following
              in his footsteps because we don't see why it should be
              additive or multiplicative, and we don't use probability
              theory or logic (crisp or logic).



              --
              M. Hubey
              Email: hubeyh@... Backup:hubeyh@...
              WWW Page: http://www.csam.montclair.edu/Faculty/Hubey.html
            • Fredric Weizmann
              There are many different IQ tests. The gold standard for tests is the Wechsler and the Stanford Binet tests, especially the former, (about which I know
              Message 6 of 20 , Dec 2, 1999
                There are many different IQ tests. The "gold standard" for tests is the
                Wechsler and the Stanford Binet tests, especially the
                former, (about which I know more). The Wechsler tests for both
                children and adults have just recently been renormed based on
                sampling different groups, age categories etc. representatively. They are
                really intended for people who are fluent in English, which means that
                (and these are tests which are given verbablly and on an individual
                basis) one must be cautious in interpreting the results when they are
                given to people who don't fit the demographic categories. There are also
                suggested supplementary tests for those who don't fit these categories.

                The Wechsler and SB tests
                are mainly used for clinical purposes, usually as part of a
                battery, and interpreted on highly individual basis by trained people,
                While they are not perfect, they give a pretty
                accurate picture of how people function intellectually (defined here in
                terms of the kinds of skills that would be useful in school; there are
                a number of people arguing for broader definitions of intelligence and
                broader ways of assessing it) at a given time.

                There are other tests in use in non-English speaking countries, and I
                suspect some of these are based on the Wechsler, but I am not sure of
                this. It is certainly necessary to norm the tests independently in each
                country in which they are used. The Wechsler tests use Canadian norms
                when used in Canada, for example, even though the test itself is the same.

                People have heard of what has been called the "Flynn Effect", which refers
                to the fact that IQs have been rising throughout the world for the last
                half-century or so. In that vein, the Wechsler
                Tests have to be renormed periodically because the scores tend to become
                higher as time goes on. By renorming the test the IQ itself tends to
                remain constant, although the raw scores may have gone up.(There are lots
                of theories about why this is the case, but the phenomenon itself is
                interesting and poses a challenge for those who believe that IQ scores are
                coextensive with what we mean by innate intelligence.) The IQ scores are
                derived from the raw scores, with the mean being set equal to 100, and a
                standard deviation beeing set equal to 15 point. One can also compare
                scores from people in one category (e.g., by age) with others in the same
                subcategory.

                In addition to the standard individual tests, there are a whole host
                of briefer, group-administered tests. Some are suitable for use with
                non-English speaking individuals (which does not mean they are "culture
                free") There are also tests--in particular
                the Ravens Progressive Matrices tests--which are used as "stand ins" for
                intelligence, even though they were not designed to be used as such.

                Fredric Weizmann
                On Wed, 1 Dec 1999, Gerry Reinhart-Waller wrote:

                > Gerry here: QUESTION: What formula is being used today to determine
                > Iq?
                > And are all ethnic groups in the US placed into the same sampling pool
                > whether or not English is their primary language? And in which language
                > is the Iq test written? Or is it administered verbally?
                > Thank you,
                > G
                >
                > Gordon M. Burghardt wrote:
                >
                > > A cute example that misses the point. Certainly the width and length
                > > can
                > > vary themselves and thus the area can vary if, as we know, IQs can
                > > vary.
                > > If length is genetics and width environment then the contribution of
                > > each
                > > to the variance in IQ or any trait can, theoretically at least, be
                > > determined. All this example illustrates is that both are necessary,
                > > not a
                > > profound point, but one used to "prove" that instincts did not exist
                > > (and
                > > why Hebb developed this argument in his 1953 paper to begin with).
                > >
                > > Gordon Burghardt
                > >
                > > At 01:36 PM 11/30/99 -0500, Fredric Weizmann wrote:
                > > >Donald Hebb made essentially this same argument years ago. He said to
                > >
                > > >argue whether genetics or environments are more important in
                > > determining
                > > >IQ is like arguing whether the length or width is more important in
                > > >determining the area of a rectangle.
                > > >
                > >
                > > -
                > > --------------------------------------------------------------------------
                > >
                > > ------------
                > > Gordon M. Burghardt
                > > Department of Psychology
                > > University of Tennessee
                > > Knoxville, TN 37996-0900
                > >
                > > Ph: 423-974-3300
                > > Fax: 423-974-3330
                > > E-mail: gburghar@...
                > > Web site: http://web.utk.edu/~gburghar/
                > >
                > > Departmental websites: PSYCHOLOGY http://web.utk.edu/~jlawler
                > > ECOLOGY & EVOLUTIONARY BIOLOGY
                > > http://www.bio.utk.edu/ecology/index.html
                > > __________________________________________________
                > >
                > > Obligatory Quotation: And he gave her a look
                > > you could have poured over a waffle (Ring Lardner)
                > > __________________________________________________
                > >
                > > --------------------------------------------------
                > > ---------------------
                > > To subscribe/unsubscribe/select DIGEST go to:
                > > http://www.egroups.com/group/evolutionary-psychology/info.html
                > > To contribute address your email to:
                > > evolutionary-psychology@egroups.com
                > >
                > > -----------------------------------
                > > ------------------------------------
                > > -- 20 megs of disk space in your group's Document Vault
                > > -- http://www.egroups.com/docvault/evolutionary-psychology/?m=1
                >
                >
                >
              • Irwin Silverman
                ... Incontestable as stated, but let s take an informal poll Purely hypothetical question - Knowing that heritability indexes based on co-twin data range from
                Message 7 of 20 , Dec 2, 1999
                  On Tue, 30 Nov 1999, Fredric Weizmann wrote:

                  > Arguments about the importance of environments and genes based on
                  > relative ontributions to variance frequently lead to that kind of mistake.
                  > Yes of course one can compute relative variances, but statements about
                  > variance (and heritabilities, which are also based on variances) are
                  > quite limited and are descriptive in nature; variances do not constrain
                  > change.

                  Incontestable as stated, but let's take an informal poll

                  Purely hypothetical question - Knowing that heritability indexes
                  based on co-twin data range from about .50 to .70, if you were planning to
                  contract for sperm or eggs for in-vitro fertilization, would you:

                  a) been completely disinterested in the I.Q. of the donor.
                  b) be casually but not critically interested.
                  c) be critically interested - in fact, insist on the score.
                • Peter Kabai
                  ... This basic model is additive, because variances are additive. Quite simple. Realistic models are far more complicated than the basic model, mainly because
                  Message 8 of 20 , Dec 3, 1999
                    > Mark Macnair wrote:
                    >
                    > The standard model is additive:
                    > > P=A+D+E+G*E+I etc
                    >
                    > I guess explanation of why it has to be multiplicative is
                    > not sufficient. Perhaps someone can explain why it has to
                    > be additive.

                    This basic model is additive, because variances are additive. Quite simple.
                    Realistic models are far more complicated than the basic model, mainly
                    because the interactive variance cannot be neglected in most cases.

                    The interpretation of heritability is also simple: the fraction of genetic
                    variance (specifically, the additive variance in the narrow sense of
                    heritability) of the total variance in a given population for a given trait
                    measured in the given way. So when discussing what heritability really means,
                    besides being critical about the methods of measuring the trait, one also
                    should keep in mind, that the estimate is restricted to the population in
                    which it was measured.

                    Heritability of the eye colour is zero in a population of individuals all
                    having brown eyes (or blue for that matter). Of course we all know, that the
                    colour of our eyes is pretty much controlled by our genes. We would know
                    that, even if there was no variation in eye colour, because we know precisely
                    how the specific genes act. Learning about the action of the genes give a
                    clue to the mechanism, while heritability estimates are purely descriptive.

                    There is hope (or worry) that combining neurobiology and genetic data the
                    normal variation in IQ will be explained on the level of mechanism.


                    Cheers, Peter Kabai
                  • PRGHOME@aol.com
                    In a message dated 12/3/99 5:09:37 AM Eastern Standard Time, pkabai@univet.hu ... ************************************* The only change needed, it seems to me,
                    Message 9 of 20 , Dec 3, 1999
                      In a message dated 12/3/99 5:09:37 AM Eastern Standard Time, pkabai@...
                      writes:

                      > Of course we all know, that the
                      > colour of our eyes is pretty much controlled by our genes. We would know
                      > that, even if there was no variation in eye colour, because we know
                      > precisely
                      > how the specific genes act. Learning about the action of the genes give a
                      > clue to the mechanism, while heritability estimates are purely descriptive.
                      >
                      > There is hope (or worry) that combining neurobiology and genetic data the
                      > normal variation in IQ will be explained on the level of mechanism.

                      *************************************

                      The only change needed, it seems to me, in this admirably succinct statement
                      of the way things are,* is in the last sentence. It would represent the state
                      of the game slightly better if it read "There is EVERY REASON NOW TO hope
                      (or worry)..."

                      _______________________________
                      *On The Way Things Are, see my obituary for T. Kuhn (and reply to R. Rorty,
                      et al.) in HIstory of the Human Sciences 10(1):125-28 (Feb.1997)
                    • H. Mark Hubey
                      ... THis is backwards. The variances are additive because the model is additive. ... That is true. Even using a simple additive model the interactive
                      Message 10 of 20 , Dec 3, 1999
                        Peter Kabai wrote:
                        >
                        > > Mark Macnair wrote:
                        > >
                        > > The standard model is additive:
                        > > > P=A+D+E+G*E+I etc
                        > >
                        > > I guess explanation of why it has to be multiplicative is
                        > > not sufficient. Perhaps someone can explain why it has to
                        > > be additive.
                        >
                        > This basic model is additive, because variances are additive. Quite simple.

                        THis is backwards. The variances are additive because the model is additive.


                        > Realistic models are far more complicated than the basic model, mainly
                        > because the interactive variance cannot be neglected in most cases.

                        That is true. Even using a simple additive model the "interactive"
                        variance has to be left out. But the interaction itself has to be
                        modeled as multiplication also, and it comes out multiplicative even
                        if the basic linear/additive model is used.

                        > The interpretation of heritability is also simple: the fraction of genetic
                        > variance (specifically, the additive variance in the narrow sense of
                        > heritability) of the total variance in a given population for a given trait
                        > measured in the given way. So when discussing what heritability really means,
                        > besides being critical about the methods of measuring the trait, one also
                        > should keep in mind, that the estimate is restricted to the population in
                        > which it was measured.

                        If the true state of IQ measurement is a multiplicative function of E and
                        G, then the heritability as measured actually is not the contribution of
                        genes but instead the environment. IT is easy to show, and it can be
                        found in my paper, the latest version of which I just put at my website.
                        The calculations are in the conclusion section.


                        --
                        Sincerely,
                        M. Hubey
                        Dept of Computer Science, Montclair State University
                        hubeyh@... http://www.csam.montclair.edu/~hubey
                      • Fredric Weizmann
                        Peter Kabai is correct. As the evolutionary ecologist H. Caswell wrote in 1989 (in a book entitled Ecological Concepts, edited by J. M. Chertok) ..it is by
                        Message 11 of 20 , Dec 3, 1999
                          Peter Kabai is correct. As the evolutionary ecologist H. Caswell wrote in
                          1989 (in a book entitled Ecological Concepts, edited by
                          J. M. Chertok) "..it is by
                          now well known that genetic variances and heritabilities are specific to a
                          particular population in a particular environment at a particular
                          time; they are not invariant properties of the species or of the
                          trait. Genetic covaniances are equally labile. They can vary with
                          environmental conditions as well as with such genetic phenomena as
                          inbreeding."

                          Insofar as the heritability of intelligence is concerned it has "bounced"
                          between .2 and .8 in various studies. Jensen
                          of course claimed it was .8, Plomin and others .5. The latter seems to be
                          generally accepted not because it represents some scientifically validated
                          heritability estimate, but because it seems to come in the middle of all
                          of the published estimates. As Henderson wrote in a 1982 review (and
                          nothing since then has
                          changed in this regard,) however, citing a midrange value is no more
                          defensible than citing an extreme one. Irwin Silverman says its somewhere
                          betweem .5 and.7. Even if one accepted that the chimera
                          of a "true" heritability lay within this range, it would still reflects a
                          difference between accounting for 25% of the variance and 49%--so thats
                          not exactly a definitive claim.

                          In answer to Irwin's challenge as to whether I would want to know the IQ
                          of a sperm donor, I have a couple of answers. The first is that that is
                          not exactly the first piece of data I would like. The second is yes it
                          would be relevant; first of all, there are genetic causes of retardation,
                          and a low IQ might be an index of that. More broadly a reasonable IQ
                          (given that I was permitted no other information) would serve at least to
                          screen out some people who could not function in society for some reasons
                          related to genes, (and IQ tests do pick up things other than
                          "intelligence".) I would be happier to accept other indices of social and
                          academic success, however.

                          I should also mention that if I were collecting sperm to
                          breed race horses I would also like to know how fast the sire ran, despite
                          the fact that the heritability of speed among race horses is pretty close
                          to zilch (because of selection).

                          Fredric Weizmann
                          On Fri, 3 Dec 1999, Peter Kabai wrote:

                          > > Mark Macnair wrote:
                          > >
                          > > The standard model is additive:
                          > > > P=A+D+E+G*E+I etc
                          > >
                          > > I guess explanation of why it has to be multiplicative is
                          > > not sufficient. Perhaps someone can explain why it has to
                          > > be additive.
                          >
                          > This basic model is additive, because variances are additive. Quite simple.
                          > Realistic models are far more complicated than the basic model, mainly
                          > because the interactive variance cannot be neglected in most cases.
                          >
                          > The interpretation of heritability is also simple: the fraction of genetic
                          > variance (specifically, the additive variance in the narrow sense of
                          > heritability) of the total variance in a given population for a given trait
                          > measured in the given way. So when discussing what heritability really means,
                          > besides being critical about the methods of measuring the trait, one also
                          > should keep in mind, that the estimate is restricted to the population in
                          > which it was measured.
                          >
                          > Heritability of the eye colour is zero in a population of individuals all
                          > having brown eyes (or blue for that matter). Of course we all know, that the
                          > colour of our eyes is pretty much controlled by our genes. We would know
                          > that, even if there was no variation in eye colour, because we know precisely
                          > how the specific genes act. Learning about the action of the genes give a
                          > clue to the mechanism, while heritability estimates are purely descriptive.
                          >
                          > There is hope (or worry) that combining neurobiology and genetic data the
                          > normal variation in IQ will be explained on the level of mechanism.
                          >
                          >
                          > Cheers, Peter Kabai
                          >
                          >
                          > ------------------------------------------------------------------------
                          > To subscribe/unsubscribe/select DIGEST go to:
                          > http://www.egroups.com/group/evolutionary-psychology/info.html
                          > To contribute address your email to:
                          > evolutionary-psychology@egroups.com
                          >
                          > ------------------------------------------------------------------------
                          > -- 20 megs of disk space in your group's Document Vault
                          > -- http://www.egroups.com/docvault/evolutionary-psychology/?m=1
                          >
                          >
                        • Fredric Weizmann
                          Irwin, That answer is a little flip, so why don t we let there be one person with blue eyes who just migrated and joined the population. Voila, its a variable
                          Message 12 of 20 , Dec 3, 1999
                            Irwin,
                            That answer is a little flip, so why don't we let there be one person
                            with blue eyes who just migrated and joined the population. Voila, its a
                            variable (but heritability is still close to zero.)

                            Fredric Weizmann

                            On Fri, 3 Dec 1999, Irwin Silverman wrote:

                            >
                            > > Heritability of the eye colour is zero in a population of individuals all
                            > > having brown eyes
                            >
                            > This seems pointless - If eyes were all one color then there
                            > wouldn't be variance to measure and it would not be a variable.
                            >
                            >
                            > ------------------------------------------------------------------------
                            > To subscribe/unsubscribe/select DIGEST go to:
                            > http://www.egroups.com/group/evolutionary-psychology/info.html
                            > To contribute address your email to:
                            > evolutionary-psychology@egroups.com
                            >
                            > ------------------------------------------------------------------------
                            > -- Easily schedule meetings and events using the group calendar!
                            > -- http://www.egroups.com/cal?listname=evolutionary-psychology&m=1
                            >
                            >
                          • Irwin Silverman
                            ... Stated more precisely, the estimate is restricted by the variance in evironmental conditions affecting the trait inasmuch as genetic variance between
                            Message 13 of 20 , Dec 3, 1999
                              On Fri, 3 Dec 1999, Peter Kabai wrote:

                              > So when discussing what heritability really means,
                              > besides being critical about the methods of measuring the trait, one also
                              > should keep in mind, that the estimate is restricted to the population in
                              > which it was measured.

                              Stated more precisely, the estimate is restricted by the variance
                              in evironmental conditions affecting the trait inasmuch as genetic
                              variance between relatives may be assumed to be constant across populations.
                              Thus, heritabilities are generally reported in terms of a range of estimates.
                              For example, studies carried out in London, Ontario have found
                              heritabilities based on co-twin studies in the area of .70 for various
                              personality traits while the Minnesota studies, based on a much broader
                              geographical base and hence a wider scope of environments, find estimates
                              of about .50 for similar traits. It is an intrinsic limitation of the
                              method but no biggie relative to the intrinsic limitations of most other
                              behavior science methods - that is, if it is understood and applied
                              appropriately.

                              > Heritability of the eye colour is zero in a population of individuals all
                              > having brown eyes

                              This seems pointless - If eyes were all one color then there
                              wouldn't be variance to measure and it would not be a variable.
                            • I. Silverman
                              ... No, the heritability is unmeasurable ... methods for estimating heritability require samples comprising relatives whose biological relationships
                              Message 14 of 20 , Dec 3, 1999
                                On Fri, 3 Dec 1999, Fredric Weizmann wrote:

                                > Irwin,
                                > That answer is a little flip, so why don't we let there be one person
                                > with blue eyes who just migrated and joined the population. Voila, its a
                                > variable (but heritability is still close to zero.)

                                No, the heritability is unmeasurable ... methods for
                                estimating heritability require samples comprising relatives whose biological
                                relationships (proportion of genetic communality) are known, which can
                                be divided into those raised in the same vs. different households from
                                birth (with the exception of one method comparing fraternal to identical twins
                                raised in the same households), and which, after such divisions are made,
                                yield sample sizes adequate for separate correlational analyses. By the
                                time enough blue eyes join the population, marry and have offspring of
                                which sufficient numbers are adopted out so that the conditions above are
                                met, we will have the makings of a heritability study and an appropriate
                                estimate of same for eye color.
                              • Paul Barrett
                                ... From: Fredric Weizmann To: Peter Kabai Sent: 03 December 1999 16:39 Subject: [evol-psych] Re: Iq genetic or
                                Message 15 of 20 , Dec 3, 1999
                                  ----- Original Message -----
                                  From: Fredric Weizmann <weizmann@...>
                                  To: Peter Kabai <pkabai@...>
                                  Sent: 03 December 1999 16:39
                                  Subject: [evol-psych] Re: Iq" genetic or environmental?

                                  > Insofar as the heritability of intelligence is concerned it has "bounced"
                                  > between .2 and .8 in various studies.
                                  ......
                                  > changed in this regard,) however, citing a midrange value is no more
                                  > defensible than citing an extreme one. Irwin Silverman says its somewhere
                                  > betweem .5 and.7. Even if one accepted that the chimera
                                  > of a "true" heritability lay within this range, it would still reflects a
                                  > difference between accounting for 25% of the variance and 49%--so thats
                                  > not exactly a definitive claim.

                                  My understanding of heritability estimates is that they represent the
                                  proportion of phenotypic variance that can be accounted for by genetic
                                  effects. As such, they are already "estimates of explained variation" - and
                                  are referred to as genetic "correlations" in studies where environments are
                                  "non-shared". The figure (7.2a) on page 179 of Jensen's book "The g Factor)
                                  shows relative proportions of genetic and environmental effects.

                                  Regards .. Paul
                                  _____________________________________________________________________
                                  Paul Barrett Direct Tel: (44)-1555-841343
                                  email: p.barrett@... Hospital Tel: (44)-1555-840293
                                  CS2000: pbarrettx1@... Fax: (44)-1555-840024
                                  http://www.liv.ac.uk/~pbarrett/paulhome.htm

                                  Chief Scientist, The State Hospital, Carstairs, Scotland, ML11 8RP, UK
                                  Senior Research Fellow, Clinical Psychology, Liverpool University, UK
                                • Peter Kabai
                                  Hi, let me briefly respond to M. Hubey and Irwin Silverman. Peter Kabai: This basic model is additive, because variances are additive. Quite simple. M. Hubey s
                                  Message 16 of 20 , Dec 4, 1999
                                    Hi, let me briefly respond to M. Hubey and Irwin Silverman.

                                    Peter Kabai:
                                    This basic model is additive, because variances are additive. Quite simple.

                                    M. Hubey's response:
                                    THis is backwards. The variances are additive because the model is additive.

                                    It might be backwards, depending from where one looks at it. However, the basic
                                    model, analysis of variance was not invented by geneticists. It was just applied
                                    to a very simplistic model: a number of independent factors affect the trait
                                    additively. In any such model, the total variance of the sum of independent
                                    factors equals the sum of the variances of these factors.

                                    > Peter Kabai: Heritability of the eye colour is zero in a population of
                                    > individuals all
                                    > > > having brown eyes
                                    >
                                    > Irwin Silverman: This seems pointless - If eyes were all one color then there
                                    > wouldn't be variance to measure and it would not be a variable.

                                    Estimating heritability for a trait without variation is indeed pointless. Zilch
                                    heritability, however, does not mean that there were no genes involved in the
                                    development of such a trait. In a random sample of families not much variation
                                    could be found for bilateral symmetry, or the number of eyes, therefore
                                    heritability of those traits could not be estimated. On the other hand, genes
                                    involved in bilateral symmetry, the formation of eyes etc. have been mapped and
                                    their action is pretty well known. Indeed, traits under higher selection pressure
                                    tend to be under more precise genetic control and show less "heritability".

                                    I think it is important to know, that IQ even within the "normal" variation has
                                    measurable genetic variance. Whether it is 0.5 or 0.8 might be irrelevant, once
                                    we learn more about the mechanism. It might be a long way to go, seems that so
                                    far we search for the key, where there is light, and not necessarily where the
                                    key might be.

                                    I am refering to studies on neurotransmitters involved in learning and problem
                                    solving or metaanalysis on the correlation of brain size and some measure of
                                    intelligence or behavioural plasticty.

                                    Some self criticism here:
                                    Stewart, M.G., Kabai, P.et al (1996). The involvement of dopamine in the striatum
                                    in passive avoidance training in the chick. Neuroscience 70:7-14.
                                    Lefebvre, L., Gaxiola, A., Dawson, S., Timmermans, S., Rozsa, L. and Kabai, P.
                                    (1998). Feeding innovations and forebrain size in Australian birds. Behaviour,
                                    135:1077-1097.

                                    I agree, that questions on mechanim, such as posed by Roger D. Masters on the
                                    effect of lead exposure are highly relevant (and fortunately not hopeless to
                                    unswere). Not because the true heritability estimates might be some points lower
                                    than previously thought, but because learning about the actual effects give a
                                    clue for intervention on the level of the individual (hope or worry).

                                    Best wishes, Peter Kabai
                                  • H. Mark Hubey
                                    ... Even if there was a large enough sample there is an additional problem. If I=G+E is used when the correct form should have been I=G*E, then (1) the
                                    Message 17 of 20 , Dec 4, 1999
                                      Irwin Silverman wrote:
                                      >
                                      >
                                      > Regarding your last point, I have argued also that traits under
                                      > high selection pressure (directional selection in any case) will show less
                                      > heritability. (see my paper on r/k theory in Ethology and Sociobiology,
                                      > 1992), maintaining that the high heritability figures for I.Q. reflect, if
                                      > anything, the lesser influence of the attributes measured by I.Q. tests in
                                      > hominid evolution. But again, I do not see that this bears on the

                                      Even if there was a large enough sample there is an additional problem.
                                      If I=G+E is used when the correct form should have been I=G*E, then
                                      (1) the heritability coefficient does not measure the contribution of
                                      G but rather that of E, and (2) both G and E have to be measured on a
                                      ratio scale.

                                      This can be seen clearly in the last section (conclusion) of my paper
                                      the latest version of which can be found on my site.

                                      http://www.csam.montclair.edu/~hubey/ZIP/intelligence.pdf


                                      --
                                      Sincerely,
                                      M. Hubey
                                      Dept of Computer Science, Montclair State University
                                      hubeyh@... http://www.csam.montclair.edu/~hubey
                                    • Irwin Silverman
                                      ... Granted there must be an adequate distribution of a trait in order to assess its heritability estimate, but I do not view this as a critique of the method
                                      Message 18 of 20 , Dec 4, 1999
                                        On Sat, 4 Dec 1999, Peter Kabai wrote:

                                        > Zilch heritability, however, does not mean that there were no genes
                                        > involved in the development of such a trait. In a random sample of
                                        > families not much variation could be found for bilateral symmetry, or
                                        > the number of eyes, therefore heritability of those traits could not be
                                        > estimated. On the other hand, genes involved in bilateral symmetry,
                                        > the formation of eyes etc. have been mapped and their action is pretty
                                        > well known. Indeed, traits under higher selection pressure
                                        > tend to be under more precise genetic control and show less
                                        > "heritability".

                                        Granted there must be an adequate distribution of a trait
                                        in order to assess its heritability estimate, but I do not view this as a
                                        critique of the method where it can be applied. For example, the N for
                                        schizophrenia in the population is sufficient for heritability estimates to
                                        be obtained and these stand at present as virtually the sole evidence for
                                        a genetic influence in the disorder. On the other hand, there are probably
                                        insufficient cases of Munchausan Syndrome for a similar analysis, but I
                                        don't see where this impacts in any way on the schizophrenic data.
                                        Regarding your last point, I have argued also that traits under
                                        high selection pressure (directional selection in any case) will show less
                                        heritability. (see my paper on r/k theory in Ethology and Sociobiology,
                                        1992), maintaining that the high heritability figures for I.Q. reflect, if
                                        anything, the lesser influence of the attributes measured by I.Q. tests in
                                        hominid evolution. But again, I do not see that this bears on the
                                        validity of the heritability measure, properly applied. Heritability is an
                                        estimate of the relative influences of genetic and environmental factors on
                                        measureable trait variances in extant populations ... albeit a crude measure
                                        which will undoubtably, eventually give way to advances in genetics ... but
                                        the best we have right now and useful in many ways. For example,
                                        developmental psychologists who follow the prevailing social learning and
                                        role modeling theories should be very interested in the data on the lack of
                                        effects of shared environmental effects on personality traits. Piagetian
                                        scholars should appreciate the data on maturational level differences in
                                        heritability estimates for specific cognitive capacities, etc ...
                                        Regrettably the method has become inappropriately entwined in
                                        contentious sociopolitical issues, whether this is a fault of its users or
                                        its critics, and when this happens objectivity inevitably goes out of the
                                        window.

                                        Irwin Silverman
                                        Psychology Department
                                        York University
                                        4700 Keele Street
                                        Toronto, Ontario M3J 1P3
                                        Canada
                                        Ph. 416-736-2100 x66213
                                        fax 416-736-5814
                                      Your message has been successfully submitted and would be delivered to recipients shortly.