MMPI II to Genomic Data Mapping

  • Posted by a hidden member.
    Log in to view his profile

    Aug 25, 2009 7:39 PM GMT
    Curious as to whether there are any Psychologist or Psychiatrist on RealJock. One of the non-profits I sit on has received a request to bring up a center/laboratory to attempt to map MMPI II results to genetic markers.

    As some of you may know, the MMPI II personality test is often used to evaluate inmates and is often part of the parole package that goes to the parole board. Since DNA is cataloged in most states for anyone entering the penal system, this is the proposed data set to be used to link DNA information to MMPI II data sets.

    Anyone know more about this?
  • cowboyathlete

    Posts: 1346

    Aug 25, 2009 8:44 PM GMT
    This sounds like an unbelievable minefield that brings up the never ending nature vs. nurture debate.
  • jrs1

    Posts: 4388

    Aug 25, 2009 9:26 PM GMT

    I would be very concerned with validity here. These personality tests are what is currently available, but perhaps a hybridized testing arrangement is in order.
    I'm not currently in the mood to do some digging of my own, but genetic markers and personality assessment/psychotherapy seem to show correlations rather than causality. This thread invokes topics concerning Raymond Cattell who was more concerned with biological traits rather than personality theories and has a 16-factor scale assessment.

    In terms of your non-profit agency, a longitudinal study where a prison cohort's genetic mapping paired with the aforementioned and suggested hybridized testing (e.g., Cattell's 16-factor trait scale combed into the MMPI II, A, RF, etc.) could be used to factor out a proposed series of null hypotheses and utilizing behavioral statistics (minding spurrious variables) to draw more reliable and valid results.
  • DanielQQ

    Posts: 365

    Aug 25, 2009 10:05 PM GMT
    What kind of information do you need to know? I'm a psychologist (ABD) specializing in personality assessment. The study sounds like a very interesting one, but i'm curious to know the goal of the study.
  • MSUBioNerd

    Posts: 1813

    Aug 25, 2009 10:17 PM GMT
    I'm not a psychologist (though one of my degrees is in psychobiology and I did work in a behavioral genetics psych lab for a while as an undergrad), so my perspective is more biological than psychological. However, I'd be concerned that such an attempt at mapping would be inherently hamstrung by the wide range of possible scores from the MMPI. Most screens for even complex traits with genetic, environmental, and interaction components (such as heart disease) are still fundamentally binary in whether the individual in question has the condition or not. When you get into personality traits such as are exhibited in the MMPI, where you have ten clinical scales (either in the 1989 revision or in the 2003 RC revision), plus the 10 validity scales and 5 further self-presentation scales.

    From a genetic standpoint, being able to do QTL mapping for a complex trait requires a very large number of individuals to even have a chance of breaking through to statistical significance. When working with data that isn't binary in nature, but instead a fuzzier numerical representation, y ou'll need an even larger number. Added to that, looking at a database of those entering into the penal system introduces a strong selection bias into your data set; you'd need a very large control sample in order to be able to tell whether the traits that seem to correlate with personality scores withint your population of those in the penal system are also good correlates when looking at the larger population of your country as a whole.
  • cowboyathlete

    Posts: 1346

    Aug 25, 2009 10:41 PM GMT
    DanielQQ saidWhat kind of information do you need to know? I'm a psychologist (ABD) specializing in personality assessment. The study sounds like a very interesting one, but i'm curious to know the goal of the study.
    Precisely. I was ABD in counseling psych before I changed careers. I am curious as to the non profit in question, and how they would interpret and use whatever comes of the analyses done.
  • Posted by a hidden member.
    Log in to view his profile

    Aug 25, 2009 11:06 PM GMT
    MSUBioNerd saidI'm not a psychologist (though one of my degrees is in psychobiology and I did work in a behavioral genetics psych lab for a while as an undergrad), so my perspective is more biological than psychological. However, I'd be concerned that such an attempt at mapping would be inherently hamstrung by the wide range of possible scores from the MMPI. Most screens for even complex traits with genetic, environmental, and interaction components (such as heart disease) are still fundamentally binary in whether the individual in question has the condition or not. When you get into personality traits such as are exhibited in the MMPI, where you have ten clinical scales (either in the 1989 revision or in the 2003 RC revision), plus the 10 validity scales and 5 further self-presentation scales.

    From a genetic standpoint, being able to do QTL mapping for a complex trait requires a very large number of individuals to even have a chance of breaking through to statistical significance. When working with data that isn't binary in nature, but instead a fuzzier numerical representation, y ou'll need an even larger number. Added to that, looking at a database of those entering into the penal system introduces a strong selection bias into your data set; you'd need a very large control sample in order to be able to tell whether the traits that seem to correlate with personality scores withint your population of those in the penal system are also good correlates when looking at the larger population of your country as a whole.



    You have a penal system data set of millions for DNA. A far smaller number for MMPI.

    How pliable do you think MMPI is over the long haul? In other words, does a persons MMPI change over their lifetime?
  • Posted by a hidden member.
    Log in to view his profile

    Aug 26, 2009 12:05 AM GMT
    MSUBioNerd saidI'm not a psychologist (though one of my degrees is in psychobiology and I did work in a behavioral genetics psych lab for a while as an undergrad), so my perspective is more biological than psychological. However, I'd be concerned that such an attempt at mapping would be inherently hamstrung by the wide range of possible scores from the MMPI. Most screens for even complex traits with genetic, environmental, and interaction components (such as heart disease) are still fundamentally binary in whether the individual in question has the condition or not. When you get into personality traits such as are exhibited in the MMPI, where you have ten clinical scales (either in the 1989 revision or in the 2003 RC revision), plus the 10 validity scales and 5 further self-presentation scales.

    From a genetic standpoint, being able to do QTL mapping for a complex trait requires a very large number of individuals to even have a chance of breaking through to statistical significance. When working with data that isn't binary in nature, but instead a fuzzier numerical representation, y ou'll need an even larger number. Added to that, looking at a database of those entering into the penal system introduces a strong selection bias into your data set; you'd need a very large control sample in order to be able to tell whether the traits that seem to correlate with personality scores withint your population of those in the penal system are also good correlates when looking at the larger population of your country as a whole.


    I would think they would focus on Pa and Pd on the clinical scale. Keep in mind incarceration does make some scores pop up on MMPI. You of course do see lots of hypomania in the prison population.

  • Posted by a hidden member.
    Log in to view his profile

    Aug 26, 2009 12:09 AM GMT
    jrs10k6 said
    I would be very concerned with validity here. These personality tests are what is currently available, but perhaps a hybridized testing arrangement is in order.
    I'm not currently in the mood to do some digging of my own, but genetic markers and personality assessment/psychotherapy seem to show correlations rather than causality. This thread invokes topics concerning Raymond Cattell who was more concerned with biological traits rather than personality theories and has a 16-factor scale assessment.

    In terms of your non-profit agency, a longitudinal study where a prison cohort's genetic mapping paired with the aforementioned and suggested hybridized testing (e.g., Cattell's 16-factor trait scale combed into the MMPI II, A, RF, etc.) could be used to factor out a proposed series of null hypotheses and utilizing behavioral statistics (minding spurrious variables) to draw more reliable and valid results.


    I have not made my way through the entire proposal, but I did notice the 16PF you cited is mentioned. Some of the data appears to be located outside the U.S., so who knows what was used.

    The main point is to create an organization to look at al of this.
  • DanielQQ

    Posts: 365

    Aug 26, 2009 12:11 AM GMT
    The MMPI-2 is a diagnostic assessment tool designed to aid clinicians in diagnosing psychopathology. There are already studies linking genetic markers to a variety of mental illnesses. I guess I'm wondering what value this study offers.
  • Posted by a hidden member.
    Log in to view his profile

    Aug 26, 2009 12:24 AM GMT
    DanielQQ saidThe MMPI-2 is a diagnostic assessment tool designed to aid clinicians in diagnosing psychopathology. There are already studies linking genetic markers to a variety of mental illnesses. I guess I'm wondering what value this study offers.


    I agree with you completely.

    So, if you were to have cash to trash, what direction would you go in if you wanted to develop predictive techniques of future psychopathology?
  • DanielQQ

    Posts: 365

    Aug 26, 2009 12:37 AM GMT
    I would continue to study DNA samples from clinically diagnosed individuals. Study the DNA of people with antisocial personality disorder, for example. Compare this DNA with samples from the "normal" population. If there are differences in DNA between the two samples, conduct a more focused analysis.

    This study seems to me a matter of laziness. It is easier to sample prisoners and give them some test vs. sampling people who have an actual clinical diagnosis.

    I would also educate the public on early childhood behavioral markers of psychopathology so that children can get treated early.

  • Posted by a hidden member.
    Log in to view his profile

    Aug 27, 2009 3:58 PM GMT

    Thanks guys. We've been able to devise a draft contract where we will fund a narrow range of things.