Discussion: View Thread

  • 1.  Considering GMA testing

    Posted 09-15-1999 00:47
    This area of general cognitive ability and, dare I mention, shotgun
    personality testing, in pre-employment selection processes seems to be
    getting out of hand in many organisations. How many companies do you
    know that simply shuffle each applicant off to a testing centre for
    "some personality and IQ tests" without ever conducting a job analysis
    (I've encountered several). I would like to bring the issue of legally
    defensible testing to the attention of many managers. Does anyone know
    of an internet resource that covers the legal issues of pre-employment
    testing?


    Allan Tudehope
    The Enhancing Performance Group

    -----Original Message-----
    From: Industrial Psychology Forum [mailto:IOOBF-L@LISTSERV.UGA.EDU]On
    Behalf Of Phil Roth
    Sent: Wednesday, 15 September 1999 3:31
    To: IOOBF-L@LISTSERV.UGA.EDU
    Subject: Re: Considering GMA testing


    Hi Paul, Kevin, and Netters,

    I think Paul has a good point in noting that cognitive ability
    tests
    typically have a large difference in scores between Blacks and White as
    well as Blacks and Hispanics. If one looks at data on the Wonderlic for

    Blacks versus Whites, there is a d of about 1.00 associated with this
    test.
    Thus, Blacks on average tend to score about one standard less than
    Whites.
    However, the results are also somewhat moderated by complexity of the
    job.
    Again looking at Wonderlic scores, the d for low complexity jobs is
    usually in the mid .80's but drops to mid .70's for moderately complex
    jobs. Thus, some of the group differences may be somewhat smaller than
    one
    might think (given that most I/O folks think the d is 1.00).

    Interestingly, as cognitive ability differences drop for job
    complexity, I
    think validity increases (if I recall Hunter and Hunter, 1984
    correctly).
    Thus, the trade-off 's of validity versus likely adverse impact are less

    severe as one moves up the job complexity levels.

    So, bottom line, I agree with Paul that you are likely to have
    adverse
    impact, but it may be a little smaller than previously thought. Further,

    the nature of the job and its complexity may be a key factor in your
    thoughts. It really boils down to a management policy decision of
    trade-offs between a validity selection mechanism (and its associated
    increase in performance) versus adverse impact concerns.

    Phil Roth
    Clemson University

    a smarter work force At 11:33 AM 9/14/99 -0400, you wrote:
    >Kevin:
    >
    > Cognitive ability tests are likely to have adverse impact against
    >blacks and possibly hispanics (depending upon your location). For that
    >reason their use is not for the faint hearted. Before using them, I'd
    do
    >some sort of job analysis to determine if CA is job relevant for each
    >position you have in mind. I would also do a validation study of some
    sort
    >to be sure it works. If you can find existing validity studies in other

    >places, that would help too, and it might not make it necessary to do
    your
    >own, although I'd be more comfortable doing so. I would also look
    >carefully at the demographics of your applicant pool to get an idea how

    >many might be adversely affected. I don't mean to discourage you from
    >using them, but if you do be very careful in being sure they are job
    >relevant and have predictive validity, and that you can demonstrate
    these
    >things.
    >
    >Paul E. Spector
    >Department of Psychology
    >University of South Florida
    >Tampa, FL 33620
    >(813) 949-6427 Voice
    >(813) 974-4617 Fax
    >spector@chuma.cas.usf.edu
    >website http://chuma.cas.usf.edu/~spector
    >


  • 2.  Considering GMA testing

    Posted 09-15-1999 08:15
    Allan,

    Please keep me informed.


    ______________________

    Randall W. Kindley The Performance Group
    President 5215 45th Ave. S.
    V: 612-721-6752 Minneapolis MN 55417-2334
    F: 847-589-5231 www.topleaders.com
    kindley@dialupnet.com

    "Building High Performance Organizations by
    Developing People and Processes"

    > -----Original Message-----
    > From: Management Education and Development Discussion
    > [mailto:MG-ED-DV@MAELSTROM.STJOHNS.EDU]On Behalf Of Allan Tudehope
    > Sent: Tuesday, September 14, 1999 11:47 PM
    > To: MG-ED-DV@MAELSTROM.STJOHNS.EDU
    > Subject: Re: Considering GMA testing
    >
    >
    > This area of general cognitive ability and, dare I mention, shotgun
    > personality testing, in pre-employment selection processes seems to be
    > getting out of hand in many organisations. How many companies do you
    > know that simply shuffle each applicant off to a testing centre for
    > "some personality and IQ tests" without ever conducting a job analysis
    > (I've encountered several). I would like to bring the issue of legally
    > defensible testing to the attention of many managers. Does anyone know
    > of an internet resource that covers the legal issues of pre-employment
    > testing?
    >
    >
    > Allan Tudehope
    > The Enhancing Performance Group
    >
    > -----Original Message-----
    > From: Industrial Psychology Forum [mailto:IOOBF-L@LISTSERV.UGA.EDU]On
    > Behalf Of Phil Roth
    > Sent: Wednesday, 15 September 1999 3:31
    > To: IOOBF-L@LISTSERV.UGA.EDU
    > Subject: Re: Considering GMA testing
    >
    >
    > Hi Paul, Kevin, and Netters,
    >
    > I think Paul has a good point in noting that cognitive ability
    > tests
    > typically have a large difference in scores between Blacks and White as
    > well as Blacks and Hispanics. If one looks at data on the Wonderlic for
    >
    > Blacks versus Whites, there is a d of about 1.00 associated with this
    > test.
    > Thus, Blacks on average tend to score about one standard less than
    > Whites.
    > However, the results are also somewhat moderated by complexity of the
    > job.
    > Again looking at Wonderlic scores, the d for low complexity jobs is
    > usually in the mid .80's but drops to mid .70's for moderately complex
    > jobs. Thus, some of the group differences may be somewhat smaller than
    > one
    > might think (given that most I/O folks think the d is 1.00).
    >
    > Interestingly, as cognitive ability differences drop for job
    > complexity, I
    > think validity increases (if I recall Hunter and Hunter, 1984
    > correctly).
    > Thus, the trade-off 's of validity versus likely adverse impact are less
    >
    > severe as one moves up the job complexity levels.
    >
    > So, bottom line, I agree with Paul that you are likely to have
    > adverse
    > impact, but it may be a little smaller than previously thought. Further,
    >
    > the nature of the job and its complexity may be a key factor in your
    > thoughts. It really boils down to a management policy decision of
    > trade-offs between a validity selection mechanism (and its associated
    > increase in performance) versus adverse impact concerns.
    >
    > Phil Roth
    > Clemson University
    >
    > a smarter work force At 11:33 AM 9/14/99 -0400, you wrote:
    > >Kevin:
    > >
    > > Cognitive ability tests are likely to have adverse impact against
    > >blacks and possibly hispanics (depending upon your location). For that
    > >reason their use is not for the faint hearted. Before using them, I'd
    > do
    > >some sort of job analysis to determine if CA is job relevant for each
    > >position you have in mind. I would also do a validation study of some
    > sort
    > >to be sure it works. If you can find existing validity studies in other
    >
    > >places, that would help too, and it might not make it necessary to do
    > your
    > >own, although I'd be more comfortable doing so. I would also look
    > >carefully at the demographics of your applicant pool to get an idea how
    >
    > >many might be adversely affected. I don't mean to discourage you from
    > >using them, but if you do be very careful in being sure they are job
    > >relevant and have predictive validity, and that you can demonstrate
    > these
    > >things.
    > >
    > >Paul E. Spector
    > >Department of Psychology
    > >University of South Florida
    > >Tampa, FL 33620
    > >(813) 949-6427 Voice
    > >(813) 974-4617 Fax
    > >spector@chuma.cas.usf.edu
    > >website http://chuma.cas.usf.edu/~spector
    > >


  • 3.  Considering GMA testing

    Posted 09-15-1999 08:35
    Allan Tudehope wrote:

    >This area of general cognitive ability and,
    dare I mention, shotgun personality testing,
    in pre-employment selection processes seems
    to be getting out of hand in many organisations.<

    How many is many? I understand that pre-employment testing is a 3 billion
    dollar annual business in the US.

    >How many companies do you know that simply
    shuffle each applicant off to a testing centre
    for "some personality and IQ tests" without ever
    conducting a job analysis (I've encountered several).<

    The job matching assessment that we use, for instance, includes a general
    abilities section but the assessment is used only to decide which qualified
    applicant is best suited for the job. By the way, higher cognitive ability
    applicants are not always preferred especially if in combination with
    certain interests and personality traits.

    What percentage of employers use pre-employment testing?

    What percentage of employers that use testing do so incompetently?

    If an employer misuses a test, is that a problem with the test or the
    employer's use of the test?

    >I would like to bring the issue of
    legally defensible testing to the
    attention of many managers.<

    When managers use validated pre-employment tests in accordance with the
    test's instructions the managers are in a legally defensible position. At
    least that is what employment law experts advise us.

    >Does anyone know of an internet resource
    that covers the legal issues of
    pre-employment testing?<

    Which test? Should we presume that all pre-employment tests whether used
    correctly or not are legally indefensible? That will be a surprise to the
    large corporations that use testing. I know our 20,000+ clients would be
    surprised. Employers satisfy themselves on the legality before they start
    using pre-employment testing.

    The problem appears to be with the misuse of testing not with the testing
    itself. I know that some people recommend and/or use tests that were
    not designed for nor validated for pre-employment use.

    I suggest a telephone call to the company that supplies the test--not the
    reseller--to verify that the test is designed for and validated for pre-
    employment use. I was quite surprised when I did this and the two companies
    I called both told me that there tests were not for pre-employment use
    yet people on the Internet are recommending their tests for such use.

    Bob
    http://go.ourworld.nu/gately


  • 4.  Considering GMA testing

    Posted 09-15-1999 12:03
    Interesting reply Bob. By the way, I was not implying that all
    pre-employment tests are legally indefensible. What I will take issue with
    however, is the claim that "Employers satisfy themselves on the legality
    before they start using pre-employment testing." Perhaps employers the USA
    are more aware of the legal issues surrounding pre-employment testing than
    they are in Australia. Then again, perhaps not.

    If all employers conducted validation studies on the tests they require
    their applicants to undergo then perhaps one might say that they have placed
    their use of pre-employment screening tests in a legally defensible
    position. That is, they would be able to demonstrate a clear relationship
    between predictor and criterion in a court of law.

    My question (can of worms) is simply whether managers can afford to continue
    to use personality or CA tests without knowing what a validity coefficient
    or job analysis is? Furthermore, while it may be necessary to determine that
    a test (any test) has demonstrated robust validity and reliability in its
    development, is this sufficient to make such a test legally defensible? My
    own feeling is that unless a test instrument has been validated against a
    specific job role in the context of the organisation's climate, culture and
    strategy, the answer is no.

    I understand fully that pre-employment testing is big business and that many
    fortunes have been made on the premise that personality testing can select
    top performing employees. Despite being a psychologist myself, I believe
    that managers should not place unquestioning faith in such sacred cows.

    Allan Tudehope
    The Enhancing Performance Group
    Perth, Western Australia.


    -----Original Message-----
    From: Management Education and Development Discussion
    [mailto:MG-ED-DV@MAELSTROM.STJOHNS.EDU]On Behalf Of Robert Gately
    Sent: Wednesday, 15 September 1999 20:35
    To: MG-ED-DV@MAELSTROM.STJOHNS.EDU
    Subject: Re: Considering GMA testing


    Allan Tudehope wrote:

    >This area of general cognitive ability and,
    dare I mention, shotgun personality testing,
    in pre-employment selection processes seems
    to be getting out of hand in many organisations.<

    How many is many? I understand that pre-employment testing is a 3 billion
    dollar annual business in the US.

    >How many companies do you know that simply
    shuffle each applicant off to a testing centre
    for "some personality and IQ tests" without ever
    conducting a job analysis (I've encountered several).<

    The job matching assessment that we use, for instance, includes a general
    abilities section but the assessment is used only to decide which qualified
    applicant is best suited for the job. By the way, higher cognitive ability
    applicants are not always preferred especially if in combination with
    certain interests and personality traits.

    What percentage of employers use pre-employment testing?

    What percentage of employers that use testing do so incompetently?

    If an employer misuses a test, is that a problem with the test or the
    employer's use of the test?

    >I would like to bring the issue of
    legally defensible testing to the
    attention of many managers.<

    When managers use validated pre-employment tests in accordance with the
    test's instructions the managers are in a legally defensible position. At
    least that is what employment law experts advise us.

    >Does anyone know of an internet resource
    that covers the legal issues of
    pre-employment testing?<

    Which test? Should we presume that all pre-employment tests whether used
    correctly or not are legally indefensible? That will be a surprise to the
    large corporations that use testing. I know our 20,000+ clients would be
    surprised. Employers satisfy themselves on the legality before they start
    using pre-employment testing.

    The problem appears to be with the misuse of testing not with the testing
    itself. I know that some people recommend and/or use tests that were
    not designed for nor validated for pre-employment use.

    I suggest a telephone call to the company that supplies the test--not the
    reseller--to verify that the test is designed for and validated for pre-
    employment use. I was quite surprised when I did this and the two companies
    I called both told me that there tests were not for pre-employment use
    yet people on the Internet are recommending their tests for such use.

    Bob
    http://go.ourworld.nu/gately


  • 5.  Considering GMA testing

    Posted 09-16-1999 14:59
    Allan Tudehope wrote:
    (Hi Allan)

    >By the way, I was not implying that all
    pre-employment tests are legally
    indefensible.<

    Great, I think we need to be clear about that when we criticize pre-
    employment testing since any criticism tends to confuse and/or
    reinforce misconceptions about testing.

    >What I will take issue with however,
    is the claim that "Employers satisfy
    themselves on the legality before they
    start using pre-employment testing."<

    Notice I didn't write what they do, just that they satisfy themselves.

    >Perhaps employers in the USA are more aware
    of the legal issues surrounding pre-employment
    testing than they are in Australia.
    Then again, perhaps not.<

    Pre-employment testing is legal and recommended by the courts and Congress
    provided, of course, that the test is job related and predicts performance.

    >If all employers conducted validation studies
    on the tests they require their applicants to
    undergo then perhaps one might say that they have
    placed their use of pre-employment screening tests
    in a legally defensible position. That is, they would
    be able to demonstrate a clear relationship
    between predictor and criterion in a court of law.<

    Maybe the law is different in the US? There are four methods for
    validating a pre-employment test but a test need only satisfy one of
    the four. If current employees who are already successful in the job
    are used as the basis for hiring, that is one of the four methods.

    >My question (can of worms) is simply whether
    managers can afford to continue to use personality<

    I suggest personality only is sufficient.

    >or CA tests without knowing what a validity coefficient
    or job analysis is?<

    Yes, if it satisfies one of the four methods of validation. If the test
    was designed so that the user need not be an expert in testing then the
    user need not be an expert in testing. Of course, the law here does require
    that the test itself be developed in accordance with acceptable standards
    and is non-discriminatory.

    >while it may be necessary to determine that
    a test (any test) has demonstrated robust validity<

    How much is robust? Is there agreement on what it should be?

    >and reliability in its development,
    is this sufficient to make such a
    test legally defensible?<

    Only if it predicts job success, it doesn't have to be perfect.

    >My own feeling is that unless a test
    instrument has been validated against a
    specific job role in the context of the
    organisation's climate, culture and
    strategy, the answer is no.<

    I agree, but then again, that is how we use our job matching assessment.

    >I understand fully that pre-employment
    testing is big business and that many
    fortunes have been made on the premise
    that personality testing can select
    top performing employees.<

    If we select employees on personality alone we will make far too many
    mistakes since there is more to job success than just personality. If
    we select employees on general ability (IQ, SAT, GPA, Alma Maters, etc.)
    alone we will also make mistakes since job success depends on more than
    just knowledge and/or IQ. But what if we used both? We could identify
    applicants who not only have the personality to be successful but also
    the ability to be successful. If we assess for interests we can identify
    those applicants who are interested in the job, who can learn the job,
    and who can do the job as needed. Not a bad end result. It works well.

    >Despite being a psychologist myself, I believe
    that managers should not place unquestioning
    faith in such sacred cows.<

    That is what I tell my clients. Conventional interviews are poor
    predictors of job success while behavioral interviews are far more
    effective. Past performance in dissimilar jobs in different companies
    are poor predictors as well as college grades, references, etc.

    Bob
    http://go.ourworld.nu/gately


  • 6.  Considering GMA testing

    Posted 09-16-1999 15:33
    >I suggest personality only is sufficient. <

    I forgot to add "in" in front of "sufficient"

    I should have written "I suggest personality only is insufficient."

    Sorry,

    Bob


  • 7.  Considering GMA testing

    Posted 09-17-1999 11:34
    About the general application of cognitive ability and personality testing ...

    Allan Tudehope wrote <<..in part .. I snipped some off >>

    >
    > If all employers conducted validation studies on the tests they require
    > their applicants to undergo then perhaps one might say that they have placed
    > their use of pre-employment screening tests in a legally defensible
    > position. That is, they would be able to demonstrate a clear relationship
    > between predictor and criterion in a court of law.

    >
    >
    > My question (can of worms) is simply whether managers can afford to continue
    > to use personality or CA tests without knowing what a validity coefficient
    > or job analysis is? Furthermore, while it may be necessary to determine that
    > a test (any test) has demonstrated robust validity and reliability in its
    > development, is this sufficient to make such a test legally defensible? My
    > own feeling is that unless a test instrument has been validated against a
    > specific job role in the context of the organisation's climate, culture and
    > strategy, the answer is no.
    >
    > I understand fully that pre-employment testing is big business and that many
    > fortunes have been made on the premise that personality testing can select
    > top performing employees. Despite being a psychologist myself, I believe
    > that managers should not place unquestioning faith in such sacred cows.
    >
    > Allan Tudehope
    > The Enhancing Performance Group
    > Perth, Western Australia.
    >

    Schmidt and Hunters research on the generality of cognitive ability test
    validity
    addresses this concern.

    We (I/O psychologists and related co-conspirators)
    used to think that a CA test had a different validity for each group, and in
    each application. Therefore, employers should conduct their own validity
    studies, when they were poorly equipped to do so. The principal constraint
    they faced was small sample size (other technical concerns notwithstanding).

    Schmidt and Hunter showed that CA tests are consistently valid across
    groups and jobs (that is, we can safely reject r = .00). To say it in short,
    if there's thinking on the job, using a cognitive ability test to screen
    potential
    employees is a good idea. See Schmidt F, Hunter J in 1998 Psychological
    Bulletin review article.

    Well, CA score means DO vary across ethnicity in U.S. samples,
    (as noted early in this discussion)
    so using a CA test may have
    the effect of improving the employment chances of
    "asian" heritage and "european" heritage folks over those
    of "latino" and "african" heritage folks.

    In my (nonlawyer) understanding, if a test has job-related validity it
    can continue to be used by the employer even if it has the apparent
    disparate impact on ethnicity. Only if plaintiffs are able to locate and
    recommend a test that is also valid and that has less adverse impact than the CA
    test,
    will the employer be required to switch tests.

    More importantly, the validity generalization research has pushed us away from
    recommending each and every employer validity the test in that location because
    the small sample size results are not robust.

    So .... I don't care if all managers understand job analysis or validity.
    Instead I care that all managers know that current job information and valid
    testing are critical for serving the dual purposes of getting the best talent
    for
    the firm and keeping on the right side of the laws and courts. If they don't
    know the technical details of how that is done, that's where the human resource
    professionals and consultants earn their keep.

    On personality testing, I recall reading work by Dr. Leatta Hough of PDRI in
    Minneapolis, who (with many others) summarized personality research in
    relation to job performance. She suggested that some personality dimensions
    consistently correlated with job success at least for sales and managerial
    samples.
    Schmidt and Hunter's 1998 Psych Bull article adds to this by addressing the
    marginal
    validity of personality tests (and other kinds) beyond cognitive ability
    testing.

    So we should HOPE for widespread use of CA tests, and somewhat more selective
    application of personality testing. And, we should hope that managers know
    enough
    about these issues to at least recognize professional evidence and advice, if
    not the
    technical guts of validation research.

    None of us should place unquestioning faith in any one hiring tool, as the
    perfect
    tool does not exist. Instead, we build our confidence by using several,
    relatively
    independent, if individually flawed, predictors of success
    (again, discussed in Schmidt and Hunter, 1998).

    The generality of CA and personality tests for employment screening in the US is

    pretty darn good. It would be reasonable to expect that such tests would also
    work
    in other places where those aggressive Brits colonized, left English (the main
    test developer's language) , and democratic government/economic capitalism took
    hold.
    Certainly Australia.

    As to their generality in other settings - I still join Allan in his cautious
    skepticism.

    Cordially,


    Howard Miller
    professor and consultant

    Query: do the Australians have a legal process parallel to the one
    defined in the US by the Civil Rights Act (1964, 1991), the Uniform Guidelines,
    and
    the Griggs v. Duke Power (1971) Supreme Court decision?

    FYI, Cynthia Fisher and James Shaw of Bond University have an excellent
    HRM text book that touches on much of this (it's coauthored with Lyle Shoenfeldt
    of
    the U.S.)