Discussion: View Thread

  • 1.  [Hrdiv_net] “Measuring Mainstream US Cu ltural Values”

    Posted 12-03-2012 17:57
    Some interesting points Michael. I'll address them as I find time.
    Romie

     
    Hope for the USA? "If something is unsustainable, it will stop."--Herb Stein, an economic adviser to Richard Nixon
    Romie F. Littrell, BA, MBA,PhD, FIAIR, An fánaí fiáin
    AUT Business School N.Z., romie.littrell@aut.ac.nz
    http://www.romielittrellpubs.homestead.com/
    Facilitator, Leadership & Management in Sub-Sahara Africa Conferences
    Contents copyright Romie F. Littrell

    From: "frone@ria.buffalo.edu" <frone@ria.buffalo.edu>
    To: Romie Littrell <littrellaom@yahoo.co.nz>
    Cc: GDO-AOM <gdo-l@aomlists.pace.edu>; HRDIV_NET <hrdiv_net@email.rutgers.edu>; hrdiv_net-bounces@email.rutgers.edu; IMD-list <imd-l@aomlists.pace.edu>; OB Listserv <ob@aomlists.pace.edu>; ORGCULT Listserver <ORGCULT-L@AOMLISTS.PACE.EDU>
    Sent: Tuesday, 4 December 2012, 10:31
    Subject: Re: [Hrdiv_net] "Measuring Mainstream US Cultural Values"


    Just an observation regarding an issue not in the limitations section:

    It would appear that generalizing to the U.S. population is important here, and you seem confident that this is possible.  On the 9th page of the paper, it is stated:

    "Comparisons with our data in Table 1 indicate our sample is adequately representative of the adult, White, non-Hispanic population in the USA."

    Yet a relatively small (by comparison to most national samples using probability sampling) convenience sample was used.  There was no probability sampling from a sampling frame that would reasonably represent some defined proportion of the US population, such as all White, non-Hispanics residing in some geographic region (lower 48) of the US.  

    Despite the attempt to suggest that most individuals in the US have access to the internet, this was not even the sampling frame.  The sampling frame was:

    "Data was gathered by collecting responses from a pool of potential subjects provided by four online retailers selling teas, coffee beans, crafts, toys, scarves, purses and jewellery."

    So one has a self-selected, convenience sample from the base of customers who shopped via the internet from four online retailers.  I have access to the internet and I am White and non-Hispanic, yet I'm quite sure that my probability of selection into this sample was zero.  So the sample may not be representative of even internet users in the U.S.  It may merely represent the 440 participants.

    Representativeness extends beyond the few demographic variables that were compared.  So for this study, the relevant issue is whether or not the values data represents the US population.  There is no way to know.   Because it is a convenience sample, there is no way to estimate the level of nonresponse (response rate)  or sampling error.  If the values data differ across internet users vs nonusers; or across customers of these 4 online-retailers vs people who are not customers; or more generally across those who had a non-zero probability of selection into the sample vs. those who had a zero probability of selection (which is almost everyone in the US.), then one cannot suggest that sample is representative of the population of interest.

    Even if the distribution of 20 demographic variables from this sample were very similar to the distribution in the US population, the only way this would be evidence that the values data were unbiased is a demographic or set of demographic variables were highly correlated with  each of the values variables (p-values are not important). In other words, the demographics would be strong proxies for the variables of interest.  This never happens, it didn't in this study, and logically it usually can't happen.  The typical approach of comparing the distribution of a few demographic variables in a sample to that in a population is generally meaningless for ruling out bias in the cental variables being analyzed.

    As nice as it would be, it is not that easy to claim with any reasonable level of certainty that one's data is representative of some population, especially a national population. Minimally one needs a probability sample drawn from a sampling frame that reasonably represents the target population.  Didn't happen here.  And one needs to be able to compute the response rate, sampling error, and provide a strong attempt at ruling out response bias.  Can't happen here.  So if the sample is representative of White, non-Hispanics in the US on the values variables, which we cannot know, it is really mere chance.  



    ****************************************************************
    Michael R. Frone, Ph.D.
    Senior Research Scientist
    Research Institute on Addictions
    State University of New York at Buffalo
    1021 Main Street
    Buffalo, New York 14203

    Office:    716-887-2519
    Fax:        716-887-2477
    E-mail:     frone@ria.buffalo.edu
    Internet:
    http://www.ria.buffalo.edu/profiles/frone.html
    ***************************************************************



    Romie Littrell <littrellaom@yahoo.co.nz>
    Sent by: hrdiv_net-bounces@email.rutgers.edu
    12/03/2012 02:56 PM
    Please respond to
    Romie Littrell <littrellaom@yahoo.co.nz>

    To
    IMD-list <imd-l@aomlists.pace.edu>, OB Listserv <ob@aomlists.pace.edu>, ORGCULT Listserver <ORGCULT-L@AOMLISTS.PACE.EDU>, HRDIV_NET <hrdiv_net@email.rutgers.edu>, GDO-AOM <gdo-l@aomlists.pace.edu>
    cc
    Subject
    [Hrdiv_net] "Measuring Mainstream US Cultural Values"





    New pub available via Springer's online early release for the Journal of Business Ethics; perhaps some insight into the recent national election results in the USA.
    "Measuring Mainstream US Cultural Values"
    Caroline Josephine Doran • Romie Frederick Littrell
    http://link.springer.com/article/10.1007%2Fs10551-012-1515-z
     
    Hope for the USA? "If something is unsustainable, it will stop."--Herb Stein, an economic adviser to Richard Nixon
    Romie F. Littrell, BA, MBA,PhD, FIAIR, An fánaí fiáin
    AUT Business School N.Z., romie.littrell@aut.ac.nz
    http://www.romielittrellpubs.homestead.com/
    Facilitator, Leadership & Management in Sub-Sahara Africa Conferences
    Contents copyright Romie F. Littrell
    _______________________________________________
    HRDiv_Net mailing list
    HRDiv_Net@email.rutgers.edu
    https://email.rutgers.edu/mailman/listinfo/hrdiv_net





  • 2.  [Hrdiv_net] “Measuring Mainstream US Cu ltural Values”

    Posted 12-03-2012 20:18
    Michael, the study provides convincing evidence with an adequate sample representative for the sub-category of US residents specified. Your comments, while appropriate for teaching a class from a textbook, are not really relevant for international field survey research with large samples, which I've been engaged in for the past 16 years across the world. I'll address your comments specifically when I get some spare time. I don't imagine many people will be interested, but if any are, please send me an email to my personal address and I'll include you. I suggest no further broadcast messages to the linkservers be sent.
    Rgds,
    Romie
    Hope for the USA? "If something is unsustainable, it will stop."--Herb Stein, an economic adviser to Richard Nixon
    Romie F. Littrell, BA, MBA,PhD, FIAIR, An fánaí fiáin
    AUT Business School N.Z., romie.littrell@aut.ac.nz
    http://www.romielittrellpubs.homestead.com/
    Facilitator, Leadership & Management in Sub-Sahara Africa Conferences
    Contents copyright Romie F. Littrell

    From: "frone@ria.buffalo.edu" <frone@ria.buffalo.edu>
    To: Romie Littrell <littrellaom@yahoo.co.nz>
    Cc: GDO-AOM <gdo-l@aomlists.pace.edu>; HRDIV_NET <hrdiv_net@email.rutgers.edu>; hrdiv_net-bounces@email.rutgers.edu; IMD-list <imd-l@aomlists.pace.edu>; OB Listserv <ob@aomlists.pace.edu>; ORGCULT Listserver <ORGCULT-L@AOMLISTS.PACE.EDU>
    Sent: Tuesday, 4 December 2012, 10:31
    Subject: Re: [Hrdiv_net] "Measuring Mainstream US Cultural Values"


    Just an observation regarding an issue not in the limitations section:

    It would appear that generalizing to the U.S. population is important here, and you seem confident that this is possible.  On the 9th page of the paper, it is stated:

    "Comparisons with our data in Table 1 indicate our sample is adequately representative of the adult, White, non-Hispanic population in the USA."

    Yet a relatively small (by comparison to most national samples using probability sampling) convenience sample was used.  There was no probability sampling from a sampling frame that would reasonably represent some defined proportion of the US population, such as all White, non-Hispanics residing in some geographic region (lower 48) of the US.  

    Despite the attempt to suggest that most individuals in the US have access to the internet, this was not even the sampling frame.  The sampling frame was:

    "Data was gathered by collecting responses from a pool of potential subjects provided by four online retailers selling teas, coffee beans, crafts, toys, scarves, purses and jewellery."

    So one has a self-selected, convenience sample from the base of customers who shopped via the internet from four online retailers.  I have access to the internet and I am White and non-Hispanic, yet I'm quite sure that my probability of selection into this sample was zero.  So the sample may not be representative of even internet users in the U.S.  It may merely represent the 440 participants.

    Representativeness extends beyond the few demographic variables that were compared.  So for this study, the relevant issue is whether or not the values data represents the US population.  There is no way to know.   Because it is a convenience sample, there is no way to estimate the level of nonresponse (response rate)  or sampling error.  If the values data differ across internet users vs nonusers; or across customers of these 4 online-retailers vs people who are not customers; or more generally across those who had a non-zero probability of selection into the sample vs. those who had a zero probability of selection (which is almost everyone in the US.), then one cannot suggest that sample is representative of the population of interest.

    Even if the distribution of 20 demographic variables from this sample were very similar to the distribution in the US population, the only way this would be evidence that the values data were unbiased is a demographic or set of demographic variables were highly correlated with  each of the values variables (p-values are not important). In other words, the demographics would be strong proxies for the variables of interest.  This never happens, it didn't in this study, and logically it usually can't happen.  The typical approach of comparing the distribution of a few demographic variables in a sample to that in a population is generally meaningless for ruling out bias in the cental variables being analyzed.

    As nice as it would be, it is not that easy to claim with any reasonable level of certainty that one's data is representative of some population, especially a national population. Minimally one needs a probability sample drawn from a sampling frame that reasonably represents the target population.  Didn't happen here.  And one needs to be able to compute the response rate, sampling error, and provide a strong attempt at ruling out response bias.  Can't happen here.  So if the sample is representative of White, non-Hispanics in the US on the values variables, which we cannot know, it is really mere chance.  



    ****************************************************************
    Michael R. Frone, Ph.D.
    Senior Research Scientist
    Research Institute on Addictions
    State University of New York at Buffalo
    1021 Main Street
    Buffalo, New York 14203

    Office:    716-887-2519
    Fax:        716-887-2477
    E-mail:     frone@ria.buffalo.edu
    Internet:
    http://www.ria.buffalo.edu/profiles/frone.html
    ***************************************************************



    Romie Littrell <littrellaom@yahoo.co.nz>
    Sent by: hrdiv_net-bounces@email.rutgers.edu
    12/03/2012 02:56 PM
    Please respond to
    Romie Littrell <littrellaom@yahoo.co.nz>

    To
    IMD-list <imd-l@aomlists.pace.edu>, OB Listserv <ob@aomlists.pace.edu>, ORGCULT Listserver <ORGCULT-L@AOMLISTS.PACE.EDU>, HRDIV_NET <hrdiv_net@email.rutgers.edu>, GDO-AOM <gdo-l@aomlists.pace.edu>
    cc
    Subject
    [Hrdiv_net] "Measuring Mainstream US Cultural Values"





    New pub available via Springer's online early release for the Journal of Business Ethics; perhaps some insight into the recent national election results in the USA.
    "Measuring Mainstream US Cultural Values"
    Caroline Josephine Doran • Romie Frederick Littrell
    http://link.springer.com/article/10.1007%2Fs10551-012-1515-z
     
    Hope for the USA? "If something is unsustainable, it will stop."--Herb Stein, an economic adviser to Richard Nixon
    Romie F. Littrell, BA, MBA,PhD, FIAIR, An fánaí fiáin
    AUT Business School N.Z., romie.littrell@aut.ac.nz
    http://www.romielittrellpubs.homestead.com/
    Facilitator, Leadership & Management in Sub-Sahara Africa Conferences
    Contents copyright Romie F. Littrell
    _______________________________________________
    HRDiv_Net mailing list
    HRDiv_Net@email.rutgers.edu
    https://email.rutgers.edu/mailman/listinfo/hrdiv_net