Discussion: View Thread

  • 1.  Multiple comparisons: article for non-statisticians on when to adjust

    Posted 05-27-2016 17:42
    Hello colleagues,

    Do any of you have a 'popular' article on multiple comparisons?  A medical colleague of mine asked for an auricle or some material on multiple comparisons that is understandable (and referenc-able) by non-statisticians. He mentioned that in his academic department, whenever someone presents an analysis on more than one endpoint, there is always a question, "shouldn't you do a Bonferroni correction?" 

    I have technical material on this topic, but nothing suitable to send to him-ditto for my Google search. 

    There is some controversy in this area, but the material does not have to cover every aspect of this topic.  

    As an aside, I lean toward no adjustment, plus a frank discussion of the number and type of analyses carried out, adding my opinion of what it all means. We have used adjusted p-values in some cases (or adjusted confidence intervals), but those cases have been infrequent. Several times we have had to address and educate peer reviewers who had a knee-jerk "Bonferroni! Bonferroni!" reaction to tables that included multiple analysis and multiple p-values or confidence intervals. 

    Does anyone have something 'popular' that they can share?

    Thank you,

    Nayak



    Nayak L Polissar, PhD
    The Mountain-Whisper-Light Statistics
    1827 23rd Ave. East
    Seattle, WA 98112
    Tel. 206-329-9325
    Fax 206-324-5915
    polissar@u.washington.edu (for university affairs only) 






    ------Original Message------

    There's been extensive literature on the topic in the past 10-15 years. There are book-length treatments (http://www.amazon.com/Synthetic-Datasets-Statistical-Disclosure-Control/dp/1461403251; http://www.amazon.com/Statistical-Disclosure-Control-Anco-Hundepool/dp/1119978157); practices by federal agencies (https://www.census.gov/srd/sdc/); review articles (http://poq.oxfordjournals.org/content/76/1/163.abstract). Pick and choose.

    ------------------------------
    Stanislav Kolenikov
    Principal Survey Scientist
    Abt SRBI
    Education Officer, Survey Research Methods Section
    ------------------------------


  • 2.  RE: Multiple comparisons: article for non-statisticians on when to adjust

    Posted 05-27-2016 19:22

    Dr. Polissar,

    I wrote a humble overview that appears in the Encyclopedia For Research Design (2010), SAGE.  I have found this text to be handy, containing many great write-ups--written by an army of experts. 

    https://us.sagepub.com/en-us/nam/encyclopedia-of-research-design/book232149

    ------------------------------
    Randy Bartlett



  • 3.  RE: Multiple comparisons: article for non-statisticians on when to adjust

    Posted 05-27-2016 20:04

    I have something to offer that's 'popular' in the sense of being a review that appeared in a non-statistics journal. The PubMed entry for it is below:

    Multiple comparisons: philosophies and illustrations.

    Curran-Everett D.

    Am J Physiol Regul Integr Comp Physiol. 2000 Jul;279(1):R1-8. Review.

    PMID: 10896857 Free Article
    ------------------------------
    Eric Siegel, MS
    Research Associate
    Department of Biostatistics
    Univ. Arkansas Medical Sciences



  • 4.  RE: Multiple comparisons: article for non-statisticians on when to adjust

    Posted 05-30-2016 11:00

    Here are two references I have handy.

    Bender, R., & Lange, S. (2001). Adjusting for multiple testing—when and how? Journal of Clinical Epidemiology, 54(4), 343-349. doi:10.1016/S0895-4356(00)00314-0

    Sainani, K. L. (2009). The problem of multiple testing. PM&R, 1(12), 1098-1103. doi:10.1016/j.pmrj.2009.10.004

    ------------------------------
    Steven Pierce
    Associate Director
    Center for Statistical Training and Consulting, Michigan State University



  • 5.  RE: Multiple comparisons: article for non-statisticians on when to adjust

    Posted 05-30-2016 14:56

    Thank you very much, Steven, that is very kind of you. I will check out those references.

    Best wishes,

    Nayak

    ------------------------------
    Nayak Polissar
    Principal Statistician
    The Mountain-Whisper-Light Statistics

    Original Message


  • 6.  RE: Multiple comparisons: article for non-statisticians on when to adjust

    Posted 06-02-2016 10:08

    A great reference on multiplicity adjustment is given by  Hsu (2010).


    Here is a publisher link.

    http://onlinelibrary.wiley.com/doi/10.1038/clpt.2010.122/abstract

    onlinelibrary.wiley.com
    J C Hsu; Article first published online: 30 JUN 2010. DOI: 10.1038/clpt.2010.122 © 2010 American Society for Clinical Pharmacology and Therapeutics
    Qing Kang

    The Statistical Intelligence Group



    ------Original Message------

    Where?

    "This message was not found. Sorry, but the message you are replying to has been deleted or removed from our system or the egroup no longer exists."

    --

    *** Wayne ***

    *** Wayne G. Fischer, BS, MS, PhD ***

    "Boat-rocker, wave-maker; paradigm-shifter; paradox-breaker."©

    ASQ Certified Manager of Quality & Organizational Excellence

    ASQ Certified Quality Engineer

    ASQ Certified Quality Auditor

    Certified "Team Skills" Instructor and Facilitator

    Certified "Improving Processes" Instructor and Facilitator

    Statistician - University of Texas Medical Branch / Galveston, TX 77555-0752

    wgfische@utmb.edu / office = 409-747-1509 / cell = 281-360-7584

    "A man of conscience, is one who never acquires tolerance, well-being, success, public standing,