Unbalanced cross-over Latin Square

From: JP (cest_raoul_at_yahoo.ca)
Date: 07/15/04

  • Next message: Lance Hoffmeyer: "Nonparametric tests with unequal group sizes" 
    Date: 15 Jul 2004 15:30:13 GMT

    Hi all,

    I need your advice on some analysis I am doing.

    We are testing 4 different products (outcomes are: taste, colour, etc.).
    The design is a cross-over.
    So the 8 patients get the 4 products.

    We used a latin square design, with 2 blocks. So each product is tasted
    twice in each position.
    (Carry over effect is not expected).

    Everything should be fine, it's a classic design.

    But something went wrong: one patient decided to stop and did not do the
    last experiement.
    We have now an unbalanced design.

    >From my understanding, I can not test the sequence effect anymore.
    How should I analyse that?

    Thanks for your help

    JP

  • Next message: Lance Hoffmeyer: "Nonparametric tests with unequal group sizes"

    Relevant Pages

    • Re: can I make 3 tabbed forms?
      ... Regarding the design of the database, ... which is filled in clinics and I want to ... Patients are categorized into two types, ... I want to input clinic name, then in the first subform, I want to input ...
      (microsoft.public.access.forms)
    • Re: Heprin Overdose Kills Two Infants - Time For a Return To Preparing Ones Own Meds
      ... Patients contracting hospital acquired infections from personnel not ... Rather, I'm saying that in addition to the five rights, ... there needs to be a system design that makes it less likely that ... "Whenever something like this happens, it is not an individual responsibility, it's an institutional responsibility," he said. ...
      (sci.med.nursing)
    • Experimental Design, proposal writing, data analysis and such
      ... machine on real patients and write a report to market this machine. ... design side. ... project, from proposal writing, experiemental design, data analysis and ...
      (sci.stat.math)
    • experimental design project proposal, data analysis and such
      ... machine on real patients and write a report to market this machine. ... design side. ... project, from proposal writing, experiemental design, data analysis and ...
      (sci.stat.consult)
    • Re: NIH Study
      ... > If Klempner put pcr-positive patients into a placebo group there are ... >> the PCR positives are proof of ongoing infection? ... Wasn't the PCR a feature of the design of the study? ...
      (sci.med.diseases.lyme)