Nut209a

            All practice problems have .sav and .txt files on
            my website.

           due Monday, October 18, 2004

             sewage / raises / caries / kidney


           due Wednesday, October 20, 2004

             soilph / potency / cadmium / foodcost



            All of the problems due Monday involve independent samples.

            For the problems due on Wednesday, pay attention to whether
        the research involves independent or paired samples.  If paired,
        analyze differences/changes.  Change should be calculated as
        (After - Before), that is, the change from where one starts.

            Begin by looking at the data (within groups) to decide
        whether they are sufficiently long-tailed to the right to
        require a logarithmic transformation to bring them closer to our
        normal ideal.  If a transformation is required, include a
        statement along the lines of "Because the distribution of <name>
        was skewed to the right, a logarithmic transformation was
        applied to the data prior to formal analysis."

            Use SPSS to calculate a 95% confidence interval for the
        population mean in the case of a single sample or a 95%
        confidence interval for the difference between the population
        means in the case of two samples.  Give the resulting interval
        in a single sentence of simple English in terms of the original
        data.  For example, "We are 95% confident that the mean daily
        caloric intake of Tufts undergraduate males is between 525 and
        982 kcal greater than that of females."  If a logarithmic
        transformation was deemed necessary, transform the result back
        to the original scale and give a statement in terms of
        percentage differences.  The scores will be determined by the
        correctness of your statements, not by your ability to use SPSS.
        (Your ability to use SPSS will determine your effectiveness at
        producing the output needed to write proper summary statements.)

            The key elements are (1) the degree of confidence (in theory
        it could be anything, but in practice it's always 95%), (2) the
        confidence limits, and (3) the direction.  It's dangerous to
        say, "The difference between A and B" because some will
        interpret it as A-B, others will see B-A, and everyone else will
        just be confused!  If a 95% CI for males - females were to go
        from -15 to +23, I would say, "We are 95% confident that the
        mean value for males is between 15 units less and 23 units more
        than that of females."