\documentclass{article}
\setlength{\oddsidemargin}{0.0in}
\setlength{\textwidth}{6.5in}
\setlength{\textheight}{8.5in}
\setlength{\topmargin}{0.0in}
\begin{document}
\Large
\begin{center}
Psychology 407
Assignment \#2
\end{center}
Suppose we are testing four ways of teaching the oxidation-reduction
method of balancing chemical equations. We randomly assign 80
students to one of four groups (group sample sizes are all 20). The
dependent variable is the number correct in a test of 15 items.
\bigskip
\begin{tabular}{cc}
Method & Group Mean \\ \hline
1 & 10.0 \\
2 & 10.0 \\
3 & 12.0 \\
4 & 6.0 \\
\end{tabular}
\bigskip
\noindent In the usual analysis of variance table, the mean square
for error was 2.0.
\bigskip
a) Assume methods 1 and 2 are variations on one particular
``modern'' technique, method 4 is the ``traditional'' technique, and
method 3 is a combination of the ``traditional'' and ``modern''
techniques. Using this substantive information, what would be a
reasonable way to formulate the analysis in terms of planned
comparisons? Carry out your formulation using confidence intervals
and control the overall error rate with the Bonferroni method.
\bigskip
b) Assuming the overall F-test was significant, carry out a post-hoc
analysis using Scheff\'{e}'s method to answer the same questions you
investigated in (a) using planned comparisons. Compare the
confidence intervals under both approaches.
\bigskip
(c) Develop confidence intervals for all pairwise differences
between the means using the Tukey, Scheff\'{e}, and Bonferroni
methods. Compare the size of the resulting confidence intervals.
\end{document}