STAT 401, FALL 2010 Eric Slud
Coverage, Topics and Problem Types for In-Class Test 2
======================================================
Coverage for Test 2 includes all material covered in class
and HW since Test 1. In the book, this covers :
Section 4.6 (probability plots)
including in-class and handout material on definitions
and interpretation of QQplots and empirical distribution
functions, and Weibull probability plots.
Section 9.5: F tests for ratios of variances based on
two-sample normal data.
Definitions of sample quantiles and boxplots, as covered in
class (and you can also find the boxplot definitions on
pp.35-38 in the text.)
Sections 10.1-10.2: one way analysis of variance, including
definitions of Total (or Corrected) Sum of Squares,
Sum of Squares Between groups (SSB or SSTr),
Sum of Squares Within groups (SSW or SSE),
F test statistic for equality of group means,
Tukey's studentized range statistic,
ANOVA Table.
Sections 12.1-12.2: simple linear regression, finding
relationships in scatter-plots and definitions of
least-squares equations for intercept and slope
coefficients, plus definition of residuals and
coefficient of determination.
Sections 14.1-14.2: goodness of fit tests based on multinomial
data, with or without estimated parameters, including
definition of chi-squared goodness of fit test, and
degrees of freedom depending on number of estimated
parameters.
===========================================================
Main problem types you could be asked on test:
DEFINITIONS: boxplots, empirical distribution function,
F_{n,m} distribution, Tukey studentized range statistic
and distribution, coefficient of determination in simple
linear regression.
BASIC CALCULATIONS: sample quantiles, calculation of chi-squared
statistic or simple linear regression coefficients,
filling in an ANOVA table based on given sample means and
variances for groups.
INTERPRETATIONS: ANOVA table, QQplot.
Hypothesis tests and CI's: for ratios of variances in 2-sample
normal data, test of fit of multinomial to specified vector
of probabilities; test of fit to a distribution like normal
to data grouped into class intervals; finding equivalent
"Tukey" groups in a one-way ANOVA layout.
============================================================
Rules for the Test, to be held in-class Wednesday, Dec. 1,
will be the same as for the previous test:
------------------------------------------------------------
One 8.5" x 11" notebook-page cheat-sheet allowed,
otherwise closed-book.
You should bring a calculator.
Most problems will require numerical expressions as answers, NOT
reduced arithmetic answers but also NOT formulas with
unevaluated symbols.