User Guide

Chapter
18
Means
The Means procedure calculates subgroup means and related univariate statistics
for depende
nt variables within categories of one or more independent variables.
Optionally, you can obtain a one-way analysis of variance, eta, and tests for linearity.
Example. Measure the average amount of fat absorbed by three different types of
cooking oi
l and perform a one-way analysis of variance to see if the means differ.
Statistics. Sum, number of cases, mean, median, grouped median, standard error
of the mean, minimum, maximum, range, variable value of the first category of
the groupi
ng variable, variable value of the last category of the grouping variable,
standard deviation, variance, kurtosis, standard error of kurtosis, skewness, standard
error of skewness, percentage of total sum, percentage of total N, percentage of sum
in, perce
ntage of N in, geometric mean, and harmonic mean. Options include analysis
of variance, eta, eta squared, and tests for linearity R and R
2
.
Data. The dependent variables are quantitative and the independent variables are
categori
cal. The values of categorical variables can be numeric or short string.
Assumptions. Some of the optional subgroup statistics, such as the mean and standard
deviation, are based on normal theory and are appropriate for quantitative variables
with sym
metric distributions. Robust statistics, such as the median, are appropriate
for quantitative variables that may or may not meet the assumption of normality.
Analysis of variance is robust to departures from normality, but the data in each
cell sh
ould be symmetric. Analysis of variance also assumes that the groups come
from populations with equal variances. To test this assumption, use Levene’s
homogeneity-of-variance test, available in the One-Way ANOVA procedure.
343