Transcript Hair_12
12: Basic Data Analysis for
Quantitative Research
Statistical Analysis
Summary Statistics
Central tendency and dispersion,
Relationships of the sample data, and
Hypothesis testing
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved.
12-2
Measures of
Central Tendency
Mean
Arithmetic
Average
Mode
Response Most
Often Given
to a Question
Median
Middle Value
of a Rank Ordered
Distribution
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved.
12-3
Measures of
Central Tendency
Each measure of central tendency
describes a distribution in its own
manner:
for nominal data, the mode is the only
possible measure.
for ordinal data, the median is generally the
best.
for interval or ratio data, the mean is
generally used.
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved.
12-4
Measures of Dispersion
Describes how close to the mean or other measure
of central tendency, the rest of the values fall
Range
Distance between the
smallest and largest
value in a set
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Standard Deviation
Measure of the average
dispersion of the values
about the mean
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved.
12-5
SPSS Output for Measures
of Dispersion
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved.
12-6
Hypothesis Testing
Independent Samples
two or more groups
of responses that are
tested as though they
come from different
populations
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Related (Matched) Samples
two or more groups of
responses that are assume
to originate from the same
population
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved.
12-7
Univariate Tests of
Significance
Tests of one variable at a time
z-test
t-test
Appropriate for interval or ratio data
Test: “Is a mean significantly different
from some number?”
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved.
12-8
Univariate Hypothesis Test Using X16
Variable (Reasonable Prices)
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved.
12-9
Bivariate
Statistical Tests
Compare characteristics (means or
frequencies) of two groups or two
variables
Cross-tabulation with Chi-Square
t-test to compare two means
Analysis of variance (ANOVA) to compare
three or more means
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-10
Cross-Tabulation:
Ad Recall vs. Gender
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-11
Chi-Square Analysis
Chi-square analysis enables the
researcher to test for statistical
significance between the frequency
distributions of two or more nominally
scaled (i.e. “categorical”) variables in a
cross-tabulation table to determine
if there is any association
between the variables
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-12
SPSS Chi-Square Crosstab
Example
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-13
Comparing means
Requires interval or ratio data
The t-test is the difference between the
means divided by the average variability of
the two random means
The t-value is a ratio of the difference
between the two sample means and the
std error of the difference in means
The t-test tries to determine whether the
difference between the two sample means
is significant or whether it occurred by
chance
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-14
Comparing Two Means with
Independent Samples t-Test
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-15
Comparing Two Means with
Paired Samples t-Test
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-16
Analysis of Variance
ANOVA determines whether three or more
means are statistically different from each
other
The dependent variable must be either
interval or ratio data
The independent variable(s) must be
categorical (i.e. nominal or ordinal)
“One-way ANOVA” means that there is
only one independent variable
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-17
F-Test
The F-test is the test used
to statistically evaluate the differences
between the group means in ANOVA
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-18
Determining Statistical
Significance using F-Test
Total variance in dataset can be separated
into Between Group and Within Group Variance.
The larger the variance Between
Groups vs. Within Groups, the larger the F-Ratio.
The higher the F-Ratio, the more likely it is that
the Null Hypothesis will be rejected
and that the means are statistically different.
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-19
SPSS One-way ANOVA example:
Likelihood of Recommending vs. Gender
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-20
Follow-up Tests
ANOVA does not tell us where the
significant differences lie – just that a
difference exists
Pairwise Comparison Tests
Tukey
Duncan
Scheffe
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-21
SPSS Scheffe Test Example
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-22
n-way ANOVA
Appropriate for multiple independent
variables
Example: men and women are shown
humorous and non-humorous ads and then
attitudes toward brand are measured. IVs =
(1) gender, and (2) ad type; DV = attitude
toward brand
Need 2-way ANOVA design here (also
called “factorial design”) because we have 2
IVs
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-23
SPSS Example: 2-way ANOVA
Likelihood of Recommending vs.
(1) Gender & (2) Distance
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-24
SPSS Example:
Repeated Measures ANOVA
Does your version of SPSS have it?
Hair/Wolfinbarger/Ortinau/Bush, Essentials of
Marketing Research 1e © McGraw-Hill/Irwin2008
Copyright © 2008 by the McGraw-Hill Companies, Inc. All rights reserved. 12-25