Lecture 10
MKF2121
1
Lecture 10
Correlation and Regression
Department of Marketing
Dr. Junzhao Ma
MKF2121
Marketing Research Methods
JM1
Today’s agenda
2
•
Quick
review
of statistical tests covered to date
–
Different tests and what they are for
–
What more can you learn from SPSS output (things are not
specifically stated in your hypotheses)
•
More statistical tests for hypothesis testing
– Correlation
– Regression
Taxonomy of different statistical tests
5/8/2017
3
Hypothesis
Testing
Test of Association
(relational RQ)
Test of Differences
(comparative RQ)
Means
Proportions
•
Different forms of t test
•
Cross tabulation
•
Correlation
•
Multiple regression

Lecture 10
MKF2121
2
Association btw 2 categorical variables
Test of association
Mean diff. btw 1 metric var. and constant
Something = 5
Mean diff. of 1 metric variable btw. 2 groups
Weight of males vs Weight of females
Mean diff. btw 2 metric variables
Attitude before vs Attitude after (same ind.)
Mean diff. of 1 metric var. among >2 groups
Prices of brands A,B and C
One sample T-test
Independent
samples T-test
Paired
samples
T-test
One-way ANOVA
Which test should you choose? – It depends
on your hypothesis
Cross-tabulation
(Chi-squared test)
Today’s agenda
5
•
Quick
review
of statistical tests covered to date
–
Different tests and what they are for
–
Interpreting SPSS output
•
More statistical tests for hypothesis testing
– Correlation
– Regression
Interpreting SPSS results - what is the
outcome of your test?
If the p
‐
value (Sig.) <
0.05
If the p
‐
value (Sig.)
≥
0.05
reject
H
0
, go to step 2
do not reject
H
0
.
State
H
0
in conclusion
Step 1
Step 2
clearly state the relationship between variables or the
differences between groups

Lecture 10
MKF2121
3
Cross-tabs example: soft drink preference
by age group (week 8)
7
•
190 customers are surveyed for their preference of
Coca-Cola and Pepsi (in paired comparison)
•
They also report their own age, which is re-coded into
two categories (“above 40” and “40 or under”)
•
Null Hypothesis: “Brand preference” has no
relationship with “Age”
SPSS output
8
P value
Strength of the relationship – this matters
only if P < 0.05
SPSS output (continued)
9
•
0 = Coca
‐
Cola
•
1 = Pepsi
Conclusion – younger adults are more likely to prefer Coca
‐
Cola over Pepsi than
older adults.
45.9% of the young adults
prefer Coke, compared to
30.4% of older adults
only when p < 0.05