MARKETING RESEARCH PART 17 pot

means or percentages of two different groups or samples.. Ch 17 21Using SPSS to Test the Difference Between Two Percentages • SPSS does not perform tests of significance of the differen

Testing for Differences Between Two Groups or Among More than Two

Groups

Ch 17 2

Why Differences are Important

within a market, there are different

types of consumers who have

different requirements, and these

differences can be the bases of

marketing strategies

Ch 17 3

Why Differences are Important

• Some differences are obvious –

differences between teens’ and baby boomers’ music preferences

• Other differences are not so obvious and marketers who “discover” these subtle differences may take

advantage of huge gains in the

marketplace

sample(s) may be assumed to exist

in the population(s) from which the random samples are drawn

Ch 17 5

Why Differences are Important

Market Segmentation

• Differences must be meaningful

marketing manager can potentially use as a basis for marketing

decisions

Ch 17 6

Why Differences are Important

Market Segmentation

• Differences should be stable

in place for the foreseeable future

• Differences must be actionable

can focus various marketing strategies and tactics, such as advertising, on the market

segments to accentuate the differences between segments

Ch 17 7

Small Sample Sizes:

The Use of a t Test or a z Test

• Most of the equations in this chapter will

lead to the computation of a z value.

• There are certain circumstances in

which the z test is not appropriate.

• The t-test should be used when the

sample size is 30 or less

• The t-test is defined as the statistical

inference test to be used with small

sample sizes (n is less than or equal to 30)

Ch 17 8

Determining Statistical Significance: The P value

• Statistical tests generate some

critical value usually identified by

some letter; i.e., z, t or F.

• Associated with the value will be a p

value which stands for probability of

supporting the null hypothesis (no

difference or no association)

• If the probability of supporting the null hypothesis is low, say 0.05 or less,

we have significance!

Ch 17 9

Determining Statistical Significance: The P value

• P values are often identified in SPSS with abbreviations such as “Sig.” or

“Prob.”

• P values range from 0 to 1.0

• See MRI 17.1 on page 491

Ch 17 10

Some Example P Values and

Their Meaning

• First, we MUST determine the

amount of sampling error we are

willing to accept and still say the

results are significant Convention is 5% (0.05), and this is known as the

“alpha error.”

means or percentages of two

different groups or samples

• Percentages are calculated for

questions with nominal or ordinal

level of measurement

• Means are calculated for questions with interval or ratio (metric level of measurement.)

Ch 17 13

Testing the Difference Between

Two Percentages

between the means being compared

difference between the compared

means

– Translate the difference into

number of standard errors from hypothesized value of 0

– Make an assessment of the

probability of support for the null hypothesis See formula…

Ch 17 15

Testing the Difference Between

Two Percentages (p 492)

• Formula for significance of the

difference between two percentages:

Ch 17 16

How do you know when the

results are significant?

• If the null hypothesis is true we would expect there to be 0 differences

between the two percentages

• Yet we know that, in any given study, differences may be expected due to sampling error

• IF the null hypothesis were true, we

would expect 95% of the z scores

computed from 100 samples to fall

between + and -1.96 standard errors

Ch 17 17

How do you know when the

results are significant?

• IF the computed z value is greater

than + or -1.96, it is not likely that the null hypothesis of no difference is

true Rather, it is likely that there is a real statistical difference between the two percentages

Ch 17 19

An Example: Testing the Difference Between Two Percentages (p 495)

• Last year a Harris Poll showed 40%

of surveyed companies were coming

to college campuses to hire seniors

• Since the z value is greater than + or

-1.96, the difference between the two

percentages is significant!

Ch 17 21

Using SPSS to Test the Difference

Between Two Percentages

• SPSS does not perform tests of

significance of the difference between the percentages of two groups, but

you can use SPSS to generate the

relevant information and perform a

hand calculation

• ANALYZE, FREQUENCIES will

produce the percentages you need

Ch 17 22

Testing the Difference Between

Two Means

• The procedure for testing the significance

of difference between two means from

two different samples is identical to the

procedure for testing two percentages

• Equations differ due to the use of a

metric (interval or ratio) scale

• Note: Only use this test with large

samples (30+)

Ch 17 23

An Example

• Sports Soft Drinks: the difference

between males (9) and females (7.5)

is significant; z =6.43.

Ch 17 24

Using SPSS to Test Differences Between Two Group Means

• The t-test is used to compare

differences between two means

(remember: “t for two”)

• But the types of t-test depends upon

whether the two groups upon which the means are calculated are

independent (separate groups) or

paired (the same group)

Ch 17 25

Using SPSS to Test Differences Between Two Group Means

• If the two groups are different, i.e.,

males vs females, you would use

INDEPENDENT SAMPLES t-test.

• If the two groups are from the same sample, you would use PAIRED

SAMPLES t-test.

Ch 17 26

An Example

• Is there a difference between

subscribers vs non-subscribers to

City Magazine on “likely patronage”?

– Since “likely patronage” is an

interval scale, we can calculate a mean score

– There are two independent groups: subscribers vs non-subscribers

Ch 17 27

An Example

– To determine if subscribers’ mean score on “likely” is different from non-subscribers’ mean on “likely,”

we should use SPSS:

INDEPENDENT SAMPLES T-TEST (See p 500.)

Ch 17 28

Ch 17 29

Ch 17 30

An Example

• Is there a difference between the

mean for “prefer simple décor” vs the mean for “prefer elegant décor”?

– Both “prefer simple décor” and

“prefer elegant décor” are intervally scaled so it is proper to calculate a mean for each question

– Under these conditions, it is

appropriate to use SPSS:

PAIRED SAMPLES T-TEST (See p

503.)

Ch 17 32

Ch 17 33

Ch 17 34

Online Surveys and Databases:

A “Significance” Challenge to Marketing

Researchers

• Sample size has a great deal to do

with statistical significance

• Sample size n appears in statistical

formulas dealing with differences,

confidence intervals, hypothesis

tests, etc

• Online surveys allow data collection

from large sample sizes, so most

tests may be found to be significant

• The difference should be meaningful

as well

Ch 17 35

Testing for Significant Differences Among More than Two Groups

• ANOVA

used when comparing the means

of three or more groups

– ANOVA will “flag” when at least

one pair of means has a statistically significant difference, but it does not tell which pair

Ch 17 36

Testing for Significant Differences Among More than Two Groups

– When the F values “Sig.” is less

than or equal to 0.05, ANOVA is telling you that “at least one pair of means is significantly different.”

– To determine which pair(s) are

different, you must rerun the test and select a POST HOC test

(Duncan)

Ch 17 37

Testing for Significant Differences Among More than Two Groups

– Assume that we wish to know if the

mean score on “likelihood of patronizing an upscale restaurant”

differs across sections of newspaper read most

– ANALYZE, COMPARE MEANS,

ONE-WAY ANOVA

– “Likely” goes in Dependent list;

“section of newspaper” goes in factor

Ch 17 38

Testing for Significant Differences Among More than Two Groups

– Output shows Sig Is 0.000

meaning at least one pair of means

is different

– Now rerun the ANOVA but select

Duncan under the POST HOC button

Ch 17 39

Ch 17 40

Ch 17 41

In Summary: Test of Differences Among More than Two Groups

• The basic logic

– ANOVA (Analysis of Variance)

– Test all pairs of averages

simultaneously

Ch 17 42

In Summary: Test of Differences Among More than Two Groups

– If no pair is different at the 95%

level of confidence, stop the analysis and say all pairs are

“Equal.”

– If at least one pair is different at the 95% level of confidence, make a

table to show what pairs are

“Equal” or “Unequal” by running post hoc test