The Global Motors Survey Differences Analysis

Case Questions

Your task is to apply appropriate differences analysis using the survey’s desirability measures from the GlobalMotors data file. Since you were not required to purchase the SPSS software I have completed the analyses and converted them to Word documents so you can access them.

You assignment is to develop a demographic profile for two of the five proposed models: the 1 SEAT ALL-ELECTRIC and the 4 SEAT GASOLINE HYBRID. You will find computer output files for each in the attached files. The variables that you must include are: Hometown size category, Gender, Marital status, Age category, Education category, and Income category. Only gender and marital status have 2 groups, so an independent samples t-test for was conducted for those variables and ANOVA’s were conducted for all others. For all ANOVAs a Duncan’s Post Hoc test was conducted to identify differences among groups. A 95% level of significance was used for the analysis. You have examples of output for these types of analyses in your text.

Case 14.2 The Global Motors Survey Association Analysis

1.) Use the TWO unique hybrid model demographic profiles that you developed in Case13.2 to determine whether or not statistically significant associations exist, and if they do, recommend the specific media vehicles for radio, newspaper, television, and magazines. Use the ZEN Motors advertising division’s preferred demographic for each medium.

This exercise requires you to revisit the demographic groups you found significantly different for each hybrid model style in Case 13.2. With Case 13.2, you identified groups within the various demographic variables that characterize each hybrid model’s target market. Here, you must determine the media preferences of those groups. I have run crosstabs for the demographic variables and the media preferences for each vehicle that are STATISTICALLY SIGNIFICANT associations. You must look at the percentages tables to determine the media vehicles that are preferred by each of the TWO model target groups you used for Case 13.2 (1 seat all-electric and 4 seat gasoline hybrid). If an association is NOT statistically significant it means that there is NO DIFFERENCE among the groups for that particular media. For example, for Radio preferences the variable GENDER is NOT statistically significant. That means that there was no difference between males and females regarding preferences for radio genres.

2. What is the life style profile of each of the possible target markets, and what are the implications of this finding for the placement of advertising messages that would “speak” to this market segment when the automobile model is introduced?

To obtain the data to answer this question a SPSS correlation analysis was completed. What you must do is to IDENTIFY the target Life Style group for EACH of the TWO proposed models that you have used in previous parts of this assignment (the 1 seat all-electric model and the 4 seat gasoline hybrid). The correlation analysis is located in one of the attached files.

CHAPTER 13 – TESTS OF DIFFERENCE

GLOBAL MOTORS CASE 13.2

1 SEAT ALL-ELECTRIC MODEL DATA – CORRECTED MARITAL STATUS

Home town size

ANOVA

Desirability: 1 Seat All Electric

Sum of Squares df Mean Square F Sig.

Between Groups 103.037 4 25.759 18.167 .000

Within Groups 1410.799 995 1.418

Total 1513.836 999

Post Hoc Tests

Homogeneous Subsets

Desirability: 1 Seat All Electric

Duncana,b

Size of home town or city N Subset for alpha = 0.05

1 2

100K to 500K 246 2.49

500K to 1 million 396 2.51

10K to 100K 190 2.55

Under 10K 40 2.63

1 million and more 128 3.48

Sig. .454 1.000

Means for groups in homogeneous subsets are displayed.

a. Uses Harmonic Mean Sample Size = 111.942.

b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed.

Gender

Group Statistics

Gender N Mean Std. Deviation Std. Error Mean

Desirability: 1 Seat All Electric Male 560 3.08 1.184 .050

Female 440 2.09 1.056 .050

Independent Samples Test

Levene’s Test for Equality of Variances t-test for Equality of Means

F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference

Lower Upper

Desirability: 1 Seat All Electric Equal variances assumed 2.293 .130 13.733 998 .000 .988 .072 .847 1.129

Equal variances not assumed 13.922 982.086 .000 .988 .071 .849 1.127

Marital Status

Group Statistics

Marital status N Mean Std. Deviation Std. Error Mean

Desirability: 1 Seat All Electric Unmarried 110 3.32 1.686 .161

Married 890 2.56 1.136 .038

Independent Samples Test

Levene’s Test for Equality of Variances t-test for Equality of Means

F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference

Lower Upper

Desirability: 1 Seat All Electric Equal variances assumed 32.819 .000 6.221 998 .000 .760 .122 .520 .999

Equal variances not assumed 4.599 121.530 .000 .760 .165 .433 1.087

Age

ANOVA

Desirability: 1 Seat All Electric

Sum of Squares df Mean Square F Sig.

Between Groups 355.843 4 88.961 76.439 .000

Within Groups 1157.993 995 1.164

Total 1513.836 999

Post Hoc Tests

Homogeneous Subsets

Desirability: 1 Seat All Electric

Duncana,b

Age category N Subset for alpha = 0.05

1 2 3 4

25 to 34 320 2.12

65 and older 75 2.40 2.40

35 to 49 440 2.77 2.77

50 to 64 145 3.06

18 to 24 20 6.05

Sig. .139 .050 .118 1.000

Means for groups in homogeneous subsets are displayed.

Education Level

ANOVA

Desirability: 1 Seat All Electric

Sum of Squares df Mean Square F Sig.

Between Groups 231.243 4 57.811 44.848 .000

Within Groups 1282.593 995 1.289

Total 1513.836 999

Post Hoc Tests

Homogeneous Subsets

Desirability: 1 Seat All Electric

Duncana,b

Level of education N Subset for alpha = 0.05

1 2 3

College degree 548 2.48

Post graduate degree 85 2.61

Some college 275 2.62

High School diploma 74 3.16

Less than high school 18 5.94

Sig. .525 1.000 1.000

Means for groups in homogeneous subsets are displayed.

a. Uses Harmonic Mean Sample Size = 57.941.

b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed.

Case 13.2 Example – Measures of difference

Here are some tips for you for Case 13.2

This chapter looks at measures of difference. That means that what you are trying to find out is whether there are differences between your various grouping variables (demographics) and the variable of interest (in this case the two proposed automobile models). You would also use this type of technique to see if there were differences among groups (say gender) and attitudes (say for example attitudes about global warming). Determining these differences is important in developing your strategies for a particular model for this case or for the global warming example you may find that women are really concerned about global warming and men are not (this is just hypothetical). So you would use this information for determining your target market and for development of communication strategies.

The statistical test that we are using is comparing differences among groups. If you have 2 groups (say males and females) you use an independent samples t-test. The “independent” part means that the responses of 1 group (males) do not affect the responses of the other group (females). When you have more than 2 groups you want to compare you use an Analysis of Variance (ANOVA) test. The statistical test for ANOVAs is the F test. So the output you have are the results of the ANOVAs for variables with more than 2 groups (like home town size) and t-tests for variables with two groups (like gender).

For the ANO VAs you look at the first table called “ANOVA.” Underneath that it lists they type of vehicle model. There are two ways to determine whether you have statistical significance. One is to calculate the value for the F test (which is 18.167 for the 1 seat all electric) and compare that to the critical value, which you look up in an F table (which you should have used in your statistics class). The other way, which is what you should use for case 13.2, is to look at the column labeled Sig. This means significance. So if you want to have a confidence level of 95%, your Sig. level should be <.05 because being 95% confident means that there is a 5% (or .05) chance that you your conclusion is wrong and 95% chance that it is right (i.e., consistent with the data). If your significance level is <.05 it means that there is a statistically significant DIFFERENCE among at least one of the groups and the other groups. In order to determine WHICH group or groups are different you do a Post Hoc Test, which in this case we are using Duncan’s Post Hoc test because it is easy to interpret. This is found in the next table with the differences identified by different columns. Since we are trying to identify target market characteristics you want those group(s) with the statistically significant HIGHER mean value. Those group(s) represent your target demographic for that demographic variable. So it might be one group or it could be more than one group.

For the t-tests you would look the first table (labeled Group Statistics) that shows the mean values for each group to determine which group has the higher mean value (first table). Then you have to look at the next table (labeled Independent Samples Test) and the two columns labeled Levene’s Test for Equality of Variances). If the column with Sig. is <.05 you assume that variances of each group are unequal and you look at the t-value (and can compare that to the critical value found in a t-table) for the “Equal Variances NOT assumed” row) OR look at the “Equal Variances assumed” row if the Sig. value in the Lenvene’s Sig. column is .05 or >. So, choose the appropriate row for equal or unequal variances and look at the Sig. column in the section entitled “t-test for Equality of Means. If is <.05 it means that the differences ARE significant. You would then look back at the first table (Group Statistics) to see which group had the higher mean and that group is your target demographic for that variable. If the Sig. value is .05 or > it means that there is NO significant difference between the groups, therefore both would be your target demographics.

I hope that helps.

The files on the 1 seat all electric and 4 seat gas hybrid models are posted in the Assignments tab.

THIS TABLE BELOW IS AN EXAMPLE OF A TARGET DEMOGRAPHIC FOR ANOTHER OF THE MODELS TO SHOW YOU WHAT YOU SHOULD COME UP WITH FOR YOUR TWO ASSIGNED MODELS.

1. 5-seat standard size gasoline

Demographic Factor Target market

Size of home town or city Under 10,000, 10,00 to 99,999,100,000 to 499,999, and

500,000 to 1 million

Gender Females

Marital status Married

Age category Between 35 and 49

Level of education Post-undergraduate degree

College degree

Job category All groups

Income category Between $75,000 and $124,999

$125,000 and higher

Dwelling type Single-family

Mobile Home

Condominium/Townhouse

Income

ANOVA

Desirability: 1 Seat All Electric

Sum of Squares df Mean Square F Sig.

Between Groups 186.817 4 46.704 35.019 .000

Within Groups 1327.019 995 1.334

Total 1513.836 999

Post Hoc Tests

Homogeneous Subsets

Desirability: 1 Seat All Electric

Duncana,b

Income category N Subset for alpha = 0.05

1 2

$50K to $74K 393 2.52

$125K and more 91 2.53

$75K to $125K 332 2.54

$25K to $49K 163 2.85

Under $25K 21 5.48

Sig. .114 1.000

Means for groups in homogeneous subsets are displayed.

a. Uses Harmonic Mean Sample Size = 71.124.

b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed.