It is already known that four attributes are important for laptop design: hard-disk size, processor speed, battery life, and selling price. The data gathered for this study are used to determine which of four laptop attributes (Hard Disk, Speed, Battery Life, and Price) are most important. It also assesses whether there are Gender or Job differences seen with these attributes.
1.
|
2.
|
3.
|
Click Select Data Table under Profile Data and select Laptop Profile.jmp. A partial listing of the Profile Data table is shown in Profile Data Set for the Laptop Example. The complete data set consists of 24 rows, 12 for Survey 1 and 12 for Survey 2. Survey and Choice Set define the grouping columns and Choice ID represents the four attributes of laptops: Hard Disk, Speed, Battery Life, and Price.
|
4.
|
Select Choice ID for Profile ID, and Add Hard Disk, Speed, Battery Life, and Price for the model effects.
|
5.
|
Select Survey and Choice Set as the Grouping columns. The Profile Data window is shown in Profile Data Dialog Box for Laptop Study.
|
6.
|
Click Response Data > Select Data Table > Other > OK and select Laptop Runs.jmp from the sample data library.
|
7.
|
Select Response as the Profile ID Chosen, Choice1, and Choice2 as the Profile ID Choices, Survey and Choice Set as the Grouping columns, and Person as Subject ID. The Response Data window is shown in Response Data Dialog Box for Laptop Study.
|
8.
|
9.
|
Choose Utility Profiler from the red triangle menu.
|
Results of this study show that while all the factors are important, the most important factor in the laptop study is Hard Disk. The respondents prefer the larger size. Note that respondents did not think a price increase from $1000 to $1200 was important, but an increase from $1200 to $1500 was considered important. This effect is easily visualized by examining the factors interactively with the Utility Profiler. Such a finding can have implications for pricing policies, depending on external market forces.
1.
|
Select Model Dialog from the Choice Model red triangle menu.
|
2.
|
3.
|
4.
|
Click Run Model.
|
Results are shown in Laptop Parameter Estimate Results with Subject Data, Laptop Likelihood Ratio Test Results with Subject Data, and Laptop Profiler Results for Females with Subject Data.
5.
|
Select Joint Factor Tests from the Choice Model red triangle menu
|
6.
|
Select Effect Marginals from the Choice Model red triangle menu.
|
The marginal effects of each level for each factor are displayed in Marginal Effects for Laptop. Notice that the marginal effects for each factor across all levels sum to zero.
The interaction effect between Gender and Hard Disk is marginally significant, with a p-value of 0.0744 (See Laptop Likelihood Ratio Test Results with Subject Data). In the Utility Profiler, check the slope for Hard Disk for both levels of Gender. You see that the slope is steeper for females than for males.
1.
|
From the Choice Model red triangle window, select Probability Profiler.
|
2.
|
3.
|
You can see in Laptop Probability Profiler Results with Baseline Effects that by increasing Hard Disk space, increasing Speed, or decreasing the price would increase the probability a subject would choose the new product over the old for females with development jobs.
You can see that by decreasing price, males with development jobs are actually less likely to choose the new product over the old. If you change the Job effect in the Baseline to Marketing, you see the same effect.
Looking at Laptop Probability Profiler Results with Baseline Effects, you can see that the steepest slope is for the Hard Disk factor.
5.
|
Change the Gender Effect in the Baseline to F and change the Hard Disk to 80 GB in the Probability Profiler.
|
You can change the subject effects to vary gender and job. The probability results are displayed in Laptop Probability Profiler Probabilities. You can see that males in development jobs are less likely than females in the same job to choose the new laptop, they are still more likely to choose the new laptop over the old. Developing a new product with increased hard disk space is more desirable to your customers than the old product.
1.
|
From the Choice Model red triangle window, select Model Dialog.
|
2.
|
Open the Subject Data section. Click Subject Data Table.
|
3.
|
Ensure that no data table is selected and click OK. By default, no data table is selected.
|
4.
|
Click Run Model.
|
5.
|
From the Choice Model red triangle window, select Multiple Choice Profiler.
|
6.
|
Click OK.
|
7.
|
For Alternative 2, set Hard Disk to 40 GB, Speed to 2.0 GHz, Battery Life to 6 hours, and Price to $1,200.
|
8.
|
For Alternative 3, set Hard Disk to 80 GB, Speed to 2.0 GHz, Battery Life to 4 hours, and Price to $1,500.
|
9.
|
By increasing hard disk space, you can increase the price of your laptop and expect a market share of about 36%. This share is almost the same as Company B’s high-performance laptop and is much better than the market share with the initial low-end settings seen in Multiple Choice Profiler with Laptop Data.
You can see in Multiple Choice Profiler with Improved Laptop that you could increase your market share even more by increasing Speed or Battery Life. If you wanted to determine what laptop configuration would give you the maximum market share, you could maximize desirability using the Choice Model red triangle menu options. In this case, however, it is fairly obvious that a high-performance laptop at a low price would be the most desirable.