Conjoint Analysis Question
Conjoint Analysis is a statistical technique used to measure consumer preferences for different products or services. This method breaks attributes down, such as price, brand, or quality. View a Conjoint Analysis example.
Creating a Conjoint Analysis Question
1. Choose Conjoint Analysis to add the question to your survey.
2. Enter the question.
3. Click on Conjoint Analysis Setting. There are three concept types to choose from:
Concept type 1 - Custom Concept: Multi-attribute Muti-horizontal Product
Upload data based on the provided data template.
Set up task setting, including the number of concepts per task (recommended 2-5 concepts, and the total number of concepts divided by the number of concepts per task should be an integer), whether to include an empty option and its text, and concept presentation rules. You can choose display all concepts, display all concepts in order, display some concepts, and display certain tasks.
Concept type 2 - Custom Concept: Simple Product
This concept type is similar to MaxDiff Analysis question. In Conjoint Analysis, you only need to select the 'best' label, whereas in MaxDiff Analysis, you need to choose both the 'best' and 'worst' labels.
You need to enter the concept code and upload the image for each concept. The minimum concepts are four.
Set up task setting, including the number of concepts per task (recommended 2-5 concepts, and the total number of concepts divided by the number of concepts per task should be an integer), whether to include an empty option and its text, and concept presentation rules. You can choose display all concepts, display all concepts in order, display some concepts, and display certain tasks.
Concept 3 - System-generated Combination Concept: Multi-attribute Muti-horizontal Product
You need to enter the attributes and horizontals in Attribute Setting.
Set up the task setting, including number of concepts per task (recommended 2-5 concepts), whether contains empty option and its text, and number of tasks.
You can also prohibited item combinations that the tasks will not show these combinations.
4. Click Save.
5. (Optional) Adjust any additional settings for questions.
6. Click Finish.
Analyzing the Data
Multi-attribute Muti-horizontal Product Report
The Conjoint Analysis report shows the attribute importance and conceptual utility.
- Attribute Importance: This table shows the importance of each attribute and level. Within the same attribute, the greater the utility value of a level, the more important that level is to the respondent. The greater the importance of an attribute, the more important that attribute is to the respondent. In here, CPU is the most important attribute for respondents.
Here is the formula for calculating the importance of an attribute:
Importance of an attribute = maximum level utility value for each attribute / sum(maximum level utility value for each attribute) x 100%
For example, the importance of CPU is 28.97% , it is calculated:
21.69 / (21.69 + 21.35 + 15.08 +10.82 + 5.93) x 100% = 28.97%
- Conceptual Utility: The preference ranking of a concept can intuitively show its importance. The higher the utility value, the more important the concept is to the respondents.
Simple Product Report
The simple product reports are presented using charts and tables to visualize the data. The following statistical analysis is provided for each attribute:
- Preference %: The percentage of times an attribute was selected as the "best" option in a task. A higher preference percentage indicates a more preferred attribute.
- Probability %: The likelihood that an attribute will be selected as the "best" option in a task. Probability scores range from 0 to 1, with higher scores indicating a higher likelihood of being chosen as the best option.
- P-Value: A p-value less than 0.05 is typically considered to be statistically significant.
- Selected Counts: The number of times an attribute was selected as the most important.
- Occurrence Counts: The number of times an attribute was displayed.
- Score: Selected Counts / Occurrence Counts. A higher score indicates a more important attribute for the respondents.
NOTE: The conjoint analysis requires a sufficient sample size to calculate relatively accurate data, so reports based on a sufficiently large sample size (number of responses > 100) are meaningful.