Create a MaxDiff experiment from a file
As an alternative to offline data collection for conjoints, you can create an experiment directly from a file. In one operation, this creates a new MaxDiff experiment with configuration inferred from the file, registers the data as an offline source, and preserves participant IDs end-to-end so downstream exports carry the original IDs.
This option is only available for Generic MaxDiff with one attribute.
Preparing your .csv data file
You will upload a single .csv file containing the responses. Each row in the file represents all the information for one respondent.
The file needs to contain three groups of columns:
- Participant.
- Design columns: One column per item per set in the format:
{Question text} : MaxDiff Set {N}: {Item}. - Selection columns: Two columns per set recording the best and worst picks. Column name format:
(MaxDiff {block}) {keyword} : MaxDiff Set {N}.
Each column should follow the following naming convention and requirements:
| Column group | Requirements | Example column name | Value type | Example values |
|---|---|---|---|---|
| Participant | Any column name. Value represents unique respondent identifier. Between 1 and 255 characters. | Participant | String | 12345678, ABC-001 |
| Design | One column per item per set. {N} must be consecutive integers starting from 1. The number of shown alternatives must be constant across all sets. Value represents whether the alternative was shown in the choice set: 0 (not shown) or 1 (shown). | Which colour is most appealing? : MaxDiff Set 1: Blue , Which colour is most appealing? : MaxDiff Set 1: Green | Integer | 0 or 1 |
| Selection (Best pick) | {block} identifies the MaxDiff exercise (e.g. 1 or 2). {keyword} must be one of: Most Important, Best, First. Value represents which alternative was chosen as the best in the set | (MaxDiff 1) Most Important: MaxDiff Set 1 , (MaxDiff 1) Best: MaxDiff Set 1 | Integer | 1, 2, 3, … |
| Selection (Worst pick) | {block} identifies the MaxDiff exercise (e.g. 1 or 2). {keyword} must be one of: Least Important, Worst, Second. Value represents which alternative was chosen as the worst in the set | (MaxDiff 1) Least Important: MaxDiff Set 1 , (MaxDiff 1) Worst: MaxDiff Set 1 | Integer | 1, 2, 3, … |
See the example .csv file for a sample of how the data should be structured.
On Basic licence, you can only upload 20 offline respondents per experiment. On Professional licence and above, you can upload up to 20,000 offline respondents per experiment. If your file contains more respondents than the limit allows, the excess respondents will not be uploaded.
Uploading your data
Navigate to utilities from the landing page and select Upload offline data for MaxDiff.
Once you have uploaded your file, Conjointly will validate the data. Any validation errors will be displayed on the same page.
Upon a successful upload:
- The Blocks detected box will indicate the number of MaxDiff sets and number of alternatives per set inferred from the data file.
- Irrelevant columns will be discarded.
- You will be able to name the experiment and choose the default language during this stage.
Navigate to the Create experiment button at the bottom right to finish the process. Upon completion:
- A new offline source is created for the uploaded responses.
- You will be redirected to the report page.
- A confirmation message will confirm how many offline responses were uploaded.
- Participant IDs from your data file will be uploaded as an external variable labelled PARTICIPANT.