The MaxDiff Simulator acts like a "voting machine" where respondents vote on a set of items that you include in a specific competitive scenario. Of course, we don't actually recontact respondents to re-ask their opinions; we use their individual-level utility scores to predict (simulate) how they would have voted if they had seen that specific set of items in the questionnaire.
Why use the Simulator instead of just refer to the average scores? If the items in your MaxDiff study are actual market offerings, then the Simulator can be more realistic and provide better predictions of market success. Sometimes, an item is well-liked (but not the most preferred) across respondents, but no market segment likes that item best. In that case, referring to average scores for the sample may not do a very good job of predicting actual market choices. The market simulator lets you put subsets of items in a competitive context (market scenario) and predicts how people would choose if those were the only items available.
Another handy outcome of the market simulator is that it makes the items included in the simulation scenario have shares of preference (percent of votes) that sum to 100%.
The items list lets you select which items to place in competition with one another in a simulation scenario. If you don't like the way the items are currently sorted in the list, use the icon to sort your items by original item number order, utility score, or label. Then, click which items to include in the competitive scenario.
In the example below, we have selected three flavors of ice cream for simulation within a competitive scenario (Chocolate, Rocky Road, and Chocolate Chip Mint).
The simulation acts like a voting machine, where respondents vote on which items they would prefer if these were the only items offered. The votes (shares of preference) sum to 100% across items included in the simulation scenario.
In the Simulator, respondent "votes" can be cast using one of two methods: "First Choice" or "Share of Preference."
oFirst Choice allows each respondent to cast just a single vote for the most preferred item,
oShare of Preference allows each respondent to split his/her vote based on the strength of the items' scores. The logit equation is used to split the votes, such that the votes sum to 100% for each respondent.
In general, we recommend you use the Share of Preference approach, as it typically does a better job leveraging your often limited sample size to stabilize the results, while recognizing that we rarely are 100% certain which item each respondent would vote for (choose).