• download all data sets or
• click on the filenames to read about the data sets and download them separately.
• If not provided, the full reference can be found in the reference list of the Latent GOLD 4.0 Choice Technical Guide

Latent GOLD Choice Data Sets 1. cbc  3. brands  5. bank  7. ratingone.sav  7A. rating   8. aim.sav  9. capt  10. coffee  11. croon1.dat  12. croon3.dat

A. Latent GOLD Choice 4.0 Data Sets

• simulated data set used by Magidson and Vermunt (2004) to illustrate LC choice-based conjoint modeling
• 8 choice sets consisting of 3 alternatuves and a none option
• 3 attributes, and two covariates

• Participants in a choice-based conjoint study are administered six 3-alternative choice sets where each set poses a choice between alternative #1: Brand A (at a certain price), alternative #2: Brand B (at a certain price) and alternative #3: a None option. Brand A represents a new brand in this market

• full ranking data obtained from a real bank segmentation study as described in Kamakura, Wedel, and Agrawal (1994)
• Respondents were asked to rank 9 checking account alternatives (from most to least preferred)

• rating-based conjoint example
• simulated data
• full factorial design (2*2*2) with 8 replications
• LC regression with ordinal dependent, 3 predictors (product attributes) and 2 covariates (individual characteristics)

• The data for this example are obtained from a hypothetical conjoint marketing study involving repeated measures. Respondents were asked to provide likelihood of purchase ratings for each of 8 different product alternatives that differ on the attributes fashion, quality and price
• simulated data utilizing a 5-point ratings scale

• example used by Vermunt (2004) to illustrate the multilevel LC ranking model
• data taken from the 1999 European Value Study (EVS): subsample of 3584 individuals from 32 countries
• a single partial ranking task performed is by each person. It concerned the well-known (post)materialism scale in which people is asked : “Select what you find the first and the second most important aim of the country”
• the multilevel LC model involves grouping countries into country classes (GClasses). An alternative is model country differences using parametric random effects (GCFactors).
• the data file constain some covariates that were not used by Vermunt in his paper on multilevel LC models

• capture-recapture analysis using latent class or random effects models
• the model that is used is a log-linear model for a 2⁶ table in which the (0,0,0,0,0,0) combination (the number of non-captured cases) is missing
• this example illustrates that Latent GOLD Choice can be used to estimate mixture and random-effects log-linear models. The log-linear model used here is a quasi-independence model.

• data from choice-based conjoint experiment in which 185 persons coffee makers out of 8 sets consisting of 3 alternatives
• there are five product attributes of interest: brand (Philips, Braun, Moulinex), capacity (6 cups, 10 cups, 15 cups), price (F39, F69, F99), filter (yes, no), thermos flask (yes, no)
• the data can be modeled using a LC Choice model or random-coefficients choice model (the latter using the CFactors option)
• data used by Vermunt in his 2004 SMABS workshop.

• simple LC model for ranking data
• same (post)materialism scale as in “aim.sav”, but from a different study and as a full ranking task
• no attributes are used (only constants)

• same data as croon1.dat but in 3-file format
• we do not need an alternatives and a sets file because there are no attributes