About this presentation
Individual differences in scale usage make it hard to identify signal from noise in segmentation studies based on choice data. The Scale-Adjusted Latent Class model (SALC) makes it easier to assign the right respondent to the right segment. We will investigate the effectiveness of the SALC model by showing examples from both artificial and real-life datasets. The SALC model will be compared to other, more common segmentation algorithms, both from a statistical perspective and from a practical, market relevance point of view.