Conjoint analysis is widely used to decompose holistic product evaluations into part-worth utilities, yet studies often rely on a single response format such as ratings. Alternative formats—including binary choices, paired comparisons, rankings, and categorical selections—are also common, but their coexistence has fragmented research and hindered cross-format comparisons. This study introduces an integrated structural equation modeling (SEM) framework that unifies diverse conjoint response formats within a single latent-variable representation. Building on the SEMWISE approach for rating data, the framework links structured preferences to observed responses through format-specific measurement models, allowing ratings, binary yes/no judgments, paired comparisons, rankings, and choice data to be analyzed coherently. A simulation study across six conditions (five single-format and one mixed-format design) demonstrates accurate parameter recovery and robustness, confirming that the unified preference structure can support estimation across heterogeneous data types. Overall, the proposed framework expands the methodological toolkit for conjoint research by enabling direct cross-format comparability, mixed-format analyses, and integration of evidence across studies.