Measurement models for the Schwartz values.

Proposed a measurement model for the Schwartz Values Inventory (SVI) that takes into account method factors as well as systematic effects of the situation of measurement and of person-situation interaction. The SVI was administered to 224 Ss (aged 18–73 yrs) on 2 occasions, 6 wks apart. Traditional latent trait models (LTRMs) and latent state-trait models (LSTMs) with method factors were specified for each of 10 value scales (VSs) and were tested with LISREL. The LTRM was inadequate as a measurement model in all cases. For 9 of the 10 VSs, a LSTM with method factors fit the data best. For 1 VS, a LTRM with method factors was acceptable. The largest proportion of variance in individual value priorities was due to trait differences. Also, for 9 value types, small but significant proportions of variance reflected systematic effects of the particular occasion of measurement. Much of the variance in the VSs was due to method effects.