23.1 Alternative to formative model

• Model with first-order constructs instead of measures, and each with single reflective measures

  • can’t estimate the loadings and unique variances of the reflective measures

• Model with first-order constructs instead of measure, but each with three reflective measures, and second-order construct also has two reflective measures.

  • sometimes, first-order construct can’t be measured multiple times (e.g., income, age)

Validity

Validity should be be established in the following order:

1. Construct Validity (also known as face validity, or concept validity): requires concepts and theory is strong

2. Convergent Validity: (e.g., AVE statistics >.5)

3. Discriminant/Divergent Validity:

4. Nomological Validity: (might also involve moderation).

5. Known Group Validity

Reliability

• Test-retest reliability

• Internal consistency reliability: standardized Cronbach $\alpha$ (> 0.7 if well establish, and .6 if newly developed)

• Composite construct reliability

Seminar paper in marketing

• Fornell and Larcker (1981)

• J.-Benedict E.  M. Steenkamp and Baumgartner (1998)

  • Propose method to assess measurement invariance in cross-national consumer research

References

Fornell, Claes, and David F. Larcker. 1981. “Structural Equation Models with Unobservable Variables and Measurement Error: Algebra and Statistics.” Journal of Marketing Research 18 (3): 382. https://doi.org/10.2307/3150980.

Steenkamp, Jan-Benedict E. M., and Hans Baumgartner. 1998. “Assessing Measurement Invariance in Cross-National Consumer Research.” Journal of Consumer Research 25 (1): 78–107. https://doi.org/10.1086/209528.