The actual focus of Analysis:

Developing a new methodological approach to proof content validity

 

Content validity is a concept of representative item sampling out of the universe of a valid item population (Fitzpatrick, 1983; Klauer, 1984). Because of the specific attributes of a Rasch scaled person parameter variance, we expect comparable Rasch Variances in any extraversion scales. Older strategies of testing about relationships between several measures are a strategy that leading mostly to weak statistical testing. Testing of equivalence (deviations) is a strategy that leads to strong statistical testing.

Methodological issues: What factors influence the True Rasch Variance?

  1. MEASUREMENT ERROR (Lord, 1983); Number of items
  2. IRT-MODELL (1PL, 2PL, Partial Credit, Graded, ...)
  3. ANSWERING FORMAT (Dichotome, Rating)
  4. ESTIMATING ALGORITHM (WLE, MLE, ...)
  5. LINK-FUNCTIONS (Logit, Normal-ogive)
  6. SOFTWARE (Winmira, Parscale, Bilog)
  7. BOTTOM-AND-CEILING-EFFECTS
  8. unknown factors

Especially point seven (Bottom and Ceiling Effects) had lead to own simulation studies, they offer the opportunity to correct a bias caused by bottom and ceiling effects.