Classical Test Theory (CTT)

Classical Test Theory (CTT)

1. Methods of estimating the internal consistency of a test in CTT

   • ·split-half reliability; ·Spearman-Brown split-half estimate; ·the Guttman split-half estimate; ·Kuder-Richardson reliability coefficients; ·coefficient alpha; ·intra-rater reliability, also known as regrounding; ·inter-rater reliability
     • Steps to split the test and check reliabailty
       • Use all the methods of splitting, then estimate the reliability coefficient for any of them, and then find the average of these coefficients

2. Based on

   • Error scores are random in the sense that they do not correlate to true scores
   • An observed score is a function of both true and error score

3. Reliabilty

   • three different reliability models
     • Internal Consistency Estimates, Stability Estimates, and Equivalence Estimates
   • Hand in hand with validity
     • Language ability

4. Main offshoots

   • The ancillary assumptions invoked and experimental procedures followed in estimating coefficients of reliability
     • The standard errors of measurement.

5. Other concepts

   • The correction for attenuation, the Spearman-Brown formula, the reliability index, the Kuder-Richardson formulas, and the Guttman's lower bounds to reliability

6. Achievements

   • A recognition of the presence of errors in measurement., A conception of that error as a random variable, A conception of correlation and how to index it.

7. Most important quality of a test is its usefulness