Course:AUDI522/goldman-fristoe presentation

From UBC Wiki

Reliability and Validity

Reliability

1. Internal reliability

High internal reliability -> test items are measuring the same same type of performance or content domain.

a) Coefficient alpha

  - Coefficient alpha is based on intercorrelation among all comparable parts of the same test.
  - The alpha reliabilities represent a measure of uniformity or homogeneity of items throughout the test and are calculated 
    using the variance of the test scores (i.e., number of errors) and the variance of the dichotomously scored items 
    (i.e., correct or incorrect production).

b) Standard errors of measurement (SEM)

  - SEM is used in the interpretation of individual test score.
  - Bands of error are used to construct confidence intervals for the obtained standard score.
  - When the individual's standard score on a test is banded by its SEM, there is a 68% chance that the individual's true 
    score falls within this band or interval.

Sounds-in-words has high internal reliability. Examiners can choose to use one of the the two common levels of reporting.

2. Test-retest reliability

Test-retest reliability is an index of temporal stability. It tells how much the individual's normative score might possibly change on retesting if a period of time has elapsed between testings.

Change could reflect the person's growth or fluctuation in the ability being measured or random differences in the performance.Change could also be due to differences in the examiner's consistency in perceiving and recording articulation errors.

  - Sounds-in-words was administered twice to 53 examinees who ranged in age from 4 years 6 months to 7 years 0 months at
    the time of the first testing.
  - All testing was done by the same examiner
  - Examinees were not aware that they would be tested again.

Sounds-in-words has high test-retest reliability.

3. Interrater reliability Sounds-in-words is highly dependent on the ability of the examiner to judge each sound production as correct or incorrect.

  - 2 examiners tested the same sample of 30 examinees twice.
  - Order of test was counterbalanced

Sounds-in-words has good interrater reliability.