Year: 2005
Type: dissertation
Abstract: ions, so one can never know the actual amounts. However, by estimating error variance, researchers can plug it into the formula and get a glimpse of the true score (Springer et al., 2002a). CTT has two drawbacks. First, it relies on several assumptions that are frequently violated in real-world practice. Second, CTT scales developed have one gross amount of error which is derived from the sample and it is assumed that the error will be the same for other samples. Therefore, CTT is sample dependent. RELIABILITY In order for a scale to be useful, it must be reliable. Reliability simply means a measure produces the same results. Kerlinger’s (1986) analogy is of two shotguns used by an expert marksman. The first gun consistently hits outside of the range of the target and the second gun hits around a clustered area near the bulls-eye. The second instrument is said to be more reliable (Kerlinger, 1986). RELIABILITY THEORY. In classical test theory, the theoretical score is the combination of true score and error score (X=T+E). However, the result is no direct way to view observed score (Springer et al., 2002a). If error is defined as the difference between what is observed and the individual’s true score and all measures contain error, “[r]eliability is defined, so to speak, through error” (Kerlinger, 1986, p. 408). There are two types of error: systematic and random. Systematic error is a bias in the instrument that consistently produces incorrect answers (Rubin & Babbie, 1997).
Author: Tara L Alexander
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W1866420892