RELIABILITY

A test is seen as being reliable when it can be used by a number of different

researchers under stable conditions, with consistent results and the results not
varying. Reliability reflects consistency and replicability over time. Furthermore,
reliability is seen as the degree to which a test is free from measurement errors,
since the more measurement errors occur the less reliable the test (Fraenkel &
Wallen, 2003; McMillan & Schumacher, 2001, 2006; Moss, 1994; Neuman, 2003). In
the same way, Maree and Fraser (2004) ask how far the same test would produce
the same results if it was administered to the same children under the same
conditions. This helps the researcher and educator to make comparisons that are
reliable. The more errors found in an assessment the greater its unreliability, and visa
versa. Reliability is a very important factor in assessment, and is presented as an
aspect contributing to validity and not opposed to validity
Reliability is more challenging to determine when dealing with national- or state-
level assessments where you are trying to generalize to a much larger population. This
generalization is not necessary in locally developed assessments for use in a classroom or
at the district level. You are not sampling a population, but you are typically including the
entire population (your entire class or your entire grade level) in your testing. “Reliability
is of more concern on standardized or high stakes testing than they are in classroom
assessment. In a classroom, students' knowledge is repeatedly assessed and this allows
the teacher to adjust as new insights are acquired” (Moskal & Leydens, 2000, p. 7).
Decisions in the classroom, made on the basis of an assessment, can easily be changed if
they appear to be wrong. Therefore, reliability is not of the same crucial importance as in
large-scale assessments, where there is no turning back (Black, 1998).
There are three types of reliability that most state or national assessments may
utilize as a means to demonstrate the reliability of their assessment, all of which use
statistics to produce a reliability coefficient: test-retest (stability), alternate-form, and
internal consistency.