preview

Validity, Reliability And Generalizability

Decent Essays
Validity, reliability and generalizability

In the surveys and randomized controlled trials, validity, reliability and generalizability are three key factors.

Validity refers to that measuring tool or approaches can accurately measure things needed to be measured. It can be considered as an extent that measured results reflect investigative contents. Measured results more tend to be identical, validity will be higher, vice versa. Guba and Lincoln (1981) argued that whole social research must include invalidity in order to acquire worthwhile data within both the rationalistic paradigm (quantitative research) and naturalistic paradigm (qualitative research). Some factors can determine the level of validity, which include bias, construct
…show more content…
Epistemological criticism argues that all approaches are unlikely to produce consistent evidence because of the inherent property. It will generate a new viewpoint. However, through abundant reading, triangulation will enlarge width and depth in qualitative research and provide an analysis with security- a overall perspective. Participants can improve the validity of research. Bloor (1978) argues that sociologists can establish a correspondence with participants, in which participants can identify, give assent and judgement to sociologists.

Reliability refers to coherence, stability and dependability in test results, generally using internal consistency to express the levels of reliability in the test. The higher reliability indicates the higher level of accordance, stabilization and dependability in test results. Reliability is the precondition of validity (Guba and Lincoln, 1981). The same findings may not generate if the same research is repeated, because many influencing factors may work in the process of research. The process of establishment in reliability research includes: the research rigorously collect and explain data in consistent investigation (internal checks); the process is transparent (sample design, field work, inquiry and rational data). Patton (1987) suggests that the use of triangulation in multiple approaches can increase the reliability in results.

Generalizability can be
Get Access