LIS 504 - Reliability and Validity

William Trochim illustrates reliability and validity with a target. Research that is reliable but not valid is like an archer who always hits about the same place but not near the bullseye. Research that is valid but not reliable is like an archer who hits various places centered around the bullseye, but not very accurately. Research that is neither reliable nor valid is like an archer who hits various places all off to the same side of the bullseye. Research that is both reliable and valid is like an archer who hits consistently close to the bullseye.
reliable not valid valid not reliable neither reliable nor valid both reliable and valid

Some types of reliability

Intra-observer reliability Does the same observer make the same observation at different times?
Inter-rater or inter-observer reliability Do different raters or observers agree on an observation or rating?
Test-retest reliability Do we get the same results when we test the same subject again?
Parallel forms reliability If essentially the same question is worded in different ways in two different questionnaires or interviews, do the two groups of respondents on average give similar answers?
Internal consistency reliability If essentially the same question is worded in different ways in a questionnaire or interview, does a respondent give consistent answers?

Some types and indicators of validity

Measurement validity Does a measure measure what it is supposed to measure?
Construct validity Does a measure make sense in the context of a theory?
External validity Can the results be generalized?
Face validity Does a measure seem "on its face" to represent the intended concept?
Internal validity Is an observed association causal in nature?

Home

Last updated July 4, 2001.
This page maintained by Prof. Tim Craven
E-mail (text/plain only): craven@uwo.ca
Faculty of Information and Media Studies
University of Western Ontario,
London, Ontario
Canada, N6A 5B7