Class Notes (1,100,000)
CA (620,000)
UOttawa (30,000)
PSY (4,000)
Lecture 10

PSY3307 Lecture 10: Reliability


Department
Psychology
Course Code
PSY 3307
Professor
Elizabeth Kristjansson
Lecture
10

This preview shows half of the first page. to view the full 2 pages of the document.
Reliability
In testing reliability refers to consistency and repeatability.
Measurement Error
-if something is unreliable then it has measurement error
-bathroom scales vary by 1 kg, baby scales in hospitals vary only 20 grams
-baby scales in hospitals are required to be more precise because they (1) weigh less and (2)
need to be precise for potential medical intervention
Reliability in measurement means it reflects true rather than chance factors.
You are able to count on it as a form of measurement.
Effecting scoring reliability:
-if to evaluators have different rubrics that would not be a reliable measure
Why do we care about reliability?
Because our test must accurately reflect our constructs.
Obtained Scores: total score gathered from the test
Error Score: how much of the data can be attributed to error
Reliability Coefficient: estimate of how errors effect your measurement, if this is low (0.2)
means it is inconsistent and there is a lot of error effecting your measurement.
Internal Consistency
Test-Retest: same test or questionnaire given twice over a long period of time (1 week, 6
months) shows how much the same the tests were
-must be done within an appropriate timeline to what you are testing
-three months to a few weeks is an appropriate timeline for attitudes
-six months may be effective for personality measures
What kinds of measures should not remain stable:
-weight
-mood
Practice Effects: with time you are able to ruminate or look up answers, you may do better on
the test the second time because you know what to expect and how to prepare
How long should an interval be between testing?
find more resources at oneclass.com
find more resources at oneclass.com
You're Reading a Preview

Unlock to view full version