Class Notes (809,569)
Psychology (6,041)
Lecture 5

# Lecture 5 - Reliability & Validity.docx

3 Pages
122 Views

School
Western University
Department
Psychology
Course
Psychology 2800E
Professor
Doug Hazlewood
Semester
Fall

Description
Reliability & Validity Prologue  Example: measuring your height  Value is 165 cm. is this an accurate (true) measure?  Random error (misreading number; slouching)  Systematic error or bias (wearing shoes?)  Observed score = True Score + & - Random Error + or – systematic error (or bias)  Systematic error systematically raises/lowers observed score from true score  Must minimize random and systematic error (so observed score = true score). How?  By maximizing the reliability & validity of measures Part 1: Reliability (Minimizing Random Error) A. The “more is better” rule  Random error will cancel itself out over repeated measurements  Example 1: Beating Vince Carter in basketball  Example 2: Grandfathers who “can’t believe it” B. We can decrease random error by increasing the reliability of our measures.  A measure is reliable if it measures things consistently C. Types of Reliability 1. Internal reliability (or internal consistency):  Relevant when measure consists of multiple item (e.g. exam)  Is there consistency between the items?  Inconsistency can be a sign of random error  Assessing internal reliability:  Item-total correlations (if random error is low, responses to any single item should be positively correlated with total score)  Eliminate items with low item-total correlations (and/or add more items)  Split-half reliability (e.g., odd-even correlation)  Combine all the odd number items together and all the even number items together, then correlate the average of the odd with the average of the even  High positive correlation = low random error  Best to use the average of all slit-halves:  E.g. KR-20 (p.133) for measure with discrete values (T-F; MC)  Cronbach’s Alpha for measures with continuous (or discrete) values  If either is low; internal reliability is low (too much random error in measure) 2. Test-retest reliability: measure produces same results consistently over time (aka “temporal stability”)  Take measure at two times; correlate  High correlation = high stability (rules out random error; which is unstable)  Note: only relevant when measuring stable properties of people/things that aren’t
More Less

Related notes for Psychology 2800E

OR

Don't have an account?

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Join to view

OR

By registering, I agree to the Terms and Privacy Policies
Just a few more details

So we can recommend you notes for your school.