Class Notes (808,204)
Canada (493,072)
Psychology (6,021)
Prof (1)

Measurement 2: Reliability and Validity

2 Pages
Unlock Document

Western University
Psychology 2800E

Week 6 ▯ Measurement 2: Reliability and   Validity  The Problem of Random and Systematic Error Prologue:  Example: Measuring your height - Value is 165cm. Is this an accurate (true) measure of your height? Random error (I misread number?; you slouched?) Systematic Error or Bias (wearing shoes?; end of tape measure was worn down a bit?; metal expands in a hot room, so if measured in a hot room you will appear to be shorter) - Observed Score = True Score + & - Random Error + or – Systematic Error (or bias)  Must minimize random and systematic error (so observed score = true score) How?  By maximizing the reliability and validity of measures Part 1: Reliability (Minimizing Random Error)  A. The “more is better” rule: random error will cancel out over repeated measurements Ex1: Beating Vince Carter in basketball - his true ability at basketball is better than yours - you can either play until one of you get ones basket or 20 baskets.. Whoever wins gets one million (20 basket game… I’m going to get lucky 20 times in a row, doesn’t work that way) – therefore, you take advantage of random chance by minimizing the number of opportunities, you take the game with one basket Ex2: Grandfathers who “can’t believe it” - One grandfather says “I have one grandchild and it’s a boy, can’t believe it” - Second grandfather says “I just can’t believe it… I have 20 grandchildren and they’re all boys”  ^ number of observations, so random error will cancel out B. We can decrease random error by ^ 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 items (exam) - is there consistency between the items? Inconsistency can be a sign of random error Assessing internal reliability:  calculate 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)  Calculate split-half reliability: (e.g. odd-even correlation) - high positive corr = low random error - best to use average of all split-halves: e.g. KR-20 (p.133) for measures with discrete values (T-F;MC questions), Cronbach’s Alpha for measures with continuous (or discrete) values - if either is l
More Less

Related notes for Psychology 2800E

Log In


Don't have an account?

Join OneClass

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

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.