Class Notes (837,550)
Canada (510,314)
Psychology (6,261)
Lecture 15

Lecture 15 - Within-group or within-subjects designs.docx

7 Pages
Unlock Document

Psychology 2800E
Doug Hazlewood

Within­Group (or “within­subjects”) Designs Prologue: Types of Experimental Designs A. Between-Group (or between-subjects) - Different groups get different conditions of IV. o E.g., 2 levels (or conditions):  Experimental group (subjects 1-10)  Control Group (subjects 11-20) o Or multiple levels (conditions), e.g.,  No treatment group (1-10)  Moderate treatment group (11-20)  High treatment group (21-30) B. Within-group or (within-subjects) - All participants exposed to all levels of IV o Experimental condition (subjects 1-10) o Control condition (same subjects 1-10) - Aka: “repeated-measures” design - Can also have “multiple conditions” if IV has more than 2 levels (all p’s will still be in all conditions) C. “Mixed-factorial” (or “mixed”) design - “factorial”: more than one IV (or “factor”) - “mixed”: 1 IV manipulated within subjects, 1 IV manipulated between groups - See next lecture (and Ch 11) for more on factorial experiments Part 1: Within­Group Designs (Conditions tested once) Example 1: Pretest-treatment-posttest design (aka: “one group pretest-posttest design” p. 263) - Better than: o One group posttest-only design (p.262) o Posttest-only with NE control groups (p.262) - BUT: has many threats to internal validity o Must add a “no treatment” control group  a “mixed” design • Pre vs. post is within subjects • Treatment vs. no treatment is between groups Example 2: Evaluating golf balls (“A” and “B”) - Type of ball is IV (each ball is a “treatment” or “condition”) - DV = distance balls travel - In between-group design, we’d have 2 groups with random assignment: o Group 1 hits ball A, Group 2 hits ball B o With 10 participants in each group, we’d need 20 participants - In within-group design, each participant hits both balls (e.g., A then B), number of participants is 10 Advantages of within-group design - Need fewer participants (e.g., 10 vs. 20) - Controls for variability from individual differences o Can be controlled in between-group designs  Use homogenous participants (e.g., “pro” golfers)  “Match” participants on golfing ability  Match on more things by using identical twins o Best to match person with self!  Removes variability caused by individual differences  Within-group design does this! Results for within-group experiment - Ball A travels 100 yards - Ball B travels 200 yards - Ball B is better? No! o New “confounds” A. Order Effect (Ball A first, Ball B second) - Order is confounded with treatment (ball) o Practice (warm-up) effect:  Hitting A first provided practice; caused B to go further? o Order can also produce “fatigue” effects - Solution 1: o Use a between-group design (but lose advantages of the within-group design) - Solution 2: o Complete Counterbalancing: Present conditions in all possible orders o Within 2 conditions, we have 2 “orders (AB, BA);  Include order as between-group factor (becomes “mixed design”) • But number of possible orders increases with number of conditions: “X!” o 2 conditions: X! = 2 x 1 = 2 o 3 conditions: X! = 3 x 2 x 1 = 6 o 6 conditions: X! = 720 possible orders - Solution 3: o “Partial” Counterbalancing: use a subset of the possible orders o Use a Latin square to control
More Less

Related notes for Psychology 2800E

Log In


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.