Lecture 8: Experimental Design I: Single-Factor Designs - PDF

# Chapter 7. Experimental Design I: Single-Factor Designs ## Chapter Objectives - Identify and understand the defining features of the four varieties of single-factor designs - independent groups, matched groups, nonequivalent groups, and repeated measures - Describe two reasons for using more tha...

# Chapter 7. Experimental Design I: Single-Factor Designs ## Chapter Objectives - Identify and understand the defining features of the four varieties of single-factor designs - independent groups, matched groups, nonequivalent groups, and repeated measures - Describe two reasons for using more than two levels of an independent variable - Decide when to use a bar graph and when to use a line graph - Describe the goals of the Ebbinghaus memory research, his methodology, and the results he obtained - Understand the logic behind the use of three special types of control groups: placebo, wait list, and yoked - Understand the ethical issues involved when using certain types of control groups - Know when to use an independent samples t-test and when to use a dependent samples t-test, when doing an inferential analysis of a single-factor, two-level design - Understand why a one-way ANOVA, rather than multiple t-tests, is the appropriate analysis when examining data from single-factor, multilevel studies - Understand why post hoc statistical analyses typically accompany 1-factor ANOVAs for single-factor, multilevel studies ## Single-Factor Designs Decision Tree ### IV between-subjects or within-subjects? **Between** - IV manipulated or subject variable? - **Manipulated** - Forming equivalent groups by... - Random assignment - Matching - Possible matching to increase equivalence - Independent groups 1-factor design - Matched groups 1-factor design - Ex post facto 1-factor design - **Subject** - Groups intrinsically not equal **Within** - IV manipulated or subject variable? - **Manipulated by definition** - How often tested per condition? - **Once** - Complete/partial counterbalance - Repeated measures 1-factor design - **> Once** - Reverse/block counterbalance - Repeated measures 1-factor design ## Single-Factor - Two Levels ### Between-subjects, single factor designs - Independent groups designs - Manipulated independent variable - Random assignment to create equivalent groups - **Research Example 11** - IV? manipulated type of note-taking - Laptop note-taking - Handwritten note-taking - DV? performance on memory test - Other concepts: conceptual replication, what's next thinking, and ecological validity - Matched Groups Designs - Manipulated independent variable - Matching to produce equivalent groups - **Research Example 12** - IV? type of social skills training - Direct teaching - Play activities - Matching variable ? Autism Quotient - DV? Social Interaction Observation Code - Other concepts: operational definitions, double-blind procedure, inter-rater reliability - Ex Post Facto Designs - Subject variable as an independent variable - Deliberate attempts to select participants to reduce nonequivalence - **Research Example 13** - IV? whether or not traumatic brain injury (TBI) had occurred - Experimental group? had experienced TBI - Control group? no TBI - DV? ability to detect insincerity others - Other concepts: matching, external validity addressed ### Within-subjects, single factor designs - Also called repeated measures designs - Famous historical example ? Stroop (Box 7.1) - Used reverse counterbalancing - Manipulated independent variable - All Ss participate in all levels of the independent variable - **Research Example 14** - IV? whether or not you share your experience with another person - Shared - Unshared - DV? ratings of liking and of flavorfulness of chocolate - Other concepts: confederate, cover story, what's next thinking ## Single-Factor - More Than Two Levels ### Between-subjects, multilevel designs - Advantage #1 ? ability to discover nonlinear effects - **A graph** - X-axis: Arousal (weak to strong) - Y-axis: Performance (weak to strong) - Graph is a bell curve, with peak at Moderate Arousal - Left of peak represents Increasing attention and interest - Right of peak represents Optimal performance, and Impaired performance because of strong anxiety - Advantage #2 ability to rule out alternative explanations - **Bransford and Johnson's (1972) 'laundry study'** - One IV, three levels, independent groups - No context (no topic presented) - Context before (topic presented before reading paragraph) - Context after (topic presented after reading paragraph) - DV? recall of paragraph's ideas ### Multilevel independent groups design - **Research Example 15** - IV number of people with children - Alone - Bystander (with 2 other children who could help) - Bystander-Unavailable (with 2 other children who could not help) - DV? whether or not child (participant) helped teacher - Other concepts: operational definitions, confederates, inter-rater reliability ### Within-subjects, multilevel designs (continued) - **Research Example 13** - Multilevel repeated measures - IV = listening experience - Listening to Mozart -Listening to a rainstorm - Control - no listening - DV = recall of digits - Other concepts: counterbalancing via 3x3 Latin square, cover story ## Analyzing Data from Single-Factor Designs ### Presenting the data - Bar graphs versus Line graphs - **Bransford and Johnson's (1972) data presented in table and graphical forms** - **A table showing the means and standard deviations for three conditions:** - Condition - Mean Score - Standard Deviation - No Topic - 2.82 - 2.47 - Topic Before - 5.83 - 2.02 - Topic After - 2.65 - 2.18 - **A bar graph** - X-axis: Context (No Topic, Topic Before, Topic After) - Y-axis: Mean Idea Units Recalled (0-18) - Graph shows bars representing each condition - No Topic = 3 - Topic Before = 9 - Topic After = 3 ### Within-subjects, multilevel designs - Nonlinear results - **Ebbinghaus forgetting curve (Box 7.2)** - **A graph** - X-axis: Retention interval (.33 hr, 1 hr, 8.8 hr, 1 day, 2 days, 6 days, 31 days) - Y-axis: % saved (0-60) - Graph shows a line declining from left to right, with points at each interval - Analyzing single-factor, two-level designs - t test assumptions - Interval or ratio scale data - Data normally distributed (or close) - Homogeneity of variance - t test for independent samples, for - Independent groups designs - Nonequivalent groups designs - t test for related samples, for - Matched groups designs - Repeated measures designs - Analyzing single-factor, multilevel designs - Multiple t-tests inappropriate - Increases chances of Type I error - one-factor Analysis of Variance (ANOVA) - One-way ANOVA for independent groups, for - Multilevel independent groups designs - Multilevel ex post facto designs - One-way ANOVA for repeated measures, for - Multilevel matched groups designs - Multilevel repeated-measures designs - Once overall significant effect found, then post hoc testing - Comparing each level of IV against each other level ## Special-Purpose Control Group Designs ### Placebo control groups - Placebo - inactive substance - Ss think they are being treated but they are not ### Waiting list control groups - To insure equivalent groups in a study of program effectiveness - **Research Example 17** - IV? exposure to subliminal tapes - Experimental? weight loss tape - Placebo control? dental pain tape (told was weight loss tape) - Wait list control? no tape until wait was over - DV? weight loss (equal amount for all three groups) - **A graph** - X-axis: Group (Subliminal, Placebo, Waiting List) - Y-axis: Weight (lbs) (163-168) - Graph shows two bars per group - Baseline - Week 1-5 - Other concepts: pilot study, Hawthorne effect, double-blind procedure ### Yoked control groups - Each control group subject "yoked" to an experimental group subject - **Research Example 18** - IV? treatment for stress - Experimental ? EMDR therapy - Yoked control ? same instructions without eye movements - Each yoked participant matched with an experimental participant in terms of session length - DV? pretest-posttest changes in self-reported stress - Other concepts: Example of failure to reject null hypothesis that nonetheless has some value ## Summary - Single-factor designs have one independent variable (factor), which can be either a between-subjects or within-subjects factor. - Independent groups and ex post facto designs are between-subject designs and repeated measures are within-subjects designs. - Single-factor designs with more than two levels can demonstrate nonlinear effects and can the addition of more than two levels can rule out alternative explanations to the main result. - Data from single-factor designs can be presented in table or graphical form (bar graph or line graph). - Data from single-factor designs can be analyzed with t-test for single-factor, two-level designs, or with one-way ANOVAs for single-factor, multilevel designs. - Special-purpose control groups may be used to compare the effect of experimental treatments to no-treatment controls.