SingleFactor Experiments

1
Single-Factor Experiments

• What is a true experiment?
• Between-subjects designs
• Within-subjects designs

2
What is a true experiment?

• Experiment study in which researcher has
  complete control over all aspects of the study
• 2 Essential features of experiments
• 2 or more levels of an IV (control condition or
  group)
• Random allocation of subjects to groups (if
  between-subjects)

3
Some Terminology

• IV what the experimenter manipulates (varies)
  in an experiment the hypothesized cause
• DV what the experimenter measures to test the
  hypothesis in an experiment the hypothesized
  effect
• Factor IV
• Level condition treatment
• One value of an IV
• Control Variable value held constant
• Counterbalancing Variable
• Confounded Variable covaries with IV
• Random variable value is randomly varied

4
Between-subjects designs

• At least 2 conditions (groups)
• Control and Experimental Condition, or
• 2 or more levels of IV without control
  condition
• Each subject is assigned to only one condition
• Random assignment of subjects to conditions

5
Between-subjects Designswith Multiple Conditions

• More than 2 levels of the IV
• Randomly assign subjects to conditions
• Design a between-subjects experiment with
  multiple levels to test the hypothesis People
  work harder individually than they do in a group.

6
Within-Subjects Designs

• At least 2 conditions
• Control and Experimental Condition, or
• 2 or more levels of IV without control
  condition
• Each subject is assigned to all conditions
• Design a within-subjects experiment with multiple
  levels to test the hypothesis that people can
  perceive a light more quickly when their
  attention is focused where the light will appear.
  (Spatial Cuing)

7
Order and Sequence Effects

• Order effects
• Result from the position in which a condition
  occurs (first, second, third, etc)
• If the order of conditions is A-B, order
  effects on B would be the effects that result
  from B being in the second position in the list
• Example practice effects
• Sequence effects
• Result from which conditions precede or follow a
  condition
• If the order of conditions is A-B, sequence
  effects on B would be the effects that result
  from B following A
• Example color perception

8
Controlling Order and Sequence Effects

• Controlling Order Effects
• Counterbalance the position in which each
  condition appears
• ABC, CBA, ACB C occurs first, second, and third
  an equal number of times
• Controlling Sequence Effects
• Counterbalance what each condition follows
• ABC, BAC C follows A half the time, and follows
  B half the time.

9
Counterbalancing in Within-Subjects Designs

• Counterbalancing within subjects
• Useful when each level of the IV occurs multiple
  times for each subject
• Each subject gets all sequences of the conditions
  (ideally)
• Counterbalancing within groups (not within each
  subject)
• Useful when there are many (more than 2)
  conditions
• Necessary if each condition occurs only once per
  subject
• An equal number of subjects get each sequence of
  the conditions (ideally)

10
Controlling Order and Sequence Effects Within
Subjects

• Only possible when each condition occurs at least
  twice for each subject
• Method depends on how many times each condition
  occurs per subject
• Many times randomize order of conditions
• A few times use block randomization of
  conditions
• Twice use reverse counterbalancing

11
Controlling Order and Sequence Effects Within
Groups

• When it is not possible or practical to control
  within subjects
• Necessary if each condition occurs only once for
  each subject
• Sequences and orders controlled within a group of
  subjects, but not within each subject
• Latin Square a technique for partial
  counterbalancing (when full counterbalancing is
  impractical)

12
Latin Square

• A square matrix with length number of
  conditions
• Each condition occurs only once on each row and
  only once in each column
• Subject 1 ABCD
• Subject 2 BCDA
• Subject 3 CDAB
• Subject 4 DABC
• Controls order, but not sequence effects

13
Balanced Latin Square

• A Latin Square in which each condition is
  preceded by every other condition exactly once
• Subject 1 ABCD
• Subject 2 BDAC
• Subject 3 CADB
• Subject 4 DCBA
• Controls both order and sequence effects

14
In-Class Group Assignment

• Hypothesis Meditation improves short-term
  memory for lists of words
• Groups of 4
• Design an experiment to test this hypothesis
• Define your IV and DV
• State your prediction (hypothesized results)
• Write a plan for the condition(s) each
  participant will be in, and in what order
• Considerations
• Number of levels of IV?
• Within or between-subjects design?
• Random assignment to conditions?
• Controlling order and sequence effects?
• Other possible confounds?