Chapter 10: Experimental Design

Question 1

What does reserach allow us to see?

Answer 1

• How the world changes when we change just one thing (all other things being equal)
• Can truly test theory
• great internal validity
• cause-and-effect conclusions

Question 2

Mueller and Oppenheimer, 2014

Answer 2

• Q: Is typing or taking notes by hand better for college students?
• Methods: Class prepped in advance (1/2 w paper, 1/2 w laptop). They watched a Ted talk and took notes, had a 30 min distractor, and then were tested on factual and conceptual information
• Results: Groups performed similarly on factual questions, but handwritten notes performed much better on conceptual questions

Question 3

“Yong 2017”

Answer 3

• Babies that watch adult models persist at difficult tasks are more likely to work thorugh difficult things
• Utilized 100 babies 13-18 months
• Parents sat next to babies- but did not help
• 1/2 trials babies watched someone take the same object out of a box, strucggle, and persevere
• 1/2 trials babies watched someone take the same object of a box w/o difficults
• Babies were then given a toy w a button that looks like it should do something but didn’t work
• Babies who observed the adult work tried to push the button more

Question 4

What is a quasi-experimental variable?

Answer 4

• ex. history of SUD
• things the researcher cannot manipulate

Question 5

What is a manipulated variable?

Answer 5

• independent variable
• at least 2 levels
• ex. note taking (laptop vs handwritten), ex effort (effort vs no effort)
• can have more than 1 in an experiment

Question 6

What is the measured variable?

Answer 6

• dependent variable/outcome measure
• the variable you operationalize
• can be more than 1 in an experiment

Question 7

manipulation of variables- goal

Answer 7

• systematically alter independent variable to predict/see changes in outcome variable (DV)
• Correctly predicted change: theory validated
• Incorrectly predicted change: theory falsified/discarded or theory modified w new info

Question 8

Experiments support causal claims

Answer 8

• covariance: results will show that changes to IV=changes to DV
• temporal precedence: causal variable (IV) comes first and outcome measure comes second
• Internal validity: the study rules out confounding variables and alternative explanations

Question 9

Control Variable

Answer 9

• Important to compare manipulated behavior to “normal” (control)
• One condition that is not manipulated or one condition to compare to
• Independent variable answer: compared to what?

Question 10

Internal validity: treat all groups equally

Answer 10

• Except for the IV
• So do random assignment, all groups should look like all other groups (age, sex, IQ). Only difference should be condition
• Matched groups: participants are matched on every measurement except IV

Question 11

Design Confounds

Answer 11

• Design confound: study’s manipulation. Influences more than one psychological construct. Each contruct impacts the DV.
• an extraneous variable within a research design that unintentionally influences the relationship between the independent and dependent variables, potentially leading to inaccurate conclusions about the study’s findings
• Ex. Flickering lights in attention study. Attention impaired b/c of manipulation

Question 12

What is an operational confound?

Answer 12

• due to poorly defined construct
• IV still causes changes in DV
• Just not sure what the DV actually is
• Interpretation of results is at risk
• Don’t know why experimental group changed

Question 13

What are nuisance “Third” Variables?

Answer 13

• systematically vary based on the IV
• strongly related to the DV
• example: time of day

Question 14

Why do we do random assignment?

Answer 14

• For “person” confounds
• Individual differences co-vary with the IV
• Individual differences are related to the DV

Question 15

What are selection effects?

Answer 15

• people in one level of the IV are different from people in the others
• ex. Lovass, 1987

Question 16

Lovass, 1987

Answer 16

• One-on-one sessions therapy for autism for 38 familes
• Experimental (new) therapy vs traditional
• Families that live close got experimental therapy- “convenience” assignment
• No random assignment
• There may have been clear differences between the groups
• Intensive treatment performed way better but also in nicer area which could have influenced it

example of selection effects

Question 17

Strength of a true experiment: eliminate confounds

Answer 17

• to decrease procedural error
• double blind everyone
• standardize procedure
• treat all participants alike

help ensure internal validity

Question 18

Independent groups design

Answer 18

• Between groups/subjects
• different participants are placed in each group
• Ex. Sugar consumption on exam performance (1 group tic tac, 1 group kale, 2 groups candy/cookie)

several types

Question 19

Posttest Only Design/ Equivalent groups

Answer 19

• A “between groups” design simply refers to comparing different groups of participants, while an “equivalent groups” design specifically means that those groups are considered to be as similar as possible, usually achieved through random assignment, so that any observed differences can be attributed to the experimental manipulation rather than pre-existing variations between the groups.
• essentially, a “between groups” design can be considered “equivalent groups” if random assignment is used to create the groups.

Question 20

Pre-test-Posttest Design

Answer 20

• participants randomly assigned to one of (at least) 2 groups
• are tested on the DV twice
• AKA mixed design(combines elements of between and within design bc comparing groups but based on individual’s first score)
• Ex. test verbal GRE score- group 1 gets mindfulness class, group 2 gets nutrition class. Then test again.

Question 21

Within-groups design

Answer 21

• repeated measures
• participants are measured more than once on the same DV
• good when you have a small group
• ex. One group tastes chocolate w confederate, then rates the chocolate. Then tastes chocolate alone, and rates the chocolate. Compare chocolate ranking

Question 22

Advantages of Within-Groups

Answer 22

• participants in the group are equal
• people act as their own control/no cohort effects “you vs you”
• make more precise estimates of differences between IV levels
• small number of participants necessary

Question 23

What are order effects?

Answer 23

• being exposed to one condition infuleces how you respond to the later condition

Question 24

What are practice effects?

Answer 24

• aka fatigue effects
• long sequences might lead participants to get better at the task or get tired of the task

Question 25

What is counterbalancing?

Answer 25

• present different manipulations in different order
• compare the different conditions to see if scores are the same despite being in different order, if they are the same you can assume no order effects

Question 26

What is construct validity?

Answer 26

• how well the variables were measured and manipulated
• Ex. Note taking. DV-each question coded by 2 graders. Questions had good face validity. There was high interrater reliability (used more frequently when the data is subjective)