Causal inference A cause must always be relative to another cause Wrong: –Eating an avocado before taking a test causes better performance on the test.

Slides:



Advertisements
Similar presentations
Experimental Design I. Definition of Experimental Design
Advertisements

Variance in Research Design Sources, threats to internal validity, and “Noise”
GROUP-LEVEL DESIGNS Chapter 9.
Research Variables.
Experience Carefully Planned: Experimental Research Designs.
Causal Designs Chapter 9 Understanding when (and why) X  Y.
Quasi-Experimental Designs
Validity, Sampling & Experimental Control Psych 231: Research Methods in Psychology.
Who are the participants? Creating a Quality Sample 47:269: Research Methods I Dr. Leonard March 22, 2010.
Experimental Research Independent variable Dependent variable Causation.
Reliability and Validity in Experimental Research ♣
Design Conditions & Variables Explicating Design Variables Kinds of “IVs” Identifying potential confounds Why control “on the average” is sufficient Characteristics.
Chapter 9 Group-Level Research Designs. CHARACTERISTICS OF “IDEAL” EXPERIMENTS Controlling the Time Order of Variables Manipulating the Independent Variable.
RESEARCH DESIGN : 1. Kinds of support for making CAUSAL interpretations of observed relationships quality of theory research design used measurement procedures.
Lecture 10 Psyc 300A. Types of Experiments Between-Subjects (or Between- Participants) Design –Different subjects are assigned to each level of the IV.
Chapter 7 Conducting & Reading Research Baumgartner et al Chapter 8 Experimental Research.
Group-Level Research Designs
Sampling and Experimental Control Goals of clinical research is to make generalizations beyond the individual studied to others with similar conditions.
Psych 231: Research Methods in Psychology
Sampling & Experimental Control Psych 231: Research Methods in Psychology.
Psychlotron.org.uk Today’s session You will learn aboutContext The difference between extraneous and confounding variables Detecting and correcting flaws.
Chapter 9 Experimental Research Gay, Mills, and Airasian
Problem identification Research questions Constructs & Variables Research design.
EVALUATING YOUR RESEARCH DESIGN EDRS 5305 EDUCATIONAL RESEARCH & STATISTICS.
Statistical Analyses & Threats to Validity
1 Psych 5500/6500 Confounding Variables Fall, 2008.
+ Controlled User studies HCI /6610 Winter 2013.
Consumer Preference Test Level 1- “h” potato chip vs Level 2 - “g” potato chip 1. How would you rate chip “h” from 1 - 7? Don’t Delicious like.
Experimental Research Validity and Confounds. What is it? Systematic inquiry that is characterized by: Systematic inquiry that is characterized by: An.
Group Quantitative Designs First, let us consider how one chooses a design. There is no easy formula for choice of design. The choice of a design should.
Chapter 14 Cooper and Schindler
The Basics of Experimentation Ch7 – Reliability and Validity.
1 Evaluating Research This lecture ties into chapter 17 of Terre Blanche We know the structure of research Understand designs We know the requirements.
Experimental Designs. Experiments are conducted to identify how independent variables influence some change in a dependent variable.
Experiment Basics: Variables Psych 231: Research Methods in Psychology.
Introduction section of article
Chapter 10 Experimental Research Gay, Mills, and Airasian 10th Edition
PowerPoint presentation to accompany Research Design Explained 6th edition ; ©2007 Mark Mitchell & Janina Jolley Chapter 14 Single-n Designs and Quasi-Experiments.
Chapter 11.  The general plan for carrying out a study where the independent variable is changed  Determines the internal validity  Should provide.
SOCW 671: #6 Research Designs Review for 1 st Quiz.
Reasoning in Psychology Using Statistics Psychology
Research Methods and Data Analysis in Psychology Spring 2015 Kyle Stephenson.
Aim: What factors must we consider to make an experimental design?
Introduction to Validity True Experiment – searching for causality What effect does the I.V. have on the D.V. Correlation Design – searching for an association.
Measurement Experiment - effect of IV on DV. Independent Variable (2 or more levels) MANIPULATED a) situational - features in the environment b) task.
Studying Behavior Variable Any event, situation, behavior, or individual characteristic that varies - that is, has at least two values.
Chapter 1: Intro to Social Psychology Part 3: Wed. Jan 21, 2015.
Political Science 30 Political Inquiry Quasi-Experiments: Good Enough for Social Science.
Today’s session You will learn aboutContext The difference between extraneous and confounding variables Detecting and correcting flaws in experiments Studies.
Validity Threats to Validity Threats to validity – definition Theory vs. measurement Types of validity – Conclusion validity – Internal validity – Construct.
CHOOSING A RESEARCH DESIGN
Quasi-Experiments: Good Enough for Social Science
Types of Research Designs
Chapter 4: Studying Behavior
DUET.
Experiments: Part 3.
Making Causal Inferences and Ruling out Rival Explanations
Experiments and Quasi-Experiments
Social Research Methods Experimental Research
How do we misinterpret? Common threats to validity
Establishing the Direction of the Relationship
Reporter For this in-class activity, each group needs to first identify a recorder/reporter for their group. This reporter will report back to the full.
Experiments and Quasi-Experiments
Experiments: Part 2.
Experiments: Part 3.
Experiments: Part 3.
Experimental Design I. Definition of Experimental Design
Non-Experimental designs: Correlational & Quasi-experimental designs
Research Methods for the Behavioral Sciences
Independent variable: Factor that experimenter changes on purpose Dependent variable: factor that responds to the manipulated change of the IV.
Presentation transcript:

Causal inference A cause must always be relative to another cause Wrong: –Eating an avocado before taking a test causes better performance on the test. Right: –Eating an avocado before taking a test causes better performance on the test compared to not eating an avocado before taking the test. –Eating an avocado before taking a test causes better performance on the test compared to eating a banana before taking the test.

Randomized experiment Experimental condition (avocado) Control condition (no avocado) = on all variables other than the avocado Random assignment of individuals into condition

Quasi-experiment Ms. Jones’ class (avocado) Mr. Smith’s class (no avocado)  on all variables other than the avocado Random assignment of pre-existing groups into condition

Observational study Students who chose an avocado for breakfast Students who chose an anchovy and mushroom omelet for breakfast  on all variables other than the avocado Self-selection into conditions: Third variable problem!

Continuum of causal inference Least confidence Greatest confidence Case study Observational study Quasi- experiment Randomized experiment Ideal experiment

Which are valid causal statements (according to Rubin)? The moon causes the tides (compared to no moon) Gender causes differences in spatial abilities (being male compared to being female) Gender causes differences in the severity of punishment for misbehaving on the playground (being male compared to being female) Eating apples causes a decrease in requests for medical attention (compared to not eating apples)

Independent variables (IVs): manipulated variables Dependent variables (DVs): measured variables

Internal validity: The confidence with which we can say that the manipulated change in the IV caused the observed change in the DV

Confounding variable: Something that changes systematically with levels of the IV

Are these confounds? It’s raining all day while the man is jaywalking The man jaywalks in a suit at noon and the work clothes at 5pm The suit has a bright red tie

Control variables: held constant for all runs

Random variables: allowed to vary between runs

External validity: How well a causal relationship can be generalized across people, settings, and times

External validity Do control variables increase or decrease external validity? Do random variables increase or decrease external validity?

Threats to internal validity: Confounding variables History Maturation Selection Mortality Testing Statistical regression

Classical Test Theory Score on a test = True score + Error

Classical Test Theory Lowest possible score Highest possible score

Internal vs. External Validity Does the IV cause the change in the DV? Can we generalize to other people, settings, and times? Confounds threaten this: Too many control variables threaten this: