Types of Experiments & Research Designs UAPP 702: Research Design for Urban & Public Policy Based on notes from Steven W. Peuquet. Ph.D.

Slides:



Advertisements
Similar presentations
PhD Research Seminar Series: Valid Research Designs
Advertisements

Validity (cont.)/Control RMS – October 7. Validity Experimental validity – the soundness of the experimental design – Not the same as measurement validity.
Increasing your confidence that you really found what you think you found. Reliability and Validity.
Defining Characteristics
GROUP-LEVEL DESIGNS Chapter 9.
Group Discussion Describe the fundamental flaw that prevents a nonequivalent group design from being a true experiment? (That is, why can’t these designs.
Experimental Research Designs
Internal Threats to Validity
Correlation AND EXPERIMENTAL DESIGN
Research Design and Validity Threats
Experiments and Quasi Experiments Profesor Alexander Settles.
Aaker, Kumar, Day Seventh Edition Instructor’s Presentation Slides
RESEARCH METHODS Lecture 35. EXPERIMENTAL RESEARCH [CONTINUED]
Group Discussion Describe the similarities and differences between experiments , non-experiments , and quasi-experiments. Actions for Describe the similarities.
Experiments Types of Experiments Laboratory Experiments Experiments of short term duration and usually conducted in a lab under very controlled conditions.
Experimental Research
McGraw-Hill © 2006 The McGraw-Hill Companies, Inc. All rights reserved. Experimental Research Chapter Thirteen.
Experimental Research
Chapter 8 Experimental Research
Experimental Design The Gold Standard?.
Research Tools and Techniques
Experimental and Quasi-Experimental Designs
I want to test a wound treatment or educational program in my clinical setting with patient groups that are convenient or that already exist, How do I.
Learning Objectives 1 Copyright © 2002 South-Western/Thomson Learning Primary Data Collection: Experimentation CHAPTER eight.
Quantitative Research Designs
Day 6: Non-Experimental & Experimental Design
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.
Learning Objectives Copyright © 2004 John Wiley & Sons,Inc Primary Data Collection: Experimentation CHAPTER Seven.
ECON ECON Health Economic Policy Lab Kem P. Krueger, Pharm.D., Ph.D. Anne Alexander, M.S., Ph.D. University of Wyoming.
Copyright ©2008 by Pearson Education, Inc. Pearson Prentice Hall Upper Saddle River, NJ Foundations of Nursing Research, 5e By Rose Marie Nieswiadomy.
Techniques of research control: -Extraneous variables (confounding) are: The variables which could have an unwanted effect on the dependent variable under.
1 Experimental Research Cause + Effect Manipulation Control.
Experimental Designs. Experiments are conducted to identify how independent variables influence some change in a dependent variable.
Introduction section of article
Research Strategies. Why is Research Important? Answer in complete sentences in your bell work spiral. Discuss the consequences of good or poor research.
Chapter 10 Experimental Research Gay, Mills, and Airasian 10th Edition
Experimental Research Methods in Language Learning Chapter 5 Validity in Experimental Research.
1 Module 3 Designs. 2 Family Health Project: Exercise Review Discuss the Family Health Case and these questions. Consider how gender issues influence.
Experimental Research Hanser and Wheeler. Principles Independent Variable Dependent Variable.
Experimental & Quasi-Experimental Designs Dr. Guerette.
Research Design. Time of Data Collection Longitudinal Longitudinal –Panel study –Trend study –Cohort study Cross-sectional Cross-sectional.
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.
Research Methods Experimental Method
 The basic components of experiments are: 1) taking action 2) observing the consequence of that action  Experimental model is most closely linked to.
Chapter Eight: Quantitative Methods
EXPERIMENTAL DESIGNS. Categories Lab experiments –Experiments done in artificial or contrived environment Field experiments –Experiments done in natural.
IREL 561: Research Methods Fall 2013 Week 10 Based largely on Neuman’s Basics of Social Research, Chapter 8 Prepared by Craig Webster, Ph.D.
Experimental and Ex Post Facto Designs
CJ490: Research Methods in Criminal Justice UNIT #4 SEMINAR Professor Jeffrey Hauck.
1. /32  A quasi-experimental design is one that looks like an experimental design but lacks the key ingredient -- random assignment. 2.
Criminal Justice and Criminology Research Methods, Second Edition Kraska / Neuman © 2012 by Pearson Higher Education, Inc Upper Saddle River, New Jersey.
School of Public Administration & Policy Dr. Kaifeng Yang 研究设计 : 实验研究的基本问题.
Research designs Research designs Quantitative Research Designs.
Research Methodology Lecture No :19 (Experimental Design-Cont)
Chapter 9 Scrutinizing Quantitative Research Design.
Experimental and Quasi-Experimental Research
EXPERIMENTAL RESEARCH
Experimental Research
The Non-Experimental and Quasi- Experimental Strategies:
Designing an Experiment
Quantitative Research
RESEARCH METHODS Lecture 35
Introduction to Experimental Design
Experimental Research
Group Experimental Design
Experimental Research
Types of Designs: R: Random Assignment of subjects to groups
Reminder for next week CUELT Conference.
Presentation transcript:

Types of Experiments & Research Designs UAPP 702: Research Design for Urban & Public Policy Based on notes from Steven W. Peuquet. Ph.D.

Types of Experiments There are four kinds of experiments: True experiments Quasi-experiments Natural experiments Naturalistic experiments

True Experiments Some true experiments are done in the lab, others are done in the field. There are five steps to follow in conducting true experiments:

True Experiments 1. Need at least two groups: an experimental group and a control group. One groups gets the intervention, the other group does not. 2. Random assignment of individuals to the groups. (The degree to which random assignment ensures equivalence of the groups is dependant upon the size of the groups.)

True Experiments 3. The groups are measured on one or more dependent variables. This is called the pretest. 4. The intervention (independent variable) is introduced. 5. The dependent variables are measured again. This is the post test.

Quasi-Experiments Are experiments where the ability to randomly assign individuals to the experimental and control groups is limited or nonexistent. Quasi-experiments are more commonly used in evaluating social programs.

Quasi-Experiments Example In a study of the effectiveness of a new math curriculum, a 6 th grade class in Porter Middle School was not given the new curriculum (the “control group”), while the 6 th grade class in Farley Middle School was given the new curriculum (experimental group).

Quasi-Experiments Example (cont.) At the end of the school year the 6 th graders in the Farley MS class had higher average math achievement test scores than the kids in the Porter MS 6 th grade class. Is this a “quasi-experiment”? Why?

Quasi-Experiments Example (cont.) Answer: Most likely -- YES Random assignment to the two classes was probably not made, hence, groups may not be equivalent. Possibility of results being confounded by socio- economic status (SES) or other factors.

Natural Experiments Natural experiments are happening around us all the time. They are not conducted by researchers, but simply evaluated by researchers. In other words, the researcher does not have control over the application of the treatment. This also means that there is no control over what groups receive the treatment and the composition of those groups.

Natural Experiments Examples: Migration patterns of people — some people migrate, some don’t. Some middle class Latino students go to college, some don’t. Note: Each of these situations tests something about human behavior, the question is: “What is being tested by what is going on?”

Naturalistic Experiments In naturalistic experiments, one contrives to collect experimental data under natural conditions. You make the data happen out in the natural world (not in the lab), and you evaluate the results.

Naturalistic Experiments The difference between a natural experiment and a naturalistic experiment is that the first just happens, the second must be contrived to happen.

Naturalistic Experiments Naturalistic experiments deviate from true experiments because group membership is not randomly assigned, and exogenous factors (confounding variables) are not controlled.

Naturalistic Experiments Examples: Links needed to connect any two randomly selected people (passage of folders) Patience of drivers behind a car that did not move soon after a red light turned green (independent variables: newness of car and dress of driver) Matched pair testing to uncover discrimination

Internal and External Validity of an Experiment

Internal Validity High internal validity means that changes in the dependent variable were caused by—not merely related to or correlated with—the treatment. High internal validity can be attained only in an ideal true experimental design, or when steps have been taken to eliminate confounding factors or influences.

Internal Validity Internal validity can be increased by: Repeating an experiment many times, and by changing or adding independent variables; Switching the control and experimental groups Controlling confounding influences.

Internal Validity Controlled experiments have the virtue of producing results with high internal validity, but they have the liability of low external validity

External Validity External validity is the extent to which the cause and effect relationship can be generalized to other groups that were not part of the experiment.

External Validity External validity is critically important in policy and program evaluation research because we would like our research findings to be generalizable. If external validity is low, then the results are not generalizable.

Threats to Validity History confound — Any independent variable other than the treatment variable, that occurs between the pre-test and the post-test.

Threats to Validity Maturation confound — Refers to the fact that as the experiment is conducted, people are growing older, and are getting more experienced. (Problem in longitudinal studies in particular)

Threats to Validity Testing confound — Refers to different responses that an experimental subject may have to questions they are asked over and over again.

Threats to Validity Instrumentation confound — Results from changing measurement instruments. (Example: use of different people to collect data, if there is no inter- rate reliability, you have changed the measurement instrument.)

Threats to Validity Regression to the mean — Groups tend to increasingly resemble the mean over time.

Threats to Validity Selection bias confound — When self-selection by individuals into groups is possible, a selection bias can result.

Threats to Validity Diffusion of treatment confound — Occurs when a control group cannot be prevented from receiving the treatment in an experiment.

Examples Let’s look at two related but different research questions, and see what types of experiments would be appropriate for each.

Final Important Principle! NEVER USE A RESEARCH DESIGN OF LESS LOGICAL POWER WHEN ONE OF GREATER LOGICAL POWER IS FEASIBLE !