DOC PREVIEW
UW-Madison SOC 357 - Defining Experiments

This preview shows page 1-2 out of 7 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 7 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 7 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 7 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

1I. Defining Experiments• By experiment we refer to that portion of research in which variables are manipulated and their effects upon other variables observed. II. Threats to Internal Validity• Internal validity means that the conclusion drawn from an experiment is valid for the respondents participating in the experiment. For any experimental design, we always require interval validity, without which experimental results are not interpretable. Example:• Testing the hypothesis that taking Soc. 134 (Problems of American Racial and Ethnic Minorities) makes white students' racial attitudes more tolerant towards minorities. • A. One-Group Pretest-Posttest Design:Time 1 Intervention Time 2 Y1 Soc 134 Y2 (Y=tolerance):2Example• If Y1< Y2, can we be certain that it is the class that increased students' racial tolerance toward minorities? • No, there are other possible explanations for the change. These competing explanations are called threats to internal validity.Common Threats to Internal Validity• 1. History, the specific events occurring between the first and second measurement in addition to the experimental variable. • 2. Maturation, processes within the respondents operating as a function of the passage of time per se (not specific to the particular events), including growing older, growing hungrier, growing more tired/bored, and the like.Common Threats to Internal Validity (Continued)• 3. Testing, the effects of taking a test upon the scores of a second testing (like soc. desirability).• 4. Instrumentation, in which changes in the calibration of a measuring instrument or changes in the observers or scorers used may produce changes in the obtained measurements.• 5. Statistical regression, operating where groups have been selected on the basis of their extreme scores (e.g., bootcamps for troubled teenagers).3A Better Design is:• Have we solved all the problems?• B: Nonequivalent control group designGroup Time 1 Intervention Time 2Treatment Y1 Soc 134 Y2 Control Y1 Y2 Common Threats to Internal Validity (Continued)• 6. Biases resulting in differential selectionof respondents for the comparison groups.• 7. Experimental mortality, or differential loss of respondents from the comparison groups.• 8. Selection-maturation interaction, which might be mistaken for the effect of the experimental variable.III. Experimental Designs• Proper experiment designs can help solve the above problems and establish internal validity. • C. Pretest-Posttest Control Group DesignGroup Time 1 Intervention Time 2R=>Treatment Y1 Soc 134 Y2 R=>Control Y1 Y24IV. Threats to External Validity• Study may be internally valid but externally invalid. • Definition: External validity asks the question of generalizability: to what extent can the experimental results be generalized? • Many factors can jeopardize external validity or representativeness. The main ones are discussed as follows.Threats to External Validity (Continued)• 9. The reactive effect of testing, in which a pretest might increase or decrease the respondent's sensitivity or responsiveness to the experimental variable and thus make the results obtained for a pretested population unrepresentative of the effects of the experimental variable for the unpretesteduniverse from which the experimental respondents were selected. Threats to External Validity (Continued) • 10. The interaction effects of selectionbiases and the experimental variable.• 11. Reactive effects of experimental arrangements, which would preclude generalization about the effect of the experimental variable upon persons being exposed to it in nonexperimental settings.5V. Quasi-Experimental Design• A. DefinitionQuasi-experimental designs are natural settings in which the researcher can introduce something like experimental design into his/her data collection procedures. B. The Time-Series Experiment • Plot out time-series graphically and look for sudden changes that may be attributable to intervention. • Example: imitation suicide• Example: effects of the Civil Rights movement on racial gaps in earnings. “Natural” Experiments • Social experiments facilitated by unpredictable events (e.g., changes in laws, wars, natural disasters, etc.).•Examples– Installation of TV in remote B.C. town– Identical twins separated at birth – nature vs. nuture6C. Statistical Control• With observation data, we cannot control for confounding by randomization. • Example: We observe that, on average, female high school teachers earn $2,000 less than men. We want to know why but cannot do experiment by randomly assigning teachers to different sexes.• We control by multivariate analysis => "statistical control." • Let y be the dependent variable, earnings ($1000), and x1through x3 be three independent variables: sex (x1), qualifications (x2), and subject (x3).In Regression Syntax •yi= β0+ β1xi1+ β2xi2+ β3xi3+ εi, • where β's are parameters, β0is the constant, or the parameter for 1.  β's are also called "partial" regression coefficients. Say β1= -1.0 (with x1i=1 for female), e.g., controlling for qualifications and subject, female teachers earn 1000 dollars less than male teachers. Different from the bivariate effect (e.g, -2 to -1). Interesting example• Does a criminal record hinder employment?• Very hard to collect useful data• Audit study• Two testers – matched on observable characteristics (e.g., race, age, etc.)• Target entry-level jobs• Compare callbacks following applications/interviews• Huge difference observed (esp. true for black men)7Matching techniques• Quasi-experimental technique• Try to generate samples (or pairs) that resemble each other in every way except the outcome/treatment of interest• Problem remains that you cannot match on unobservables• If unobservables are important correlates of both treatment and outcome, then you have trouble with


View Full Document

UW-Madison SOC 357 - Defining Experiments

Documents in this Course
Syllabus

Syllabus

12 pages

Sampling

Sampling

35 pages

Class 7

Class 7

6 pages

Review

Review

3 pages

Load more
Download Defining Experiments
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Defining Experiments and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Defining Experiments 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?