Variable, Variable
Independent IV
Any characteristic or attribute that varies (e.g., gender, income, etc.) The hypothesized cause of variation in a dependent variable They hypothesized effect of variation in an independent variable
Dependent Variable (outcome variable), DV
Extraneous Variable
A variable that is not explicitly included in a hypothesis but has a causal impact on the DV
A research study in which participants are randomly assigned to conditions which are identical except with respect to some variable being manipulated at different levels
Or, put differently: Manipulate the independent variable in order to observe effects on the dependent variable, while trying to hold constant all extraneous variables
The Classic Example
Medical research Medicine vs. Placebo
Another Example
Does academic status of
(Is this man an intellectual fraud?)
the author affect the evaluation of an unintelligible text?
Manipulation
Conditions
Term for when you turn the IV on and off, or put it at 2 or more levels Different circumstances created by the manipulation of a variable Medicine vs. Placebo
Experimental Group, Control Group Random Assignment
Helps control for participantlevel factors that are hard to control Even ones you wouldn't think of
Before Random Assignment
R After Random Assignment
Group 1 Experimental Groups
Colors symbolize any differentiating attribute among the individuals (e.g., personality type)
Group 2
In Experiments You Must Manipulate an IV Creating 2 or More Conditions One Condition Studies are NOT experiments
Breaching "Experiments" The Baking Soda Volcano
Control Are all aspects of the setting held constant besides the IV being manipulated? Experimental Realism How engaged and into the study is the subject? Mundane Realism How much is the setting like the real world? Not always good! real The world is messy. (Why else would you be here?) "Audit studies," Devah Pager
The "Real World"
Control
By carefully controlling for other factors, we can feel confident regarding causality By creating setting we can create revealing circumstances that might rarely occur
Artificiality
Easily replicated "Keep it simple, stupid" But the method has drawbacks and doesn't suit all research topics...
"Can You Really Study an Army in the Laboratory?" (Zelditch 1969)
All Methods have Advantages and Disadvantages So be a Methodological Pluralist! Demand Effects
Selfpresentation "Hawthorne Effect" Reversedemand effects
Experimenter Effects
Experimenter unconsciously gives cues to subjects regarding how to act
External Validity
A.K.A. "Generalizability" Often a consideration in sampling How well does a lab result generalize? Depends on setting and participants
Sociological SPists say this is bogus
Experiments test theory No empirical study is directly generalizable to other settings
Both are sort of right
Demand Characteristics
Experimenter Effects
Safeguards Have a good "cover story" (if the study involves deception) Conduct poststudy interviews Safeguards Standardize procedures (use video) Limit experimentersubject interaction Keep research personnel "blind" to condition, now say "unaware"
Pain (is bad) Deception in experiments
Informed consent
Milgram study, replications
Debriefing
The special case of economics Is important
Due Thursday, January 29th Write a short description of a recent conflict you had with someone else (or any significant, interesting, or memorable one) Explain causes and content of the conflict Give your perspective; feel free to go off on him/her Keep it short: 2 pages, doublespaced