Class Notes (1,100,000)
CA (630,000)
UW (20,000)
PSYCH (2,000)
Lecture

PSYCH291 Lecture Notes - Dependent And Independent Variables, Internal Validity, Psychology Today


Department
Psychology
Course Code
PSYCH291
Professor
Joanne Wood

This preview shows page 1. to view the full 5 pages of the document.
Threats to Internal Validity
What is internal validity?
*did X cause Y?
*Did changes in the predictor/independent variable cause the changes in the dependent
variable?
Threats to internal validity
*things that jeopardize our ability to conclude that X caused Y
*Things that jeopardize our ability to draw conclusions abt causality
*Things other than the predictor variable that may had lead to effect
*A factor other than X that may have caused Y
Correlational Designs
*Especially susceptible to threats to internal validity
*Two issues in correlational designs that threaten validity:
-3rd variable and temporal order
Types of correlational designs
*Typical correlational design: use continuous measures to measure both the predictor
and dependent variables
*Continuous – ranked order from low to high
*Categoral – discrete categories were talked abt
Correlational Designs
*Typically, continuous measures used for both the predictor and dependent variables.
*Two new types of correlational designs
*One group pre-test post-test design – pre test- exposed to x- get a pre-test -
examine single group
*Static groups design – compare two groups, one group gets the x, one grou has
not
*Two different ways of looking for the effects of X
*Static groups design
*Compare two groups—
*One that has the X
*One that has not
You're Reading a Preview

Unlock to view full version

Only page 1 are available for preview. Some parts have been intentionally blurred.

*If difference between them, is it the X?
*One group pre-test post-test design
*Examine single group
*Before an X occurs and after an X occurs
*Is difference from before to after due to X?
*Want to conclude that the difference from before to after is due to X
*But could it be something other than X? (threats to internal validity)
*Two ways of making comparison – 1. looking the same group before gettiong
the x, one
*(better than loooking only at yes/yes cell – no comparison there)
Threats to internal validity
*Both of these designs have their own threats to internal validity.
*Both correlational designs no manipulation
*Today we will focus on threats that apply to the one group pre-test post-test design
Example of OGPTPT design
The Globe and Mail, June 25, 2001
BRAIN FOOD
“In one large-scale analysis of approximately one million students enrolled in the New
York City school system,” reports Psychology Today, “researchers examined IQ
scores before and after preservatives, dyes, colourings, and artificial flavours were
removed from cafeteria lunch offerings. They found a 14 percent improvement after
the removal. And the improvement was greatest for the weakest students.”
*PV: chemicals on the food IV:IQ scores
BRAIN FOOD
How requirements addresed in OGPTPT design?
*Covariation
-before A, B lower
-As A changed, B changes
2. Temporal Order (check)
3. Uniqeness
*School learning (third variable)
Or was it…
*maturation
*history
*testing
*Instrumentation
*mortality
*regression to the mean
You're Reading a Preview

Unlock to view full version