PSY 3128 Lecture Notes - Lecture 3: White Matter, Stereotype Threat, Ageism
5 Jun 2018
School
Department
Course
Professor
Experimental Research: we manipulate a variable and measure its effect on another variable.
Correlational Research: we attempt not to influence any variables, and simply measure the
relationship between two things.
Experimental Studies: most direct way to test a hypothesis about a cause-effect relationship
between variables.
Experimenter controls one variable: independent, hypothesized to cause effect on another variable.
The other is observed and measured: dependent – usually behavior, hypothesized to be affected.
All other variables are held constant – control group.
You only want one thing to be different between experimental and control groups.
Confound: another variable you didn’t notice or control that process an alternative explanation for
your findings.
Placebo Effect: you just expect things to be different even though they might not be.
Random Assignment: randomly assign treatments
Blinding: minimize expectancy by eliminating knowledge about experimental conditions.
i. Single blind – participant doesn’t know
ii. Double blind – person administering experiment doesn’t know
iii. Triple blind – those coding do not know what the study is about.
Age is not a true independent variable.
- You cannot give random assignment for age.
- Independent variable: the variable that is being manipulated.
- Dependent variable: the variable that is being measured.
Studies on ageing use quasi-experimental designs and compare groups on predetermined
characteristics.
ie. Sleeping with shoes on = waking up with a headache. Not related, confounded by alcohol.
Simpsons Paradox: incredibly challenging pattern that can show up.
When you make a scatter plot and run a correlational analysis, you look at things in the context
within another variable
X Age
Y Happiness
Z Sex/Gender
Multiple Correlational Designs: can make up for shortcomings of simple correlational designs.
i. Control for confounds other than age: other variables can be added into analysis that
the research suspects may be linked with age.
ii. Allow investigations of causality: paths can actually be drawn from one variable to
another, and their significance tested.
iii. Provide ways to examine change over time: can use statistical methods that model
changes in growth.
Example of structural equation modelling predicting mobility from flexibility and balance in older
adults.
- Circles are latent variables.
- Squares are variables that were measured.
- Assuming relationships between construct and others
Cross Sectional: testing 2 or more groups of participants at the same point in time to make
longitudinal conclusions.
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Experimental research: we manipulate a variable and measure its effect on another variable. Correlational research: we attempt not to influence any variables, and simply measure the relationship between two things. Experimental studies: most direct way to test a hypothesis about a cause-effect relationship between variables. Experimenter controls one variable: independent, hypothesized to cause effect on another variable. The other is observed and measured: dependent usually behavior, hypothesized to be affected. All other variables are held constant control group. You only want one thing to be different between experimental and control groups. Confound: another variable you didn"t notice or control that process an alternative explanation for your findings. Placebo effect: you just expect things to be different even though they might not be. Blinding: minimize expectancy by eliminating knowledge about experimental conditions. Double blind person administering experiment doesn"t know. Triple blind those coding do not know what the study is about. You cannot give random assignment for age.
