Wednesday, August 4 Chapter Eight Educational Research
Vocabulary
1. Predictor variable: (text pg. 196) predicts the criterion variable.
Since high school GPA precedes college GPA, it is called the predictor variable.
2. Criterion variable: (text pg. 196) the predicted dependent variable.
College GPA would be the criterion variable when using high school GPA as a predictor for college success.
3. Multiple regression analysis: (text pg. 197) combines several predictor variables.
When trying to predict college success, researchers used multiple regression analysis be combining the predictor variables of SAT scores, high school GPA and extra curricular activities.
4. Logistic regression: (text pg. 197) combines several variables to predict a dichotomous outcome.
When a researcher uses pass/fail as the dependent variable combined with various predictor variables, they use logistic regression to explore the relationship.
5. Attenuation: (text pg. 200) lowering of correlation because of unreliable measures.
As the measures used in a study become unreliable, the correlation lowers and is called attenuation.
6. Causal-comparative: (text pg. 202) nonexperimental studies designed to determine cause and effect.
When the researchers cannot control the program or study, they use a causal-comparative design such as different ways the principals evaluate teachers.
7. Ex post facto: (text pg. 202) presumed cause that occurred in the past.
Sometimes researchers cannot have any involvement in a study, thus instead they look into the past to see if for instance class size impacts achievement.
Reflection
While we catch up to the reading in class, I find myself struggling a bit with reliability and validity. Maybe if I write it out here I can gain some more understanding. When I look at whether a instrument is reliable or not, I want to see if it has been used over and over in studies. Then I want to make sure that there are as few chances for errors as possible. When I am looking for validity, I want to make sure that variable chosen in valid for that study. Sometimes an independent variable can be valid for one dependent variable, but not for another. I think I still need some examples to fully understand this. Hopefully, I will get them in class Thursday.
