Her Review Flashcards
What does an F test tell us?
How well our model fits the data
Compared regression model to a mean model
What does the model summary tell us?
Variance accounted for in our DV by our Independent predictor (overlap)
Also the R2
Outcome= b0 + b1x + error
What is b0 and what is b1
B0= intercept
B1= slope of X
Cold symptoms = 3.30 - 0.67 vitamins +error
What does the 3.30 mean?
What does the 0.67 mean?
3.30 is the number of cold symptoms with zero vitamins
-0.67 means for every vitamin taken, cold symptoms reduce by .67 symptoms
What does the F test tell us for multiple regressions
How well our model fits the data
How well do ALL our PREDICTORS explain the DV
What is an R2?
How much variance in our outcome, our combined predictors explained
What is a residual?
Difference between predicted and observed data
Based on prediction on combination of IVs (X1, X2, X3)
What are we more interested in for a multiple regression, a semi partial correlation or a partial correlation?
Semi partial (unique prediction above all other IVs)
Square your “part” value and x100 for a percentage
What does an F test tell us for a hierarchical multiple regression
How well the model fits the data at:
Step 2: how do IVs in step 2 explain DV beyond IVs in step 1
What is a simultaneous regression
All IVs entered at the same time
What is a hierarchical regression
IVs entered in steps/blocks
What is a stepwise regression?
IVs entered based on association to DV
What is an R2 change?
Whether our next steps does a better job as a model in explaining our dependent variable then the previous step
How do we know if our model is a good fit when looking at the ANOVA table in regressions?
If the p value is less then .05
What does the F ratio test in an independent ANOVA
Model fit/group difference amongst 3 groups (or more)
Are the differences between the groups larger then within
If the F ratio is large what does this mean?
Group differences are larger than individual differences within group
When do you use a planned contrast?
Before you have a specific hypotheses
When do you use a post hoc?
After you expect differences but no specific hypothesis
You’re testing the hypothesis that a health program will have higher acamdeics then no program. You have three schools, one has no program, one has a physical health program, and one has a mental health program?
Do you use planned contrast or post hoc?
Planned contrast because we have a hypothesis
You’re testing the hypothesis that a health program will have higher acamdeics then no program. You have three schools, one has no program, one has a physical health program, and one has a mental health program.
What is your control and experimental group?
Control : no program
Experimental : physical health group, mental health group
You’re testing the hypothesis that a health program will have higher acamdeics then no program. You have three schools, one has no program, one has a physical health program, and one has a mental health program?
What is your first and second null?
First null (contrast): no difference in academics between no program compared to physical/mental program
Second null (contrast): no difference in academics between physical and mental health program
When looking at the table “Tests of Homogeneity of Variance” when have you met the assumption?
When the sig (p value) is greater then .05
When a table says “Multiple Comparisons” is that a planned contrast or a post hoc?
Post hoc
In a 2 way factorial ANOVA, how many IVs are there? DVs?
2 IV
1 DV
