Chys 3P15 Exam Review

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 1

CHYS 3P15 EXAM REVIEW- OUTLINE OF EXAM-

55 multiple choice questions


Cumulative :(
Main focus is after the midterm → measures of association (Chi square) and onward
“Here are the results→ what does it mean”
Bring a calculator for math questions
We will get a calculation sheet
o Effect size (eta or mega squared)
More things on the exam then being mentioned!
Should be able to interpret the output
Should know assumptions
(measurement association) Correlation and chi squared
o Are variables related
o Could get a scatter plot and asked which is the appropriate determination
o Scatter plot with results asking what the conclusion is
o Effect sizes (small, medium, large)’
o Directionality and causality (at least 3 ways they can be related)
 Correlation does not equal causation
o Correlation output and significance
o Partial and semi-partial (partial is bigger)
o Measures of association with categorical data (Chi square)
 Contingency table
o Chi-square test of independence (frequencies in each)
o Assumptions of chi-square (independence, expected frequencies larger than 5)
o Odds ratios (effect size)
 Linear Regression
o Closely associated with correlation→ used for description
o Regression is used for prediction
o Interception, equation, positive and negative slope, etc.
o Regression line satisfies all the errors (adds to 0) if we squares all the differences there would
be a minimum
o Anova (sums of squares we can, and cannot predict)
 Mean squared error
o R2→ variance accounted for
o Output: model summary (R and R2), Anova (df, f, sig), coefficients (unstandardized [original
units] standardized
o Multiple regressions (using variables to predict an additional variable)
o Slope values
o Methods of regression (hierarchical, simultaneous [forced entry], stepwise [computer only])
hierarchical is the best?
o Analysis of Variance: anova
o Partial and semi-partial
o Standardized beta values and standard deviations
 ANOVA
o Groups
 Levene’s test of homogeneity of variance (if not sig, it’s ok), anova, coefficients,
descriptions
o Contrasts or Post-Hoc?
 Contrast: only 2 chunks, independent, K-1
o Planned constrasts
o post-Hoc: test output (Bonferroni, Dunnet’s C)
o Effect size

You might also like