Undergraduate Statistics Final Exam

What is the purpose of correlation?

examines strength of relationship between two variables

What are the two types of correlation?

Bivariate and Partial

What are the four types of bivariate correlation?

Pearson, Spearman, Point-biserial, Biserial

Bivariate: What are Pearson and Spearman Correlations?

Pearson:

• Parametric Data

• Normal distribution

• ratio/Interval vars

Spearman

• Ranked means

• non-paramatric

• scale of measurement: interval, ratio, or ordinal

How to find effect size for correlations?

r²

What are the assumptions for Pearson bivariate correlation?

Parametric, normally distributed, two scale variables

What are the two types of partial correlation?

Partial and semi partial

What are the assumptions of partial correlation?

• IV: interval/ratio

• DV: interval or ratio

• control variable: interval, ratio, categorical (dichotomous)

• parametric data

• normal distribution

• linear relationship between variables and homoscedasticity (scatter matrix)

• no univariate outliers

• no multivariate outliers

What is the hypothesis for partial correlation?

H0: there is no correlation between IV and DV when controlling for Z. ρ = 0

What is the purpose of simple linear regression?

• find magnitude of relationship between two variables

• find model that best predicts an outcome

What are the IV and DV in a linear regression also called?

IV: predictor, exogenous

DV: outcome, endogenous

What is the hypothesis of a simple linear regression?

The coefficient of the slope equals zero; β1 = 0

What are the variables used in a simple linear regression?

1 predictor- continuous or categorical dichotomous

1 outcome- continuous, ratio/interval

What are the assumptions of simple linear regression?

parametric (norm dis and int or ratio)

linearity, homoscedasticity (Q-Q and scatter plot)

Independence of observations (durbin watson)

normality of residuals (scatterplot)

no univariate outliers

correlation between .3 and .9

What is the equation?

Y^ = B0(int. constant) + B1Xi (slope B)

What is the effect size of simple linear regression?

r²

What are the variables for multiple linear regression?

2+ predictors (cont. or dichotomous)

1 outcome (cont.)

What are the assumptions of MLR?

• normality

scatterplot matrix:

• linearity

• homoschedasticity

in results:

• correlation pearson’s r

• multicollinearity- tolerance; VIF

• normality of residuals- histogram, P-P, Q-Q

• outliers of residuals -casewise +/- 3

• independence of observations - durbin watson close to two

effect size for MLR?

adjusted R²

What is the purpose of a one-way ANOVA?

analyze variance of groups to assess difference in means between three or more groups

What are the assumptions of planned contrasts ANOVA?

• one-tailed hyp

• 1 IV w at least 3 levels categorical

• 1 DV parametric (ratio/int)

• Normal dist

• sample size roughly equal

• homogeneity of variance Levene’s test

What are the two types of planned contrasts?

orthogonal: control group

non-orthogonal: no control group

What variables for repeated measures ANOVA?

1 IV 2 or more levels

1 DV parametric

What are the assumptions for repeated measures ANOVA?

• normally distributed

• within-groups variance sphericity Mauchly

• no missing data

• if sphericity not met greenhouse or hyuhn

effect size for nonparametric test?

z/squareroot of n

What are the four nonparametric tests?

mann-whitney (between-subjects t-test), wilcoxon (within subjects t-test), kruskal-wallis (between subjects anova), friedman (within subjects anova)

What variables for multi-factorial anova?

2 or more IV

1 DV parametric

assumptions for multifact anova?

normal dist for each level of IV

no outliers

at least 20 in each level

equal sample sizes in each IV

homogeneity of variances (levene’s)