Correlation and Regression

Correlations

100

What is the verbal null and alternative hypothesis?

There is no significant linear relationship between [variable 1] and [variable 2].

There is a significant linear relationship between [variable 1] and [variable 2].

200

What does correlation not imply?

Causation

300

What are the assumptions?

1) Two quantitative

2) No outliers

3) A probable linear relationship

400

Identify the ranges between a weak, moderate, and strong correlation.

Weak: (+/-) 0.01-0.29

Moderate: (+/-) 0.30-0.69

Strong: (+/-) 0.70-1.00

500

What is the outline of the conclusion?

In our sample of N = [sample size] [sample description] we have determined that we can [reject/fail to reject] the null hypothesis that their is not a linear relationship between [variable 1] and [variable 2], R([N-2])= [value], p[=/<] [p-value].

Regression

100

What is the symbolic null and alternative hypothesis?

Null: H₀: B₁=0

Alternative: H_A: B₁≠0

200

What is the null and alternative verbal hypothesis?

Null: [Predictor variable] has no linear effect on [outcome variable].

Alternative: [Predictor variable] has a linear effect on [outcome variable].

300

What are the assumptions?

1) The sample is representative of the population

2) The relationship is believed to be linear

3) Variables are normally distributed

4) There is no multicollinearity

400

What is the model for regression?

B₀+B₁^*X

500

What is the outline of the conclusion if we reject the null hypothesis?

In our model [predictor variable] account for a portion of the variance in [outcome variable], R² = [.--]. The test statistic F(k-1, n-k) = [value], [p-value] is less than our chosen alpha of a=0.05. Because this test statistic was in the rejection region, I will reject the null hypothesis that [predictor variable] has no linear effect on [outcome variable]. I conclude that there is a statistically significant linear effect of [predictor variable] on [outcome variable], B= [b-value], p [p-value associated with B].

Random

100

True or false

Correlations cannot be negative they can only be positive.

False: Correlations can either be positive or negative

200

What kind of correlation is this?

The more I exercise the healthier I'll be. 

Positive correlation

300

What cell, in SPSS coefficients^a, tells me the B-value?

Row: Constant

Column: Unstandardized B

400

In what data table can I find B₀ and B₁?

Coefficients^a

500

What data table do I use to find the F-Statistic values for regression in SPSS?

ANOVA^a

Symbols

100

X

Predictor Variable

200

Y

Outcome Variable

300

B₁

The effect of the predictor variable on the outcome variable

400

R/r

Correlation symbol

500

In regression:

What is (k-1)?

What is (n-k)?

K-1 is df regression

N-K is df residual

Vocab

100

Define an independent variable

A variable that does not depend on another.

200

Define a dependent variable

A variable that depends on another

300

Define Standard Deviation

The average distance to the mean.

400

Define Variability

The way in which participants answers fluctuate

500

Define Spurious

No real existential relationship