Home
Explore
Exams
Login
Get started
Home
cr1 lecture 16- correlation and simple regressioon
cr1 lecture 16- correlation and simple regressioon
0.0
(0)
Rate it
Studied by 0 people
Call Kai
Learn
Practice Test
Spaced Repetition
Match
Flashcards
Knowt Play
Card Sorting
1/66
There's no tags or description
Looks like no tags are added yet.
67 Terms
View all (67)
Star these 67
1
New cards
Correlation
Measures the strength and direction of a linear relationship between two variables.
2
New cards
Pearson's r
Sample correlation coefficient, ranging from −1 to +1.
3
New cards
Positive correlation
As X increases, Y tends to increase.
4
New cards
Negative correlation
As X increases, Y tends to decrease.
5
New cards
No linear correlation
Pearson's r is close to 0; there is little or no linear relationship.
6
New cards
Perfect positive correlation
r = +1; all observations lie on an upward-sloping line.
7
New cards
Perfect negative correlation
r = −1; all observations lie on a downward-sloping line.
8
New cards
Correlation direction
The sign of r (+ or −) shows the direction of the relationship.
9
New cards
Correlation strength
The absolute value of r shows the strength of the linear relationship.
10
New cards
Correlation assumptions
Both variables are at least interval level and their relationship is linear.
11
New cards
Scatterplot
Shows the relationship between two numerical variables using plotted points.
12
New cards
Correlation theoretical H₀
There is no linear correlation between X and Y in the population.
13
New cards
Correlation theoretical H₁ (two-sided)
There is a linear correlation between X and Y in the population.
14
New cards
Correlation statistical hypotheses (two-sided)
H₀
15
New cards
Positive one-sided correlation hypotheses
H₀
16
New cards
Negative one-sided correlation hypotheses
H₀
17
New cards
r versus ρ
r is the sample correlation; ρ is the population correlation.
18
New cards
Distribution for testing correlation
The t-distribution.
19
New cards
Correlation in SPSS
Analyze → Correlate → Bivariate → Pearson.
20
New cards
One-sided p
predicted direction
21
New cards
One-sided p
opposite direction
22
New cards
p < .05
Reject H₀; the result is statistically significant.
23
New cards
p > .05
Do not reject H₀; insufficient evidence against it.
24
New cards
Correlation effect size
Pearson's r is already an effect-size measure.
25
New cards
Correlation strength benchmarks
|r| < .30 weak; .30–.50 moderate; > .50 strong.
26
New cards
Correlation and causality
A significant correlation does not prove that X causes Y.
27
New cards
Simple regression
Uses one predictor X to predict a numerical outcome Y.
28
New cards
Correlation versus regression
Correlation measures a relationship; regression predicts Y from X.
29
New cards
Simple regression equation
Ŷ = b₀ + b₁X.
30
New cards
Ŷ
The predicted value of the dependent variable.
31
New cards
b₀ (constant/intercept)
The predicted value of Y when X = 0.
32
New cards
b₁ (slope)
The predicted change in Y for a one-unit increase in X.
33
New cards
Positive regression coefficient
A one-unit increase in X predicts an increase in Y.
34
New cards
Negative regression coefficient
A one-unit increase in X predicts a decrease in Y.
35
New cards
Unstandardised B in SPSS
Used to interpret the slope and construct the fitted regression equation.
36
New cards
Regression prediction
Substitute a value of X into Ŷ = b₀ + b₁X.
37
New cards
Ordinary Least Squares (OLS)
Finds the regression line that minimises the sum of squared residuals.
38
New cards
Residual
Actual Y minus predicted Y
39
New cards
Sum of squared residuals
Sum of squared prediction errors; smaller values indicate a better fit to the same data.
40
New cards
Overall regression model H₀
The model does not improve prediction of Y in the population.
41
New cards
Overall regression model H₁
The model improves prediction of Y in the population.
42
New cards
Overall regression model test
F-test, found in the SPSS ANOVA table.
43
New cards
Regression coefficient test
t-test, found in the SPSS Coefficients table.
44
New cards
Regression coefficient hypotheses (two-sided)
H₀
45
New cards
Positive one-sided regression hypotheses
H₀
46
New cards
Negative one-sided regression hypotheses
H₀
47
New cards
b₁ versus β₁
b₁ is the sample regression coefficient; β₁ is the population coefficient.
48
New cards
Regression 95% confidence interval
A range of plausible values for the population regression coefficient.
49
New cards
95% CI excludes zero
The coefficient is significant in a two-sided test at α = .05.
50
New cards
R² (R-squared)
Proportion of variation in Y explained by the regression model.
51
New cards
R² formula
R² = SS regression / SS total.
52
New cards
R² and correlation in simple regression
R² = r² when the model has one predictor and an intercept.
53
New cards
R² effect-size benchmarks
Below .09 small; .09–.25 moderate; above .25 strong.
54
New cards
R² = .634
The model explains 63.4% of the variation in Y.
55
New cards
Simple regression in SPSS
Analyze → Regression → Linear.
56
New cards
Regression SPSS variables
Numerical Y in Dependent; predictor X in Independent(s).
57
New cards
SPSS Model Summary
Provides R² and other measures of model fit.
58
New cards
SPSS ANOVA table in regression
Provides F, degrees of freedom and p for the overall model.
59
New cards
SPSS Coefficients table
Provides constant, B, standard error, t, p and confidence intervals.
60
New cards
Dichotomous variable
A categorical variable with exactly two categories.
61
New cards
Dummy variable
A dichotomous variable coded 0 and 1 for use in regression.
62
New cards
Dummy regression intercept
The mean of the group coded 0.
63
New cards
Dummy regression slope
The mean of group 1 minus the mean of group 0.
64
New cards
Dummy regression and independent t-test
Both test the same two-group mean difference with the usual equal-variance t-test.
65
New cards
Regression with dummy X = 0
Predicted Y equals b₀.
66
New cards
Regression with dummy X = 1
Predicted Y equals b₀ + b₁.
67
New cards
Regression and causality
Regression alone does not establish causality; research design matters.