Week Seven - Correlation & Regression Flashcards
What is Correlational Research?
A form of nonexperimental research in which VARIABLES ARE MEASURED NOT MANIPULATED.
What are the two goals of Correlational Research?
Describe
Predict
CAN NOT explain
Internal Validity is what in regards to correlational research and WHY?
Low, due to data collection methods and their lack of control. Ability to make claims is low.
External Validity is what in regards to correlational research and WHY?
High, due to o strong control.
When is Correlational Research used? (5)
Manipulation of independent variable is not possible
Manipulation of independent variable is not feasible
Manipulation of independent variable is not ethical
Establishing reliability and validity
Exploratory research
What variables can Correlational Research include?
Correlational research can involve continuous variables and/or categorical variables.
What does Pearson’s Correlation do?
Quantifies strength of correlation and association between two continuous variables
Range of Pearson’s r
From −1.00 (Strongest Possible Negative Relationship), Through 0 (No Relationship), to +1.00 (Strongest Possible Positive Relationship).
±.50-1.00 = Strong ±.30-.49 = Moderate ±.10-.29 = Weak 0-±.10= Trivial
What is the correlation coefficient?
Range = -1.00 to +1.00
Magnitude indicates strength of relationship
Sign indicates direction of relationship
Pearson’s product moment correlation (Pearson’s r).
What does a Scatterplot represent?
Visual presentation of the relationship between two variables.
Two measurements for each individual are represented by a single point.
How should we interpret a strong correlation?
Does not mean there is a causal relationship
Third variable
Causality may be reversed
Might reflect influence of outliers
How should we interpret a weak correlation?
Does not mean there isn’t a causal relationship
Might reflect truncated range
Might reflect curvilinear relationship
Might reflect outliers.
What are the 4 assumptions for Pearson’s r? What do they mean?
Linearity
If there is a relationship it must be linear (straight line), not curved.
Interval or Ratio data
The data for the Pearson’s r correlation must be interval or ratio level data.
If data are ordinal, use Kendall’s tau or Spearman’s rho
Univariate normality
Both variables should be normally distributed.
No outliers.
Bi-variate normal
Distribution of variables in combination should be normal.
How to calculate a confidence interval for Pearson’s r
r has to be converted to z-scores to calculate a CI
Writing up correlation, what does it look like?
r(N = sample size) = .r stat, p =, 95% CI [lower, upper].
Plain Englisht to describe relationship i.e., strength
include stats
explain what findings mean
What is Regression?
Attempts to determine the strength and character of the relationship between one dependent variable (usually denoted by Y) and a series of other variables (known as independent variables)
What does regression allow?
Allows prediction of one variable, from one or more measured variables.
What is Simple Regression?
Mathematical procedure that determines the straight line that best ‘fits’ the data.
What does Simple Regression include?
One predictor variable and one outcome variable.
Minimizes the distance between the outcome variable (Y) the predicted value of the regression line for each predictor value (X).
What is the Intercept (SR)?
Value of outcome variable (Y) when value of the predictor variable (X) is 0
- Not always meaningful or interpretable.
- Centring predictor makes the intercept equal to the mean of the outcome variable.
What is the B-Coefficient (SR)?
How much the outcome variable (Y) changes for each unit change of the predictor variable (X).
What is R-squared? What does it indicate?
How much of the variance in the outcome variable (Y) is accounted for by predictor variable (X).
R-squared indicates how much information the predictor variable provides about the value of the outcome variable
What does the SR stat formula look like?
F(df1, df2) = , p = , R = , R 2 =.
How can we check for Univariate Normality?
Univariate normality can be tested through inspection of statistics such as skewness, but a simple and powerful approach is to is to visually inspect figures of the data using histograms, box plots, or other graphical methods such as stem-and-leaf plots.
How do we check for Bivariate normality and outliers?
A simple method to check this requirement is to plot the variables against one another in a scatter plot.
