Regression I
Hypothesis testing using two variables
Bivariate hypothesis tests
Default hypothesis that the relationship between two variables is zero.
Null hypothesis
Variables associated or correlated with both X and Y that, if ignored, may make a spurious relationship look real.
Confounding variables
As we approach an infinite number of samples, the sample means approach a normal distribution around the true population mean.
Central limit theorem
Technique for estimating the parameters of a regression model
Ordinary Least Squares (OLS)
Two methods of bivariate hypothesis testing
Difference of means and correlation coefficient.
Type of figure that best depicts the relationship between two continuous variables
scatterplot
The ideal in which we want two parallel universes in which everything is identical except for X.
Counterfactual ideal
Bell shaped distribution with a single peaked mean that equals the median.
Normal distribution
Multiple regression
The probability that we would see the relationship that we are finding because of random chance.
p-value
Metric reflecting the idea that we will gain confidence in an observed pattern as the amount of data on which that pattern is based increases.
Degrees of freedom
Type of assignment to treatment and control groups that deals with confounding variables
Random assignment
Two parameters of a probability distribution
Mean and standard deviation
Goodness-of-fit measure that ranges between 0-1 indicating the proportion of the variation in the DV that is accounted for by the model.
R-Squared
Standard p-value most social scientists use to qualify results as "significant"
0.05
Statistical way of summarizing the general patterns of association between two continuous variables.
Covariance
Moving from abstract concepts to empirical measures.
Operationalization
As the sample size goes to infinity, the sample averages converges to the population average
Stochastic component equal to the difference between the actual value of the DV and the predicted value of the DV.
Residual
Type of DV/IV variables needed for a difference of means test.
Continuous DV and categorical IV.
Type of DV/IV variables needed for a correlation coefficient test.
DV and IV are continuous
The extent to which the measurement technique yields the same results when repeated
Reliability
Type of normal distribution when μ = 0 and σ = 1
Standard normal distribution
Two components that play a role in determining the magnitude of uncertainty in a regression model.
Individual residuals and sample size