t-Statistics and Correlation
sample mean should approximate population mean
Estimated standard error=square root of sample variance/ sample size
sm=square root of sample variance/sample
t-statistic is used when the population mean and variance are unknown
t-statistics are bell-shaped and symmetrical
null hypothesis states treatment has no effect
Hypothesis testing (4 parts)
- State hypothesis and select an alpha level
- Locate the critical region
- Calculate the test statistic
- Make a decision regarding hypothesis (H0)
Two assumptions for t-Statistic
- population sampled must be normal
- Values in the sample must have independent observations
Positive and/or negative signs depict the positive or negative direction of the relationship.
Perfect correlation: correlation of 1.0
intermediate values between 0 and 1 indicate the degree of tendency
Pearson correlation for a sample= degree to which X and Y vary together/ degree to which X and Y vary seperately
Correlation measures the degree of relationship between two variables on a scale from 0-1.0
Sum of products: measure the amount of covariability between two variables
A correlation does NOT necessarily imply a cause-and-effect outcome between two variables.
Partial correlation measure relationship between two variables while observing/controlling the influence of a third variable by holding it constant