Study
We surveyed a random sample sample from the population of North Olmsted High School juniors and seniors. We wanted to obtain data from students who maintain a job outside of school so we decided to sample juniors and seniors instead of the whole population at North Olmsted High School. If we were to sample freshman and sophomores as well, the data would be biased because most freshman and sophomores most likely don’t have a job. We surveyed a total of seventy juniors and seniors at random.
The type of test used was the linear regression
We are testing β.
β = the slope of the population regression of the number of hours worked vs. the student’s GPA.
Ho: There is no linear correlation between number of hours worked and the student’s GPA
Ha: There is a linear correlation between number of hours worked and the student’s GPA
test-statistic: linear regression t-test=(b- β)/(Sb)
Assumptions:
Linear relationship
Multivariate normality
no or little multicollinearity
no auto-correlations
homoscedasticity
GPA = 3.12 - 0.00428 HOURS WORKED
Constant 3.1223 0.1231 25.37 0.000
HOURS WORKED -0.004275 0.008895 -0.48 0.633
Regression 1 0.0896 0.0896 0.23 0.633
Residual Error 40 15.5195 0.3880
Total 41 15.6091
Obs WORKED GPA Fit SE Fit Residual St Resid
10 30.0 1.7000 2.9940 0.2129 -1.2940 -2.21R
22 35.0 2.9000 2.9726 0.2534 -0.0726 -0.13 X
34 0.0 1.5000 3.1223 0.1231 -1.6223 -2.66R
X denotes an observation whose X value gives it large leverage.
linear regression correlation of determination: 0.006
weak correlation r-square value is 0.00-0.49
moderate correlation r-square value: 0.50-0.80
strong correlation r-square value: 0.81-1.00