In: Statistics and Probability
Below is the ANOVA table for a linear regression analysis:
Source |
SS |
df |
MS |
Lin Regression |
75 |
1 |
75 |
Residual |
1500 |
60 |
25 |
Total |
1575 |
61 |
(a) How many people were tested?
N = ________
(b) What are the tabled (α = .05) and calculated F-values?
Ftabled = ___________
Fcalc = ___________
(c) What are the two possible values for the product moment correlation (r)?
r = _______ or _______
(d) Without calculating the interval, will the 95% confidence interval that estimates β include 0? Why or why not?
Solution:
Given:
Source | SS | df | MS |
Lin Regression | 75 | 1 | 75 |
Residual | 1500 | 60 | 25 |
Total | 1575 | 61 |
Part (a) How many people were tested?
we have
dftotal = N – 1
61 = N - 1
N = 61 + 1
N = 62
Part b) What are the tabled (α = .05) and calculated F-values?
dfnumerator = dfregression = 1
dfdenominator = dferror = 60
α = 0.05
Ftabled = 4.00
Fcalc = MSregression / MSresidual
Fcalc = 75 / 25
Fcalc = 3.00
Part (c) What are the two possible values for the product moment correlation (r)?
thus
thus r = -0.2182 or 0.2182
Part d) Without calculating the interval, will the 95% confidence interval that estimates β include 0? Why or why not?
Since Fcalc = 3.00 < Ftabled = 4.00, regression model is insignificant. That is regression model can not be used to predict.
that also means regression slope β is insignificant or it is 0.
Thus answer is:
Yes, the 95% confidence interval that estimates β will include 0.