In: Statistics and Probability
3. Data for magnitude and longitude for a sample of 24 earthquakes are below. The objective is to determine whether the magnitude of an earthquake is associated (predicted) by its longitude coordinates.
Longitude |
Magnitude |
-118.85 |
2.54 |
-122.20 |
2.9 |
-119.80 |
2.81 |
-104.82 |
3.5 |
-115.09 |
3.4 |
-97.15 |
3.2 |
-122.81 |
2.5 |
-121.75 |
2.7 |
-114.96 |
3.4 |
-115.54 |
3.2 |
-70.60 |
3.8 |
-113.4 |
2.8 |
-117.73 |
2.9 |
-117.66 |
2.8 |
-115.76 |
2.9 |
-116.46 |
2.8 |
-115.27 |
2.6 |
-115.14 |
2.6 |
-115.77 |
3.3 |
-121.49 |
3.1 |
-104.97 |
2.8 |
-121.02 |
2.8 |
-115.08 |
3.4 |
-122.19 |
2.8 |
3a. Specify the null and alternative hypotheses regarding model fit (1 point)
3b. Calculate the test statistic for model fit (3 points) and state your conclusion using an alpha of 5% (1 point)
Let’s compare the results you calculated for Q3b with results from a multiple linear regression.
4a. Would additionally controlling for ‘depth’ and ‘latitude’ be helpful? In other words, is a model that includes ‘depth’, ‘latitude’ and ‘longitude’ superior in model fit to a model that includes only ‘longitude’? Output for a multiple linear regression which includes longitude, depth, and latitude is provided below. (2 points)
4b. Interpret the parameter estimate for ‘longitude’ from the multiple linear regression output. (1 point)
Analysis of Variance |
|||||
Source |
DF |
Sum of |
Mean |
F Value |
Pr > F |
Model |
3 |
1.07870 |
0.35957 |
4.38 |
0.0159 |
Error |
20 |
1.64090 |
0.08204 |
||
Corrected Total |
23 |
2.71960 |
Root MSE |
0.27953 |
R-Square |
0.3966 |
Dependent Mean |
2.98200 |
Adj R-Sq |
0.3104 |
Coeff Var |
9.37398 |
Parameter Estimates |
|||||
Variable |
DF |
Parameter |
Standard |
t Value |
Pr > |t| |
Intercept |
1 |
4.86602 |
0.85582 |
5.69 |
<.0001 |
Depth |
1 |
0.00131 |
0.01084 |
0.12 |
0.9049 |
Latitude |
1 |
0.00564 |
0.01108 |
0.51 |
0.6157 |
Longitude |
1 |
0.01849 |
0.00561 |
3.29 |
0.0035 |