Research Article

Efficiency Comparison of New Adjusted Nonparametric and Parametric Statistics Interval Estimation Methods in the Simple Linear Regression Model

Table 2

The average width of the confidence interval for an independent variable has a scale-contaminated normal distribution with mean of 1 and variance of 1 and an error has a scale-contaminated normal distribution with mean of 0 and variance of 1, 4, 9, 16, 25, 36, 49, 64, 81, and 100; sample sizes are 10, 30, 50, 70, 100, and 200.

nVarianceAverage width of confidence interval
LSBSJKTOTATSSLSBSJKTOTATSS

101
42.830
94.334
165.832
257.290
368.748
4910.206
6411.664
8113.122
10014.580
3011.0131.5731.0281.136
41.6582.3632.0572.489
92.8533.2903.085
164.2404.114
254.4735.2225.142
366.2236.170
499.4527.199
648.2128.227
819.2459.256
10010.28010.284
5010.6790.7310.742
41.343
9
16
25
364.004
494.325
64
81
100
7010.4540.4570.507
40.9040.911
9
16
25
36
493.533
643.777
81
100
10010.3600.3630.394
43.7510.7170.7390.848
91.070
16
259.055
36
49
643.360
813.564
1003.757
20010.2630.2650.263
40.5250.5280.526
93.9430.7850.7900.789
161.0431.0541.052
251.2971.3151.315
361.5471.5801.578
491.7921.8421.841
642.0322.1092.104
812.2662.3702.367
1002.4942.6352.630

Note: the bolded value refers to the average width of the narrowest confidence interval, and the underlined value means that the average width of the confidence interval cannot be calculated because the coefficient of confidence interval is not within the specified range.