12. Extreme Value Test - Grubbs

The Grubbs test is developed to check the presence of values extremes in sample observations.

Example 1:

Data
99.1
110
97.6
98.2
102.1
103.2
103.4
102
98.6
98.4

We will upload the data to the system

To perform an extreme value test, the following setup is performed, as shown in the following figure.

Then click Calculate to get the results. You can also generate the analyses and download them in Word format.

The results are:

Summary of Outlier Test - Grubbs Test

	Value
Mean	101.260
Standard Deviation	3.781
Significance level	0.050
Critical value	2.290

Outlier Test Table

Sample	Z	Status
99.100	0.571	ok
110.000	2.312	Outlier
97.600	0.968	ok
98.200	0.809	ok
102.100	0.222	ok
103.200	0.513	ok
103.400	0.566	ok
102.00	0.196	ok
96.600	0.704	ok
98.400	0.756	ok

The result of batch 2 is considered an outlier, according to the analysis performed in the Action software. The statistic $Z$ obtained was 2.312, compared to the critical value $Z_c$ of 2.29 for n=10. Since $Z>Z_c$, it is concluded that the result of batch 2 is an outlier.

Example 2:

Data
1.90642
2.10288
1.52229
2.61826
1.42738
2.22488
1.69742
3.15435
1.98492
1.99568