EnviroAutomate
← All validation

Hypothesis Testing validation

Free hypothesis testing for environmental datasets: one- and two-sample t-tests, the Wilcoxon-Mann-Whitney test and one-way ANOVA with Bonferroni post-hoc comparisons.

Scenarios
4
Values agreeing
15
Values differing
0
Validated against
R EnvStats / nortest, USEPA ProUCL

Open Hypothesis Testing

The one-sample t-test, Welch two-sample t-test and Wilcoxon-Mann-Whitney test are compared against R EnvStats. Analysis of variance is compared against USEPA ProUCL, agreeing on the F statistic and p-value to thirteen decimal places.

Scenarios

Non-parametric site vs background comparison

Typical use

Input

Group Site (8 values)

value
12.4
15.1
11.8
18.2
14.6
16.9
13.2
19.4

Group Background (7 values)

value
8.1
9.4
7.7
10.2
8.8
9.1
11.5

Settings: test = mannWhitney, tail = twoSided, significance = 0.05

Against R EnvStats / nortest (EnvStats 3.1.0, nortest 1.0-4, R 4.6.1)

ValueR EnvStats / nortestEnviroAutomateDifferenceAgreement bandVerdict
Test statistic testStatistic000within 0.000001 relativeAgree
p-value pValue0.001460061916242150.00146006191636271.21e-13within 1e-9Agree

Not run for this scenario

  • USEPA ProUCL - ProUCL does not expose this test in a form that can be compared directly, so it is checked against R instead.

One-sample t against a standard, upper tail

Typical use

Input

Group Site (8 values)

value
12.4
15.1
11.8
18.2
14.6
16.9
13.2
19.4

Settings: test = oneSampleT, tail = greater, standard = 12, significance = 0.05

Against R EnvStats / nortest (EnvStats 3.1.0, nortest 1.0-4, R 4.6.1)

ValueR EnvStats / nortestEnviroAutomateDifferenceAgreement bandVerdict
Test statistic testStatistic3.2800494207597913.28004942075979144.44e-16within 0.000001 relativeAgree
Degrees of freedom degreesOfFreedom770within 0.000001 relativeAgree
p-value pValue0.0067430702244169230.0067430702244168923.12e-17within 1e-9Agree

Not run for this scenario

  • USEPA ProUCL - ProUCL does not expose this test in a form that can be compared directly, so it is checked against R instead.

Site vs background, unequal variances (Welch)

Typical use

Input

Group Site (7 values)

value
12.4
15.1
11.8
18.2
14.6
16.9
13.2

Group Background (6 values)

value
8.1
9.4
7.7
10.2
8.8
9.1

Settings: test = twoSampleTWelch, tail = twoSided, significance = 0.05

Against R EnvStats / nortest (EnvStats 3.1.0, nortest 1.0-4, R 4.6.1)

ValueR EnvStats / nortestEnviroAutomateDifferenceAgreement bandVerdict
Test statistic testStatistic5.9450037439127365.9450037439127358.88e-16within 0.000001 relativeAgree
Degrees of freedom degreesOfFreedom7.9575684116735627.957568411673561.78e-15within 0.000001 relativeAgree
p-value pValue0.000351060578381760.0003510605783818579.61e-17within 1e-9Agree

Not run for this scenario

  • USEPA ProUCL - ProUCL does not expose this test in a form that can be compared directly, so it is checked against R instead.

One-way ANOVA across three groups of unequal size

Boundary case

Input

Group A (5 values)

value
12.4
15.1
11.8
18.2
14.6

Group B (7 values)

value
8.1
9.4
7.7
10.2
8.8
9.1
9.9

Group C (4 values)

value
21.3
19.8
22.6
20.1

Settings: test = anova, significance = 0.05

Against R EnvStats / nortest (EnvStats 3.1.0, nortest 1.0-4, R 4.6.1)

ValueR EnvStats / nortestEnviroAutomateDifferenceAgreement bandVerdict
Test statistic testStatistic66.8042840835160766.804284083516125.68e-14within 0.000001 relativeAgree
Degrees of freedom degreesOfFreedom220within 0.000001 relativeAgree
p-value pValue1.44742476870365e-71.44742476893001e-72.26e-17within 1e-9Agree

Against USEPA ProUCL (ProUCL 5.2.0)

ValueUSEPA ProUCLEnviroAutomateDifferenceAgreement bandVerdict
Test used testUsedanovaanova-exact matchAgree
Test statistic testStatistic66.8042840835162866.804284083516121.56e-13within 0.000001 relativeAgree
Degrees of freedom degreesOfFreedom220within 0.000001 relativeAgree
p-value pValue1.44742476893001e-71.44742476893001e-70within 1e-9Agree

Evidence generated 2026-08-26 from a full comparison run against the versions named above.