ht_twosample_tnonpooled(A, B, alpha, type)

Nonpooled T-test for two population means. Hypothesis test for the means of two independent samples A and B. alpha is the confidence level. Parameter type can be 0 (left tail), 1 (right tail), or 2 (both tails).
Assumes samples are independent and that populations are normal or samples are large.

See also ht_twosample_std ht_twosample_tpaired ht_twosample_tpooled 

Example

(See NOTE below)

a=(370, 345, 360, 450, 510, 505, 445, 335, 295, 280, 315, 325, 490, 500)

( 370 345 360 450 510 505 445 335 295 280 315 325 490 500 )

b=(430, 455, 445, 490, 455, 535 )

( 430 455 445 490 455 535 )

ht_twosample_tnonpooled(a, b, .01, 0)

Two-sample Non Pooled T Test.

Hypothesis test for the means of two independent samples.

Ho : m1==m2

Ha : m1

Significance level : 0.01 (1.0%)

Results:

m1 : 394.64285714285717

m2 : 468.3333333333333

s1 : 84.74995542804105

s2 : 38.16630276391292

D : 17

Test statistics T : -2.680420404239568

Critical left value found is -2.566933982674698

P value is : 0.007903784417611432

Reject the null hypothesis.

NOTE: The data used in this example was taken from the textbook Introductory Statistics 6th Ed. by Neil A. Weiss with kind permission of Addison Wesly Longman, Inc. (AWL). AWL does not endorse nor it supports the Calcugator.