ttest_wizard.testRun.runTTest

Description

Calculates one App-owned t-test with Base MATLAB beta functions. Difference direction is always A minus B. Scientific validation failures return a result row with ok=false and a stable status instead of fabricating a statistic.

Inputs

vectorA

Finite numeric vector with at least two values.

vectorB

Finite numeric vector with at least two values.

options

Scalar struct with method, alternative, alpha, labelA, and labelB. method accepts a label from testRun.choices or welch, pooled, or paired. alternative accepts a choice label or two_sided, greater, or less. alpha is a finite scalar strictly between zero and one.

Outputs

result

Scalar struct created by emptyResult. Successful results include vector snapshots, summaries, A-minus-B estimate and confidence bounds, standard error, t statistic, degrees of freedom, and p-value.

Units

A and B must use the same measurement unit. Means, standard deviations, the A-minus-B estimate, standard error, and confidence bounds retain that unit. The t statistic, degrees of freedom, and p-value are dimensionless.

Assumptions

Independent tests require observations that are independent within and between A and B. The pooled method additionally assumes equal population variances. The paired method requires displayed A(k) and B(k) to describe the same experimental unit; it analyzes A(k)-B(k). Test choice and independence remain the caller's scientific responsibility.

Failure Behavior

Invalid options throw ttest_wizard:InvalidTestOptions. Nonfinite vectors, sample sizes below two, unequal paired lengths, or zero standard error return ok=false with status invalid_input, insufficient_n, unequal_pairs, or zero_standard_error.

Example

options = struct( ...
"method", "welch", ...
"alternative", "two_sided", ...
"alpha", 0.05, ...
"labelA", "Condition A", ...
"labelB", "Condition B");
result = ttest_wizard.testRun.runTTest( ...
[1.2 1.4 1.3 1.5], [1.8 1.7 2.0 1.9 1.6], options);
assert(result.ok)