ttest — One-sample and paired-sample Student-t tests.
Test a population mean with unknown variance, or test the mean of paired differences.
Syntax
[h,p,ci,stats] = ttest(x,Name,Value,...)
[h,p,ci,stats] = ttest(x,reference,Name,Value,...)Inputs
x- Type
NumericArray- Required
- Yes
- Default
- —
- Description
- Real single or double sample array.
name_value_pairs- Type
Any- Required
- Variadic
- Default
- —
- Description
- Alpha, Dim, and Tail name/value pairs.
reference- Type
NumericArray- Required
- No
- Default
0- Description
- Scalar population mean or same-sized paired sample.
Returns
h- Type
LogicalArray- Description
- Reject when p is less than Alpha.
p- Type
NumericArray- Description
- Student-t tail probability.
ci- Type
NumericArray- Description
- Confidence limits for the population mean or paired mean difference.
stats- Type
Any- Description
- Scalar structure containing tstat, df, and sd arrays.
Errors
RunMat:ttest:InvalidArgument- When
- Sample class, paired geometry, option grammar/value, device residency intent, or output count is unsupported.
- Message
ttest: invalid argument
Samples and hypotheses
ttest(x) tests mean zero. ttest(x,m) tests a finite scalar population mean m. A nonscalar second sample selects the paired test and must have exactly the same size as x; paired observations are subtracted before estimating variance. Samples and numeric controls accept real single or double values. Native integer, logical, complex, object, and distributed sample inputs are unsupported rather than implicitly coerced.
NaN observations are omitted; a paired observation is omitted if either member is NaN. Each lane uses its observed count n, sample standard deviation sd, and n-1 degrees of freedom. Fewer than two observations produce NaN probabilities and a false decision. Computed confidence endpoints are NaN in that case; the untested endpoint of a one-sided interval remains infinite. Zero standard error gives the formula's signed infinite or undefined t statistic, and finite confidence bounds can collapse to the observed mean. Infinite observations can produce undefined statistics.
The default tests the first nonsingleton dimension. Dim selects a positive, platform-representable 1-based integer dimension; dimensions beyond the input rank are singleton dimensions. Dimension-derived metadata must have representable geometry and a successful allocation. Result arrays have length one in that dimension; confidence limits have length two there. No broadcasting is performed for nonscalar paired samples.
Options and results
Name/value options are case-insensitive Alpha, Dim, and Tail. Alpha defaults to 0.05 and must lie strictly between zero and one. Tail is both (default), right, or left. Unknown options, incomplete pairs, and unsupported values raise an error; later duplicate options replace earlier ones.
[h,p,ci,stats] returns a logical decision h (p < Alpha), the selected Student-t tail probability, the confidence interval for the original population mean or paired mean difference, and a scalar structure with tstat, df, and sd arrays. The one-sided interval has an infinite bound on its untested side. Numeric outputs use host binary64, including single-input tests; this is not a claim of MATLAB single-output-class parity.
Execution and resource limits
Computation runs on the host. Only authenticated automatic residency may use the context-owned host gather fallback, for samples and numeric controls alike. Explicit gpuArray or unproven device intent is rejected before any argument transfer; call gather explicitly to request host analysis. There is no provider-native test kernel, fusion producer, distributed implementation, or separately enabled explicit-host-fallback extension. Gather errors and typed cancellation propagate without retry or relabeling.
Input scans and lane statistics cooperate with the owning execution context. Finite host-memory requests are rejected before gather and workspace allocation because this implementation has no complete lane/result allocation certificate; it does not fabricate a finite-memory grant.
Examples
Test a population mean
[h,p,ci,s] = ttest([1;2],0);Test paired differences
x = [2;4;6]; y = [1;2;3]; [h,p,ci,s] = ttest(x,y,'Tail','right');Using ttest with coding agents
Open a RunMat example with live inputs, then ask the agent to explain how ttest changes the result.
Run a small ttest example, explain the result, then change one input and compare the output.
Related Stats functions
Summary
anova1 · binocdf · binscatter · boxplot · cdf · cdfplot · chi2cdf · corr · corrcoef · corrcov · cov · cov2corr · dummyvar · ecdf · filloutliers · fitdist · geomean · grpstats · harmmean · icdf · isoutlier · ksdensity · kstest · kurtosis · lsline · mad · mode · nanmax · normalize · normcdf · norminv · normpdf · onehotdecode · onehotencode · pca · pdf · prctile · qqplot · quantile · refline · rmse · scatterhist · skewness · tabulate · tcdf · tiedrank · tinv · tpdf · ttest2 · wblinv
Ml
bayesopt · classify · confusionmat · crossvalind · cvpartition · fitclinear · fitctree · fitlm · kmeans · knnsearch · lasso · lassoglm · linkage · lscov · mnrfit · nlparci · optimizableVariable · pdist · pdist2 · perfcurve · predict · regress · ridge · squareform · test · training · tsne
Random
binornd · bootstrp · datasample · dividerand · exprnd · gamrnd · gpurng · lhsdesign · mvnrnd · normrnd · random · randsample · rng · trnd · unidrnd · unifrnd · wblrnd
Hist
About RunMat
RunMat is a runtime that executes MATLAB-syntax code fast on any GPU. The runtime is free for individuals, including freelancers; students and faculty at an academic institution, for teaching or research there; contributors to an open-source project, for work on that project; organizations with 250 or fewer people and US$10M or less in annual revenue; and anyone, for non-production use. See runmat.com/license.
- RunMat automatically optimizes your math for GPU execution on Apple, Nvidia, and AMD hardware. No code changes needed. Simulations that took hours now take minutes.
- Start running code in seconds. RunMat runs in the browser, on the desktop, or from the CLI. No license server, no IT ticket.