ChakataStat 0.11.0

ChakataStat 0.11.0 — Linux (.deb, x64).

Stable release - see the README for install steps and how to verify a
download against SHA256SUMS.txt.

Confirmatory factor analysis arrives: state the measurement model you
want to test — which observed indicators load on which latent factors —
and fit it by maximum likelihood, instead of letting the structure
emerge as exploratory Factor Analysis does. The output carries the free
loadings with their standard errors, the completely standardized
solution, factor correlations and variances, and the conventional fit
block: the model chi-square with its degrees of freedom and p value,
CFI, TLI, RMSEA with a 90% confidence interval, and SRMR. Your model is
checked for identification before anything is fitted, convergence is
reported honestly rather than assumed, and Heywood cases are flagged
instead of hidden. The numbers are validated against established
reference implementations. Find it under Analyze > Dimension Reduction >
Confirmatory Factor, with a full User Guide page on specifying a model
and reading the fit.

Meta-analysis gains a second random-effects estimator: REML pooling now
sits alongside DerSimonian-Laird in the Meta-Analysis dialog.

As always, everything new is reproducible: the same procedure is
available from the menus, syntax, the AI assistant, MCP and batch
scripts alike.

ChakataStat 0.10.0

ChakataStat 0.10.0 — Linux (.deb, x64).

Stable release - see the README for install steps and how to verify a
download against SHA256SUMS.txt.

Meta-analysis arrives: combine per-study effect sizes — one row per
study, an effect and its standard error on a common scale — into
fixed-effect and DerSimonian-Laird random-effects pooled estimates with
confidence intervals, plus the standard heterogeneity statistics
(Cochran's Q, I², τ²). Two new chart types come with it: the forest plot
draws every study with its confidence interval and a weight-sized marker
plus the pooled diamond, and the funnel plot maps each effect against
its precision with Egger's regression test for small-study asymmetry
reported alongside. Everything is validated against established
reference implementations. Find it under Analyze > Meta-Analysis, with a
full User Guide page covering the data layout and how to read the plots.

As always, everything new is reproducible: the same procedure — charts
included — is available from the menus, syntax, the AI assistant, MCP
and batch scripts alike.

ChakataStat 0.9.0

ChakataStat 0.9.0 — Linux (.deb, x64).

Stable release - see the README for install steps and how to verify a
download against SHA256SUMS.txt.

Mediation analysis arrives: estimate total, direct and indirect effects
(X on Y through one or several parallel mediators, with optional
covariates), with path coefficients for every component model and
bootstrap confidence intervals (percentile or bias-corrected BCa) for
each indirect effect — validated against reference implementations. Find
it under Analyze > Regression > Mediation.

Bar charts can now be stacked: with a cluster factor set, a new Stacked
bars option draws the second factor's segments in one bar per category —
the part-to-whole reading — instead of side by side.

This release also rides a full memory audit: the native engine and FFI
bridge verified leak-clean, several small memory leaks fixed in the UI
and on the engine boundary, and permanent leak guards added to the test
suite.

Windows is distributed through the Microsoft Store only; macOS remains
deferred.

ChakataStat 0.8.0

ChakataStat 0.8.0 — Linux (.deb, x64).

Stable release - see the README for install steps and how to verify a
download against SHA256SUMS.txt.

This release closes the expected-by-default gaps: five new chart types —
Bar Chart, Error Bar / means plot, Q-Q Plot (normal / Student t /
chi-square), a scree plot on Factor Analysis and a dendrogram on
Hierarchical Cluster; one-click regression diagnostic plots (residuals vs
fitted, residual Q-Q) on Linear Regression; and effect sizes reported by
default across the nonparametric tests (rank-biserial r, epsilon-squared,
Kendall's W) and the ANOVA family (partial eta-squared on factorial GLM,
Repeated Measures and MANOVA).

Windows is distributed through the Microsoft Store only; macOS remains
deferred.

ChakataStat 0.7.0

ChakataStat 0.7.0 — Linux (.deb, x64).

Stable release - see the README for install steps and how to verify a
download against SHA256SUMS.txt.

This release's theme is chart presentation control: histogram bin-limit
labels (edges / range / bar-start / custom numbering), legend placement and
series styling options on the Line Chart, and measured, backgrounded chart
legends that no longer clip long series names.

Windows is distributed through the Microsoft Store only; macOS remains
deferred.

ChakataStat 0.6.0

ChakataStat 0.6.0 — Linux (.deb, x64).

Stable release - see the README for install steps and how to verify a
download against SHA256SUMS.txt.

Windows distribution has moved to the Microsoft Store only as of this
release, so no direct-download Windows build is provided here. macOS
remains deferred.

ChakataStat 0.5.0

ChakataStat 0.5.0 — Windows (x64) and Linux (.deb, x64).

Stable release - see the README for install steps (including the
Windows SmartScreen notice) and how to verify a download against
SHA256SUMS.txt.