The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.
CoTiMA 1.0.7
- Join plot output filenames to their directory without relying on a
trailing separator.
CoTiMA 1.0.6
- Normalize temporary directory paths in the plotting test for
Windows.
CoTiMA 1.0.5
- Remove the obsolete variational Bayes setting from Stan defaults and
fit calls.
- Explicitly select Stan when the installed ctsem fitting API accepts
a backend.
- Omit historical optimizer defaults when ctsem has retired them.
CoTiMA 1.0.4
ctmaTables() now accepts fontSize = 10
(points) for titles, headers, cells and notes in Word, XML, TeX and PDF
exports.
Regular Stan table
compatibility (2026-10-29)
- Reads joint population DRIFT arrays from regular optimized and
sampled Stan fits.
- Reconstructs missing optimized-fit caches from retained raw draws
through the fitted Stan program, with Kalman filtering disabled and no
model refitting.
- Correctly labels native ML, MAP and Bayesian fits from their direct
args fields.
- Retains native-summary precision for four-decimal tables, avoids
partial cache name matching, and preserves a wrapper’s CT representation
during DT refresh.
- Validated against regular ctsem 3.11.1 and an existing 2000-draw
CoTiMA benchmark.
Table presentation
options (2026-10-29)
- layout selects portrait (long, default) or landscape (wide) pages
consistently across Word, XML, TeX and PDF exports. Oversized tables
continue in column panels.
- sortBy groups ctmaInit estimates by study (default) or by
parameter.
- rowLayout keeps one parameter per row (default), combines DRIFT
entries, or combines all selected parameters with grouped
parameter/statistic headers.
- Numeric long-format values and safe staged output publication are
preserved.
Saved-fit table summaries
(2026-10-29)
- ctmaTables defaults to four display decimals.
- Missing statistics and DT uncertainty can be regenerated from saved
native fits with summary(timeinterval=lag), with one consent for all
primary studies.
- recomputeSummary=“yes” permits batch use; no models are refitted and
raw point estimates are never treated as uncertainty draws.
Table
selectors and primary-study summaries (2026-10-28)
- Accepts bare, case-insensitive matrix families such as
parameters=c(drift, diffusion). Quoted selectors and character-vector
variables remain supported.
- Presents every primary study from a ctmaInit aggregate summary with
original study labels, retained uncertainty and explicit aggregate
fit-statistic notes; no refitting.
- Preserves unavailable statistics as unavailable and aligns
study/parameter text columns in Word and TeX exports.
APA tables (2026-10-28)
- Adds ctmaTables() with checked summary agreement, named/matrix
parameter selection, source-aware statistics, interactive selection and
APA-style Word/XML/TeX/PDF exports.
- Adds joint-draw DT effects at selected fitted-time lags and
coordinate-aligned model comparisons.
- Exports are staged; existing outputs require explicit overwrite.
Unsupported summaries and unavailable statistics fail explicitly rather
than fabricating transformed values.
- Documents optional Word/PDF dependencies and the limits of legacy
summary provenance.
Power analysis
refactoring (2026-10-28)
- Separates validation, invariant fitting/loading, matrix extraction,
implied moments, regression effects, power probabilities, integer
sample-size search and result assembly.
- Verifies unrestricted Gaussian regression eligibility for direct
covariance calculations. Supports explicit restricted Gaussian
regression syntax through a lavaan fallback, with console and stored
summary notes identifying the calculation route.
- Corrects negative-effect p-values, two-wave and higher-dimensional
calculations, lag bounds, reduced hypotheses, root boundaries and tied
optimal-lag selection.
- Fast default power includes both rejection tails. useMBESS=TRUE
retains the legacy probability convention. Missing post-hoc powers stay
NA instead of becoming zero.
- Restores original time units from invariant-fit scaling metadata.
Reduced-fit caches now validate covariance/model metadata and are reused
across requested power levels.
Plotting corrections
(2026-10-28)
- Correct forest pooled confidence intervals and funnel confidence
limits.
- Correct time-range indexing, axis positions, power-only plots and
fractional-lag selection.
- Handle multiple fits, legacy example objects and per-fit
moderator/lag mappings.
- Export uniquely named PNG files in the requested directory,
including headless sessions.
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.