Inference Suite: Discover and Bundle Every Applicable Inference Class for a Design
Source:R/inference_suite.R
InferenceSuite.RdA lightweight coordinator (not itself an Inference subclass, and not
part of the Inference R6 hierarchy) that pairs a single completed
Design object with the full set of concrete
Inference classes compatible with it. On construction, the suite
consults package-level inference metadata to discover every exported,
non-abstract Inference subclass whose declared response-type,
matched-design (KK), blocking, and censoring requirements are all satisfied
by des_obj, storing the resulting sorted class-name vector in
applicable_design_classes. Because discovery is driven by metadata
lookups rather than by actually attempting to construct each candidate
class, the applicable list automatically stays current as new inference
classes are registered elsewhere in the package, without this class needing
any changes, and without risking side effects (e.g. an optional-package
load failure inside some class's constructor) from a doomed construction
attempt.
Compatibility rules (see
is_inference_class_compatible_with_design_metadata(), also used by
Design's own
applicable_inference_class_names()): a candidate class is
excluded if it is abstract or not exported; if it declares no compatible
response types, or none match des_obj's response type; if its name
contains "KK" (a matched-design-only class) but des_obj does
not support KK matching; if the class's requires_blocking_design()
is TRUE (currently only InferenceIncidExtendedRobins –
InferenceIncidCMH works on both blocking and non-blocking designs
via different standard-error estimators, so it is not excluded)
but des_obj does not support blocking; or if des_obj has any
left-/interval-censored subjects (a finite y_R) but the class's
supports_interval_or_left_censored_data() is FALSE –
both of the latter two mirror Inference$initialize()'s own
construction-time gate exactly, via each class's registered
requires_blocking_design/supports_general_censoring
metadata (see infer_inference_requires_blocking_design()/
infer_inference_supports_general_censoring() in
inference_class_registry.R). Ordinary right-censoring alone never
excludes a class. The KK-name-pattern rule is still hardcoded in this
class rather than looked up from a central registry.
A class that is design-compatible but whose registered
required_packages are not all installed is excluded from
applicable_design_classes and reported separately, by class name,
in unavailable_due_to_missing_packages – so callers can tell "not
applicable to this design" apart from "applicable, but an optional
dependency isn't installed" (see the Discovery section of
fix_inference_hierarchy.md). Package availability is never a
reason a class is treated as design-incompatible.
Construction itself does not compute any estimates, p-values, or
confidence intervals – it only discovers and validates which inference
classes are applicable and does not eagerly construct any of them. This
class's run_all_inference() method does that: it constructs and
fits every applicable class and returns a uniform
comparison across them (see that method's own documentation for the
output schema, the CI/p-value method selection policy, and the
screen/html/plots/pdf/
save_results_as_JSON output options). lock_objects = FALSE
allows ad hoc fields to be attached to an instance after construction.
Every row this class discovers and fits is a test about the
same outcome variable, by construction. response_type is a
required, immutable constructor argument on Design
(read-only thereafter via get_response_type()), and this class
discovers every candidate in applicable_design_classes from one
attached Design object's one response_type – there is no
code path here that spans two response types in a single instance. This
matters beyond bookkeeping: a planned comparison-across-classes feature
(a single combined-evidence p-value summarizing every row, via a
dependence-robust combination test) relies on every combined class
sharing one sharp null of "no treatment effect on this outcome" –
which only holds when every test concerns the same outcome variable
(combining, say, a survival model's p-value with an unrelated
continuous-outcome model's p-value would not be valid, since a real
effect on one with none on the other is entirely plausible). That
precondition is guaranteed here structurally, not by caller discipline.
Same-Y does not mean same estimand, and that matters for
interpretation. Rows discovered here routinely test genuinely different
\(\theta_i\) on the same outcome – a mean difference, a log-odds
ratio under one link, a rank-based effect, a quantile shift – each
\(\theta_i\) its own parameterization's own null. The combined
\(H_0: \theta_1 = 0 \cap \dots \cap \theta_k = 0\) is one coherent
claim (Fisher's unit-level sharp null: no treatment effect
whatsoever) only under a randomization-based/exact procedure,
where every possible summary of "effect" is simultaneously zero by
construction. Asymptotic/likelihood-based procedures instead test a
weak, population-level null specific to their own
parameterization, and weak nulls for genuinely different summaries are
not generally nested – treatment can shift a distribution's variance or
a high quantile while leaving its mean exactly unchanged, so "mean
difference = 0" does not imply "quantile effect = 0" outside
location-shift-style models. This is comparatively safe across
different link functions for the same latent effect (e.g.
cauchit/probit/cloglog/logit on the same binary/ordinal response, which
typically share one underlying latent-variable null) and comparatively
riskier across different kinds of summary (a mean-difference test
combined with a rank-based test combined with a quantile-regression
test). Under that weak-null reading, a rejection is more honestly read
as "at least one specific summary of this outcome's distribution
differs" rather than "there is one coherent nonzero effect" – reinforcing,
not loosening, the interpretation caveat below.
Combined Evidence interpretation caveat (read before using
combined_evidence$pval): the Cauchy combination test is a
union-intersection test of \(H_0: \theta_1 = 0 \cap \theta_2 = 0 \cap
\dots \cap \theta_k = 0\) against the alternative that at least
one \(\theta_i \neq 0\). A significant combined_evidence$pval
is therefore evidence of an effect in at least one of these
senses, not evidence for a specific estimate or direction – it does
not say which class's estimand is nonzero, nor does a small combined
p-value imply every (or even most) constituent p-values were small. Do
not report combined_evidence$pval as if it estimated a single
effect size, and do not treat it as validating any one class's estimate
over another's; its only valid use is as evidence that *some* legitimate
way of looking for a treatment effect on this outcome found one.
This guarantee assumes every constituent p_i is itself valid (i.e.
uniform under its own null) – the Cauchy combination's dependence-robust
size control says nothing about which such p_i is small,
so a single miscalibrated or misspecified procedure (e.g. an asymptotic
approximation that breaks down for this sample) can dominate the combined
result the same way it would dominate a plain minimum-p-value test, even
when every other procedure shows nothing. A significant
combined_evidence$pval is worth cross-checking against the
per-estimand CI-forest plot (plots$ci_forest) before trusting it –
one outlying interval sitting apart from a cluster of concordant ones is a
miscalibration flag, not confirmed evidence.
This same-Y precondition is guaranteed within one
InferenceSuite instance structurally (one Design, one
response_type), but the architecture cannot stop a caller from
manually combining raw pvals pulled from two separate
InferenceSuite objects' results_tables outside
cct_combine_pvalues() – doing so is outside this function's
validity guarantee and is not a supported use of this metric.
References
Madigan, D., Ryan, P. B., and Schuemie, M. (2013), "Does design matter? Systematic evaluation of the impact of analytical choices on effect estimates in observational studies," Therapeutic Advances in Drug Safety, 4(2), 53-62, PMID 25083251 – the motivating finding behind this class's "every legitimate way to look for an effect, compared honestly" default output.
Liu, Y. and Xie, J. (2020), "Cauchy combination test: a powerful test
with analytic p-value calculation under arbitrary dependency
structures," Journal of the American Statistical Association,
115(529), 393-402 – run_all_inference()'s Combined Evidence
Metric.
Public fields
applicable_design_classesCharacter vector of applicable inference class names derived during initialization.
unavailable_due_to_missing_packagesA named list, one entry per otherwise-design-compatible class whose registered
required_packagesare not all installed: names are class names, values are the character vector of missing package names. These classes are excluded fromapplicable_design_classesbut are reported here separately from plain design/response-type incompatibility (see the class-level docs' "Discovery" rules), so callers can distinguish "not applicable to this design" from "applicable, but an optional dependency isn't installed." Empty named list if every design-compatible class has all required packages available.
Methods
InferenceSuite$new()
Discover every Inference class applicable to des_obj
(see the class-level documentation for the compatibility rules) and validate
any per-class constructor overrides in inference_params, storing
applicable_design_classes for later use. This constructor does
not instantiate any Inference object itself; callers are
expected to construct the specific classes they need (optionally passing the
validated inference_params) from the discovered list.
Usage
InferenceSuite$new(des_obj, model_formula = NULL, inference_params = list())Arguments
des_objA completed
Designobject (validated viais(des_obj, "Design")when assertions are enabled; seetoggle_asserts).model_formulaAccepted for interface/future-extension purposes but currently not used anywhere in this method's body – supplying a non-
NULLvalue has no effect on discovery, validation, or any stored state. Do not rely on this parameter to affect covariate adjustment; a design's own model formula and design matrix are what individualInferenceclasses actually consult when later constructed from this suite's discovered class list.inference_paramsA named list of lists supplying additional constructor arguments for specific inference classes. Each name must be the name of a concrete
Inferencesubclass that is applicable todes_obj(checked againstapplicable_design_classesonce discovered – an inapplicable class name raises an error); the corresponding list contains keyword arguments (beyonddes_obj) forwarded to that class'sinitialize, and every argument name supplied must match a formal parameter of that class'sinitializemethod (other thandes_objand...) or an error is raised naming the unknown argument(s) and the valid ones. Defaults to an empty list (no extra arguments for any class).
InferenceSuite$run_all_inference()
Construct and fit every class in applicable_design_classes,
and report one uniform comparison row per class – estimate, SE, CI, p-value
(each via the highest-priority available method; see
inference_suite_inspect.md's "Method Selection Policy"), likelihood
tier, estimand (where declared), fit time, captured warnings, and status.
Unlike the constructor, this method fits models and is not free; call it
explicitly when you want the comparison, not automatically.
A single class's construction or fit failure never aborts the report – it
is caught and recorded as that class's status/message (see
"Per-Class Failure Isolation" in the design doc).
Side effects (v1.0.0 slice): screen prints each row as its
class finishes fitting (computation order), not buffered to the end, with a
percent-done/estimated-time-remaining progress bar line underneath each row
(the ETA is the mean per-class elapsed time so far times classes remaining),
followed by a footer listing classes excluded for missing optional packages.
html = TRUE writes a self-contained, timestamped HTML report (the
same table plus the same footer) to output_dir
and opens it via browseURL; it requires the
knitr package. The plots ggplot2 visualizations, and their
embedding into this HTML report, are not yet implemented
(inference_suite_inspect.md TODO-7); pdf output is not yet a
parameter of this method.
Usage
InferenceSuite$run_all_inference(
screen = TRUE,
html = FALSE,
alpha = 0.05,
save_results_as_JSON = FALSE,
plots = screen,
pdf = FALSE,
classes = NULL,
exclude_classes = character(),
max_secs_per_class = NULL,
num_cores = 1L,
formulas = NULL,
methods = NULL,
basic_bootstrap = FALSE,
compute_conf_intervals = FALSE,
output_dir = "~",
combined_evidence_estimands = NULL,
combined_evidence_weighting = c("estimand_grouped", "equal", "custom"),
combined_evidence_weights = NULL
)Arguments
screenPrint results to the console as each class finishes. At least one of
screen/htmlmust beTRUE.htmlRender, save (
output_dir, timestamped filename), and auto-open a self-contained HTML report of the results.alphaSignificance level: confidence intervals are computed at
1 - alphaandalphais the significance threshold used anywhere the report flags significance. Default0.05.save_results_as_JSONIf
TRUE, serialize the return object (excluding plot objects) to a timestamped JSON file inoutput_dir. Requires the optional jsonlite package; if it is not installed, awarning()is issued and this artifact is skipped rather than erroring. DefaultFALSE.plotsIf
TRUE, build and display (on the current graphics device) one visualization per estimand: an annotated confidence-interval forest plot (p-value left of each interval, interval width right of it, class/method label right-aligned on each row, color-keyed to significance atalpha) stacked over its own “Estimates” box-and-whisker subplot – a free x-axis (same label, same log10/linear choice as the forest, but its own limits) summarizing the point estimates, one point per inference class/formula, collapsed over method/type since those share one estimate. That subplot scales with how many distinct estimates there are: none for a single estimate (redundant with the forest's own dot), dots alone for 2-5, and dots over a box-and-whisker for more than 5. Built with ggplot2 and stacked into a single gtable grob (draw withgrid::grid.draw()); requires the optional ggplot2 package, if it is not installed, awarning()is issued and plotting is skipped rather than erroring. Defaults to the value ofscreen.pdfIf
TRUE, save the visualization to one timestamped multi-page PDF file inoutput_dir(one page per estimand; page height scales with the largest estimand's number of CI rows). Same ggplot2 dependency and missing-package handling asplots. DefaultFALSE.classesOptional character vector restricting which applicable classes to fit – e.g. re-running against only the few classes a user is actually deciding between, without reconstructing the suite. Every name must already be in
applicable_design_classesor this errors, naming the unknown name(s) and the valid ones.NULL(default) fits every applicable class.exclude_classesOptional character vector of applicable classes to skip, applied after
classes. Same validation asclasses. Default none.max_secs_per_classOptional per-class elapsed-time limit in seconds (via
setTimeLimit), after which that class's row getsstatus = "timeout"instead of hanging the whole report. Protects against one pathological class (e.g. a bootstrap/randomization method with many replicates) blocking every other class. Known limitation: R's time limits are checked at R-level interrupt points, so this reliably cuts off slow R-level work but is not guaranteed to interrupt one very slow single native (C/C++/BLAS) call with no intervening R-level check.NULL(default) means no limit.num_coresIf greater than
1, fit classes in parallel across this many forked workers (makeForkCluster) – Unix/Linux only; on other platforms this falls back to sequential (num_cores = 1) with awarning(). Screen output changes under parallel execution: fitting is a single blocking call that only returns once every worker has finished, so there is no meaningful per-class ETA to show while running –screen = TRUEinstead prints a "fitting N classes across K workers" message up front, then every result row together once complete, then a total-elapsed-time summary line (a deliberate design choice, not a degraded default – seeinference_suite_inspect.md's TODO-13). Default1L(sequential, with the normal incremental streaming/progress bar).formulasNULL(default), a single formula (~ .), a single formula string ("~ ."), or a collection of either – includingc(~ 1, ~ .), which base R already returns as a plainlistofformulaobjects (formulas have noc()method of their own), or a character vector (c("~ .", "~ age + sex * smoking")).NULLmeans each class fits once with its own default formula – identical to omitting this argument entirely, sincemodel_formula = NULLat construction already resolves todes_obj$get_design_formula()(default~ .). When non-NULL, only classes whose constructor syntactically accepts amodel_formulaargument are fit once per formula informulas(oneresults_tablerow each, disambiguated inresultsby"<class>[<formula>]"names); classes without amodel_formulaconstructor argument at all still fit exactly once, ignoringformulas. Note this is a syntactic check (does the constructor accept one), not a semantic one (does the fit actually use it) – some classes accept-and-ignoremodel_formula(e.g.InferenceAllSimpleAverageDiff's unadjusted Welch's t-test); seefix_inference_hierarchy.md'sadjusts_for_covariatesregistry-metadata audit, which makes thecov_modelcolumn semantics-aware wherever that audit has landed.methodsNULL(default), a character vector of method sentinel strings, or (TODO-22) a named list,sentinelto a character vector of requestedtypevalues, orNULL, restricting which inference method(s) – and, for the three resampling sentinels marked "typed" below, which resampling/CI- constructiontypeflavor(s) – get fit and reported per class.NULLconsiders every sentinel inEDI_INFERENCE_SUITE_METHOD_SENTINELS, and for each typed sentinel, everytypevalue that class supports (queried at runtime via its ownget_supported_bootstrap_ {pval,ci}_types()/get_supported_bayesian_bootstrap_ {pval,ci}_types()/get_supported_rand_bootstrap_ {pval,ci}_types()accessor – never a hardcoded type table in this package), except class/method/type combinations declared inEDI_COMPREHENSIVE_SLOW_PATHS. Those implemented but prohibitively slow paths are omitted only from this default selection. Supplyingmethodsexplicitly opts into the named sentinel/type combinations even when the registry marks them slow. Thus the default remains broad without allowing known multi-minute paths to dominate a routine report; it is not a single "best available" cascade. List-shaped example:methods = list(bootstrap = c("percentile", "bca"), rand_bootstrap = NULL)fits only"bootstrap"(restricted to the"percentile"/"bca"types that class actually supports) and"rand_bootstrap"(every type that class supports); a sentinel present as a list name with valueNULLstill means "every valid type" for it, exactly like the flat-vector shape – to not fit a sentinel at all, simply don't name it. Requesting atypefor a sentinel with notypeaxis (any sentinel not marked "typed" below) errors. Valid sentinels, each corresponding to one asymptotic/exact/ randomization/resampling inference family a class may (or may not) implement:"wald"Asymptotic Wald inference –
compute_asymp_confidence_interval()/compute_asymp_two_sided_pval()(capability"wald"). The standard closed-form normal-approximation CI/test."exact"Exact inference –
compute_exact_confidence_interval()/compute_exact_two_sided_pval_for_treatment_effect()(capability"exact_test"). Finite-sample-exact methods (e.g. Fisher's exact test, exact binomial)."rand"Randomization inference –
compute_rand_confidence_interval()/compute_rand_two_sided_pval()(capabilities"randomization_ci"/"randomization_test"– distinct capability names for the CI vs. p-value side, since a class can support one without the other). Design-based inference via re-randomizing the observed treatment assignment."rand_bootstrap"(typed)Randomization-bootstrap inference –
compute_rand_bootstrap_confidence_interval()/compute_rand_bootstrap_two_sided_pval()(capabilities"randomization_bootstrap_ci"/"randomization_bootstrap"– distinct capability names for the CI vs. p-value side, since a class can support one without the other). Resamples under the randomization null rather than the usual iid-resampling bootstrap.type(both sides agree on the same four values, unlike"bootstrap"/"bayes_boot"below):"percentile","studentized","symmetric-percentile-t","smoothed"."jackknife"Jackknife-Wald inference –
compute_jackknife_wald_confidence_interval()/compute_jackknife_wald_two_sided_pval()(capability"jackknife"). Leave-one-out variance estimate feeding a Wald-style CI/test."score"Score (Rao) test –
compute_score_confidence_interval()/compute_score_two_sided_pval()(capability"likelihood_tests", one of its three independent sub-procedures)."lik_ratio"Likelihood-ratio test –
compute_lik_ratio_confidence_interval()/compute_lik_ratio_two_sided_pval()(capability"likelihood_tests")."gradient"Gradient test –
compute_gradient_confidence_interval()/compute_gradient_two_sided_pval()(capability"likelihood_tests").
(The plain, "best available" auto-selecting
compute_lik_ratio_bartlett_two_sided_pval()/
compute_lik_ratio_bartlett_confidence_interval()
dispatcher is deliberately not its own sentinel -- it only picks
between the two explicit variants below depending on which this
class implements, so it is never a distinct inference procedure.)
"lik_ratio_bartlett_approx"Bartlett-corrected likelihood-ratio test, Monte-Carlo-approximated correction factor pinned explicitly (for reproducibility) –
compute_lik_ratio_bartlett_approx_confidence_interval()/compute_lik_ratio_bartlett_approx_two_sided_pval()(capability"likelihood_tests"; degrades toNAfor classes without an approximate Bartlett factor)."lik_ratio_bartlett_exact"Bartlett-corrected likelihood-ratio test, closed-form analytic correction factor pinned explicitly –
compute_lik_ratio_bartlett_exact_confidence_interval()/compute_lik_ratio_bartlett_exact_two_sided_pval()(capability"likelihood_tests"; degrades toNAfor classes without an exact Bartlett factor)."param_boot"Bootstrap-calibrated likelihood-ratio test –
compute_lik_ratio_bootstrap_confidence_interval()/compute_lik_ratio_bootstrap_two_sided_pval()(capability"parametric_likelihood_bootstrap")."param_boot_direct"Direct parametric-bootstrap estimate/CI/pval for the treatment coefficient itself –
compute_param_bootstrap_confidence_interval()/compute_param_bootstrap_pval()(capability"parametric_likelihood_bootstrap"; distinct from"param_boot"above, which is a bootstrap-calibrated likelihood-ratio test, not a direct estimate)."bayes_boot"(typed)Bayesian bootstrap inference –
compute_bayesian_bootstrap_confidence_interval()/compute_bayesian_bootstrap_two_sided_pval()(capability"bayesian_bootstrap"). CI-sidetype:"percentile","basic","wald","studentized","bootstrap-t","bca"; pval-sidetypeswaps"basic"for"symmetric"(all others the same)."bootstrap"(typed)Nonparametric bootstrap inference –
compute_bootstrap_confidence_interval()/compute_bootstrap_two_sided_pval()(capability"nonparametric_bootstrap"). CI-sidetype:"percentile","basic","studentized","bootstrap-t","symmetric-percentile-t","bca","prepivoted","double-bootstrap","calibrated","smoothed"; pval-sidetypeis a smaller set –"percentile","symmetric","studentized","bootstrap-t","bca"– neither"basic"nor the other CI-only variants apply on the pval side.
For the three "typed" sentinels above, an exhaustive
typelist is documented here for orientation only – the actual set consulted at runtime always comes from that class's own accessor (see the top of this section), so a class need not support every value listed. ("likelihood_ratio"/"estimating_equation_likelihood_ratio"are deliberately not separate sentinels – both capabilities gate the exact same method pair"lik_ratio"above already covers.) For each class, only sentinels the class has any CI or p-value capability for (among the requestedmethods) get a row; for a typed sentinel, one row pertypethat class actually supports (intersected with any requested type subset) – a class with zero applicable sentinels, or a typed sentinel with zero resulting types, still gets exactly one row withmethod/type = NA_character_(mirrors the pre-methods"no capability" row) rather than being silently dropped. A class contributing more than one applicable-sentinel row is disambiguated inresults/results_tableby"<class>{<method>}"or"<class>{<method>:<type>}"(or with a"[<formula>]"tag too under simultaneousformulasfan-out) names. Unlike the removed cascade,ci_method/pval_methodon a given row now always match that row's ownmethod(or areNAif this class lacks that half of the sentinel's capability, including whentypeis valid on one side but not the other) – there is no fallback to a different sentinel within one row.basic_bootstrapFALSE(default). Convenience flag: whenTRUE, restricts every typed sentinel ("bootstrap"/"bayes_boot"/"rand_bootstrap") to just that class's first (i.e. default)typevalue instead of fitting everytypeit supports – "just run the default bootstrap flavor for nonparametric/Bayesian/randomization resampling" without having to spell outmethods = list(bootstrap = ..., bayes_boot = ..., rand_bootstrap = ...)by hand. Only takes effect for a typed sentinel the caller didn't already restrict via an explicitmethodslist entry – an explicittyperequest there always wins over this flag. No effect on non-typed sentinels ("param_boot"/"param_boot_direct"included – neither has atypeaxis, so they already run their one procedure).compute_conf_intervalsFALSE(default). WhenFALSE, every task's confidence-interval side (ci_a/ci_b/ci_method) is skipped entirely – only the p-value side runs. Several sentinels' CI search (Bartlett-approx,"rand","rand_bootstrap","param_boot") re-invokes the same expensive machinery as its own p-value roughly 15-40 times per bound during root-finding, by far the dominant cost of a fullrun_all_inference()run for those sentinels; skipping it can cut total runtime dramatically.ci_a/ci_b/ci_methodstay present but alwaysNAinresults_table(stable schema either way) and are omitted entirely from the live/print/HTML display tables whenFALSE. SetTRUEto compute CIs as before.output_dirDirectory for the
html/pdf/save_results_as_JSONoutput files. Default"~"(the user's home directory), not the current working directory – these calls are routinely made from inside the package's own source tree (a demo/dev script run fromR/EDI/), and a stray timestamped report left ingetwd()there is exactly the kind of untracked file that can end up bundled into a source tarball (R CMD build) or committed by accident. Pass an explicit path (e.g. the current directory, or a temp directory) to write elsewhere.combined_evidence_estimandsNULL(default: include every declaredestimand), or a character vector ofestimandvalues to restrict the Combined Evidence p-value/weights to. Validated argument-time against theestimandvalues actually declared amongclasses/exclude_classes-filtered candidates.combined_evidence_weightingOne of
"estimand_grouped"(default –w_i = 1 / (G * m_i)),"equal"(flatw_i = 1/k), or"custom"(caller suppliescombined_evidence_weights). Seeinference_suite_inspect.md's TODO-15.combined_evidence_weightsNamed numeric vector (
inference_classname -> weight), required when and only whencombined_evidence_weighting = "custom". Names must be a subset of the classes being fit; an unnamed usable class defaults to weight0(excluded). Need not pre-sum to 1.
Returns
Invisibly, an object of class c("EDIInferenceSuiteResults", "list")
with elements results (one named sub-list per class, in computation
order), results_table (the same rows as a flat data.frame,
sorted/grouped by estimand – NA_character_ last – with a
secondary sort by inference_class; includes the weight
column driven by combined_evidence_weighting/
combined_evidence_estimands), combined_evidence
(list(pval, stat, method = "cauchy_combination", n_classes_used,
n_estimand_groups, estimands_used, weighting, weights_used,
classes_used) – the Cauchy-combination-test p-value/statistic
across all usable rows under the resolved weighting policy;
weights_used/classes_used are keyed/valued by each
row's results name, not results_table$inference_class
directly, since that column can repeat under formulas;
pval = stat = NA_real_ if fewer than 2 rows are usable),
design, alpha, unavailable_due_to_missing_packages,
plots (list(ci_forest); ci_forest is a named list
of one gtable grob per estimand – the CI forest
stacked over its “Estimates” box-and-whisker subplot; draw
with grid::grid.draw() or pass to ggplot2::ggsave() –
possibly empty – rather than a single plot, since the visualization
is split one-per-estimand, per user request, 2026-08-19; the former
separate estimates plot became that subplot, 2026-08-21),
files (list(html, pdf, json), each a
path or NULL; pdf is one multi-page PDF with one page
per estimand), timestamp, total_secs, and
edi_version.
Examples
# \donttest{
seq_des = DesignSeqOneByOneBernoulli$new(n = 20, response_type = "continuous")
for (i in 1:20) {
seq_des$add_one_subject_to_experiment_and_assign(data.frame(x = rnorm(1)))
}
seq_des$add_all_subject_responses(rnorm(20))
suite = InferenceSuite$new(seq_des)
suite$applicable_design_classes
#> [1] "InferenceAllSimpleAverageDiff" "InferenceAllSimpleMeanDiffPooledVar"
#> [3] "InferenceAllSimpleWilcox" "InferenceContinLin"
#> [5] "InferenceContinOLS" "InferenceContinQuantileRegr"
#> [7] "InferenceContinRobustRegr"
# Fit and compare every applicable class:
results = suite$run_all_inference(screen = TRUE)
#> inference cov estimand est se pval pval method status
#> class mod
#> ===========================================================================================
#> Classes 0/154 [ 0% ] Status: Estimating...
Avg Δ mean Δ -0.153 0.569 7.95e-01 wald ok
#> Classes 1/154 [ 0% ] Estimated Time Left: 3s
Avg Δ mean Δ -0.153 0.569 7.43e-01 rand ok
#> Classes 2/154 [ 1% ] Estimated Time Left: 3s
Avg Δ mean Δ -0.153 0.569 7.03e-01 rand boot (%i… ok
#> Classes 3/154 [ 1% ] Estimated Time Left: 9s
Avg Δ mean Δ -0.153 0.569 2.81e-01 rand boot (st… ok
#> Classes 4/154 [ 2% ] Estimated Time Left: 11s
Avg Δ mean Δ -0.153 0.569 9.48e-01 rand boot (sy… ok
#> Classes 5/154 [= 3% ] Estimated Time Left: 13s
Avg Δ mean Δ -0.153 0.569 7.54e-01 rand boot (sm… ok
#> Classes 6/154 [= 3% ] Estimated Time Left: 17s
Avg Δ mean Δ -0.153 0.569 7.97e-01 jackknife ok
#> Classes 7/154 [= 4% ] Estimated Time Left: 15s
Avg Δ mean Δ -0.153 0.569 7.78e-01 bayes boot (%… ok
#> Classes 8/154 [= 5% ] Estimated Time Left: 16s
Avg Δ mean Δ -0.153 0.569 NA bayes boot (b… ok
#> Classes 9/154 [= 5% ] Estimated Time Left: 14s
Avg Δ mean Δ -0.153 0.569 7.67e-01 bayes boot (w… ok
#> Classes 10/154 [== 6% ] Estimated Time Left: 14s
Avg Δ mean Δ -0.153 0.569 7.35e-01 bayes boot (s… ok
#> Classes 11/154 [== 7% ] Estimated Time Left: 14s
Avg Δ mean Δ -0.153 0.569 6.79e-01 bayes boot (t) ok
#> Classes 12/154 [== 7% ] Estimated Time Left: 13s
Avg Δ mean Δ -0.153 0.569 7.75e-01 bayes boot (b… ok
#> Classes 13/154 [== 8% ] Estimated Time Left: 14s
Avg Δ mean Δ -0.153 0.569 7.62e-01 bayes boot (s… ok
#> Classes 14/154 [== 9% ] Estimated Time Left: 14s
Avg Δ mean Δ -0.153 0.569 7.47e-01 boot (%ile) ok
#> Classes 15/154 [=== 9% ] Estimated Time Left: 14s
Avg Δ mean Δ -0.153 0.569 NA boot (basic) ok
#> Classes 16/154 [=== 10% ] Estimated Time Left: 13s
Avg Δ mean Δ -0.153 0.569 NA boot (symm t) ok
#> Classes 17/154 [=== 11% ] Estimated Time Left: 12s
Avg Δ mean Δ -0.153 0.569 7.22e-01 boot (bca) ok
#> Classes 18/154 [=== 11% ] Estimated Time Left: 13s
Avg Δ mean Δ -0.153 0.569 NA boot (prepiv) ok
#> Classes 19/154 [=== 12% ] Estimated Time Left: 12s
Avg Δ mean Δ -0.153 0.569 NA boot (dbl-boo… ok
#> Classes 20/154 [==== 12% ] Estimated Time Left: 11s
Avg Δ mean Δ -0.153 0.569 NA boot (calib) ok
#> Classes 21/154 [==== 13% ] Estimated Time Left: 11s
Avg Δ mean Δ -0.153 0.569 NA boot (smth) ok
#> Classes 22/154 [==== 14% ] Estimated Time Left: 10s
Avg Δ mean Δ -0.153 0.569 8.08e-01 boot (symm) ok
#> Classes 23/154 [==== 14% ] Estimated Time Left: 11s
Avg Δ Pooled … mean Δ -0.153 0.483 7.56e-01 wald ok
#> Classes 24/154 [==== 15% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 7.43e-01 rand ok
#> Classes 25/154 [===== 16% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 6.91e-01 rand boot (%i… ok
#> Classes 26/154 [===== 16% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 2.34e-01 rand boot (st… ok
#> Classes 27/154 [===== 17% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 9.44e-01 rand boot (sy… ok
#> Classes 28/154 [===== 18% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 7.27e-01 rand boot (sm… ok
#> Classes 29/154 [===== 18% ] Estimated Time Left: 11s
Avg Δ Pooled … mean Δ -0.153 0.483 7.97e-01 jackknife ok
#> Classes 30/154 [====== 19% ] Estimated Time Left: 11s
Avg Δ Pooled … mean Δ -0.153 0.483 7.31e-01 bayes boot (%… ok
#> Classes 31/154 [====== 20% ] Estimated Time Left: 11s
Avg Δ Pooled … mean Δ -0.153 0.483 NA bayes boot (b… ok
#> Classes 32/154 [====== 20% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 7.61e-01 bayes boot (w… ok
#> Classes 33/154 [====== 21% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 7.03e-01 bayes boot (s… ok
#> Classes 34/154 [====== 22% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 6.81e-01 bayes boot (t) ok
#> Classes 35/154 [======= 22% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 8.06e-01 bayes boot (b… ok
#> Classes 36/154 [======= 23% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 7.37e-01 bayes boot (s… ok
#> Classes 37/154 [======= 24% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 7.39e-01 boot (%ile) ok
#> Classes 38/154 [======= 24% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 NA boot (basic) ok
#> Classes 39/154 [======= 25% ] Estimated Time Left: 10s
Avg Δ Pooled … mean Δ -0.153 0.483 7.90e-01 boot (stud) ok
#> Classes 40/154 [======== 25% ] Estimated Time Left: 27s
Avg Δ Pooled … mean Δ -0.153 0.483 8.10e-01 boot (t) ok
#> Classes 41/154 [======== 26% ] Estimated Time Left: 44s
Avg Δ Pooled … mean Δ -0.153 0.483 NA boot (symm t) ok
#> Classes 42/154 [======== 27% ] Estimated Time Left: 43s
Avg Δ Pooled … mean Δ -0.153 0.483 7.94e-01 boot (bca) ok
#> Classes 43/154 [======== 27% ] Estimated Time Left: 42s
Avg Δ Pooled … mean Δ -0.153 0.483 NA boot (prepiv) ok
#> Classes 44/154 [======== 28% ] Estimated Time Left: 41s
Avg Δ Pooled … mean Δ -0.153 0.483 NA boot (dbl-boo… ok
#> Classes 45/154 [========= 29% ] Estimated Time Left: 39s
Avg Δ Pooled … mean Δ -0.153 0.483 NA boot (calib) ok
#> Classes 46/154 [========= 29% ] Estimated Time Left: 38s
Avg Δ Pooled … mean Δ -0.153 0.483 NA boot (smth) ok
#> Classes 47/154 [========= 30% ] Estimated Time Left: 37s
Avg Δ Pooled … mean Δ -0.153 0.483 8.02e-01 boot (symm) ok
#> Classes 48/154 [========= 31% ] Estimated Time Left: 36s
Wilcox HL shift -0.0378 0.636 9.69e-01 wald ok
#> Classes 49/154 [========= 31% ] Estimated Time Left: 35s
Wilcox HL shift -0.0378 0.636 9.26e-01 rand ok
#> Classes 50/154 [========== 32% ] Estimated Time Left: 35s
Wilcox HL shift -0.0378 0.636 7.58e-01 rand boot (%i… ok
#> Classes 51/154 [========== 33% ] Estimated Time Left: 34s
Wilcox HL shift -0.0378 0.636 9.06e-01 rand boot (st… ok
#> Classes 52/154 [========== 33% ] Estimated Time Left: 43s
Wilcox HL shift -0.0378 0.636 7.98e-01 rand boot (sy… ok
#> Classes 53/154 [========== 34% ] Estimated Time Left: 52s
Wilcox HL shift -0.0378 0.636 9.02e-01 boot (%ile) ok
#> Classes 54/154 [========== 35% ] Estimated Time Left: 51s
Wilcox HL shift -0.0378 0.636 NA boot (basic) ok
#> Classes 55/154 [=========== 35% ] Estimated Time Left: 49s
Wilcox HL shift -0.0378 0.636 9.64e-01 boot (stud) ok
#> Classes 56/154 [========== 36% ] Estimated Time Left: 1m 9s
Wilcox HL shift -0.0378 0.636 9.54e-01 boot (t) ok
#> Classes 57/154 [========== 37% ] Estimated Time Left: 1m 26s
Wilcox HL shift -0.0378 0.636 NA boot (symm t) ok
#> Classes 58/154 [========== 37% ] Estimated Time Left: 1m 24s
Wilcox HL shift -0.0378 0.636 9.74e-01 boot (bca) ok
#> Classes 59/154 [========== 38% ] Estimated Time Left: 1m 22s
Wilcox HL shift -0.0378 0.636 NA boot (prepiv) ok
#> Classes 60/154 [========== 38% ] Estimated Time Left: 1m 20s
Wilcox HL shift -0.0378 0.636 NA boot (dbl-boo… ok
#> Classes 61/154 [=========== 39% ] Estimated Time Left: 1m 18s
Wilcox HL shift -0.0378 0.636 NA boot (calib) ok
#> Classes 62/154 [=========== 40% ] Estimated Time Left: 1m 16s
Wilcox HL shift -0.0378 0.636 NA boot (smth) ok
#> Classes 63/154 [=========== 40% ] Estimated Time Left: 1m 14s
Wilcox HL shift -0.0378 0.636 9.72e-01 boot (symm) ok
#> Classes 64/154 [=========== 41% ] Estimated Time Left: 1m 12s
Lin ~. mean Δ -0.174 0.603 7.76e-01 wald ok
#> Classes 65/154 [=========== 42% ] Estimated Time Left: 1m 10s
Lin ~. mean Δ -0.174 0.603 7.27e-01 rand ok
#> Classes 66/154 [============ 42% ] Estimated Time Left: 1m 8s
Lin ~. mean Δ -0.174 0.603 7.11e-01 rand boot (%i… ok
#> Classes 67/154 [============ 43% ] Estimated Time Left: 1m 7s
Lin ~. mean Δ -0.174 0.603 7.70e-01 rand boot (st… ok
#> Classes 68/154 [============ 44% ] Estimated Time Left: 1m 6s
Lin ~. mean Δ -0.174 0.603 7.64e-01 rand boot (sy… ok
#> Classes 69/154 [============ 44% ] Estimated Time Left: 1m 5s
Lin ~. mean Δ -0.174 0.603 7.62e-01 rand boot (sm… ok
#> Classes 70/154 [============ 45% ] Estimated Time Left: 1m 4s
Lin ~. mean Δ -0.174 0.603 7.85e-01 jackknife ok
#> Classes 71/154 [============ 46% ] Estimated Time Left: 1m 2s
Lin ~. mean Δ -0.174 0.603 7.33e-01 score ok
#> Classes 72/154 [============ 46% ] Estimated Time Left: 1m 0s
Lin ~. mean Δ -0.174 0.603 7.33e-01 lik_ratio ok
#> Classes 73/154 [============= 47% ] Estimated Time Left: 59s
Lin ~. mean Δ -0.174 0.603 7.33e-01 gradient ok
#> Classes 74/154 [============= 48% ] Estimated Time Left: 57s
Lin ~. mean Δ -0.174 0.603 7.07e-01 LR ≈Bartlett ok
#> Classes 75/154 [============= 48% ] Estimated Time Left: 56s
Lin ~. mean Δ -0.174 0.603 7.20e-01 param boot ok
#> Classes 76/154 [============= 49% ] Estimated Time Left: 55s
Lin ~. mean Δ -0.174 0.603 NA param boot ok
#> Classes 77/154 [============= 50% ] Estimated Time Left: 53s
Lin ~. mean Δ -0.174 0.603 6.63e-01 bayes boot (%… ok
#> Classes 78/154 [============= 50% ] Estimated Time Left: 52s
Lin ~. mean Δ -0.174 0.603 NA bayes boot (b… ok
#> Classes 79/154 [============= 51% ] Estimated Time Left: 51s
Lin ~. mean Δ -0.174 0.603 7.24e-01 bayes boot (w… ok
#> Classes 80/154 [============= 51% ] Estimated Time Left: 50s
Lin ~. mean Δ -0.174 0.603 7.98e-01 bayes boot (s… ok
#> Classes 81/154 [============= 52% ] Estimated Time Left: 49s
Lin ~. mean Δ -0.174 0.603 7.56e-01 bayes boot (t) ok
#> Classes 82/154 [============= 53% ] Estimated Time Left: 47s
Lin ~. mean Δ -0.174 0.603 6.81e-01 bayes boot (b… ok
#> Classes 83/154 [============= 53% ] Estimated Time Left: 46s
Lin ~. mean Δ -0.174 0.603 7.21e-01 bayes boot (s… ok
#> Classes 84/154 [============= 54% ] Estimated Time Left: 45s
Lin ~. mean Δ -0.174 0.603 7.47e-01 boot (%ile) ok
#> Classes 85/154 [============= 55% ] Estimated Time Left: 44s
Lin ~. mean Δ -0.174 0.603 NA boot (basic) ok
#> Classes 86/154 [============= 55% ] Estimated Time Left: 43s
Lin ~. mean Δ -0.174 0.603 7.90e-01 boot (stud) ok
#> Classes 87/154 [============= 56% ] Estimated Time Left: 47s
Lin ~. mean Δ -0.174 0.603 7.60e-01 boot (t) ok
#> Classes 88/154 [============= 57% ] Estimated Time Left: 51s
Lin ~. mean Δ -0.174 0.603 NA boot (symm t) ok
#> Classes 89/154 [============= 57% ] Estimated Time Left: 50s
Lin ~. mean Δ -0.174 0.603 7.99e-01 boot (bca) ok
#> Classes 90/154 [============= 58% ] Estimated Time Left: 49s
Lin ~. mean Δ -0.174 0.603 NA boot (prepiv) ok
#> Classes 91/154 [============= 59% ] Estimated Time Left: 48s
Lin ~. mean Δ -0.174 0.603 NA boot (dbl-boo… ok
#> Classes 92/154 [============= 59% ] Estimated Time Left: 46s
Lin ~. mean Δ -0.174 0.603 NA boot (calib) ok
#> Classes 93/154 [============= 60% ] Estimated Time Left: 45s
Lin ~. mean Δ -0.174 0.603 NA boot (smth) ok
#> Classes 94/154 [============= 61% ] Estimated Time Left: 44s
Lin ~. mean Δ -0.174 0.603 7.68e-01 boot (symm) ok
#> Classes 95/154 [============= 61% = ] Estimated Time Left: 43s
OLS ~. mean Δ -0.173 0.496 7.31e-01 wald ok
#> Classes 96/154 [============= 62% = ] Estimated Time Left: 42s
OLS ~. mean Δ -0.173 0.496 7.23e-01 rand ok
#> Classes 97/154 [============= 62% = ] Estimated Time Left: 41s
OLS ~. mean Δ -0.173 0.496 7.27e-01 rand boot (%i… ok
#> Classes 98/154 [============= 63% = ] Estimated Time Left: 40s
OLS ~. mean Δ -0.173 0.496 7.50e-01 rand boot (st… ok
#> Classes 99/154 [============= 64% = ] Estimated Time Left: 39s
OLS ~. mean Δ -0.173 0.496 7.35e-01 rand boot (sy… ok
#> Classes 100/154[============= 64% == ] Estimated Time Left: 38s
OLS ~. mean Δ -0.173 0.496 7.11e-01 rand boot (sm… ok
#> Classes 101/154[============= 65% == ] Estimated Time Left: 37s
OLS ~. mean Δ -0.173 0.496 7.80e-01 jackknife ok
#> Classes 102/154[============= 66% == ] Estimated Time Left: 36s
OLS ~. mean Δ -0.173 0.496 7.27e-01 score ok
#> Classes 103/154[============= 66% == ] Estimated Time Left: 35s
OLS ~. mean Δ -0.173 0.496 7.27e-01 lik_ratio ok
#> Classes 104/154[============= 67% == ] Estimated Time Left: 34s
OLS ~. mean Δ -0.173 0.496 7.27e-01 gradient ok
#> Classes 105/154[============= 68% === ] Estimated Time Left: 33s
OLS ~. mean Δ -0.173 0.496 7.46e-01 LR ≈Bartlett ok
#> Classes 106/154[============= 68% === ] Estimated Time Left: 32s
OLS ~. mean Δ -0.173 0.496 6.50e-01 param boot ok
#> Classes 107/154[============= 69% === ] Estimated Time Left: 31s
OLS ~. mean Δ -0.173 0.496 NA param boot ok
#> Classes 108/154[============= 70% === ] Estimated Time Left: 30s
OLS ~. mean Δ -0.173 0.496 6.63e-01 bayes boot (%… ok
#> Classes 109/154[============= 70% === ] Estimated Time Left: 29s
OLS ~. mean Δ -0.173 0.496 NA bayes boot (b… ok
#> Classes 110/154[============= 71% ==== ] Estimated Time Left: 29s
OLS ~. mean Δ -0.173 0.496 7.14e-01 bayes boot (w… ok
#> Classes 111/154[============= 72% ==== ] Estimated Time Left: 28s
OLS ~. mean Δ -0.173 0.496 7.50e-01 bayes boot (s… ok
#> Classes 112/154[============= 72% ==== ] Estimated Time Left: 27s
OLS ~. mean Δ -0.173 0.496 7.07e-01 bayes boot (t) ok
#> Classes 113/154[============= 73% ==== ] Estimated Time Left: 26s
OLS ~. mean Δ -0.173 0.496 6.86e-01 bayes boot (b… ok
#> Classes 114/154[============= 74% ==== ] Estimated Time Left: 25s
OLS ~. mean Δ -0.173 0.496 7.27e-01 bayes boot (s… ok
#> Classes 115/154[============= 74% ===== ] Estimated Time Left: 24s
OLS ~. mean Δ -0.173 0.496 7.31e-01 boot (%ile) ok
#> Classes 116/154[============= 75% ===== ] Estimated Time Left: 24s
OLS ~. mean Δ -0.173 0.496 NA boot (basic) ok
#> Classes 117/154[============= 75% ===== ] Estimated Time Left: 23s
OLS ~. mean Δ -0.173 0.496 7.66e-01 boot (stud) ok
#> Classes 118/154[============= 76% ===== ] Estimated Time Left: 24s
OLS ~. mean Δ -0.173 0.496 7.39e-01 boot (t) ok
#> Classes 119/154[============= 77% ===== ] Estimated Time Left: 25s
OLS ~. mean Δ -0.173 0.496 NA boot (symm t) ok
#> Classes 120/154[============= 77% ====== ] Estimated Time Left: 25s
OLS ~. mean Δ -0.173 0.496 6.98e-01 boot (bca) ok
#> Classes 121/154[============= 78% ====== ] Estimated Time Left: 24s
OLS ~. mean Δ -0.173 0.496 NA boot (prepiv) ok
#> Classes 122/154[============= 79% ====== ] Estimated Time Left: 23s
OLS ~. mean Δ -0.173 0.496 NA boot (dbl-boo… ok
#> Classes 123/154[============= 79% ====== ] Estimated Time Left: 22s
OLS ~. mean Δ -0.173 0.496 NA boot (calib) ok
#> Classes 124/154[============= 80% ====== ] Estimated Time Left: 21s
OLS ~. mean Δ -0.173 0.496 NA boot (smth) ok
#> Classes 125/154[============= 81% ======= ] Estimated Time Left: 20s
OLS ~. mean Δ -0.173 0.496 7.50e-01 boot (symm) ok
#> Classes 126/154[============= 81% ======= ] Estimated Time Left: 19s
Median Regr ~. median ef… 0.0102 0.680 9.88e-01 wald ok
#> Classes 127/154[============= 82% ======= ] Estimated Time Left: 19s
Median Regr ~. median ef… 0.0102 0.680 8.82e-01 rand ok
#> Classes 128/154[============= 83% ======= ] Estimated Time Left: 18s
Median Regr ~. median ef… 0.0102 0.680 8.26e-01 rand boot (%i… ok
#> Classes 129/154[============= 83% ======= ] Estimated Time Left: 17s
Median Regr ~. median ef… 0.0102 0.680 8.18e-01 rand boot (st… ok
#> Classes 130/154[============= 84% ======== ] Estimated Time Left: 17s
Median Regr ~. median ef… 0.0102 0.680 7.82e-01 rand boot (sy… ok
#> Classes 131/154[============= 85% ======== ] Estimated Time Left: 16s
Median Regr ~. median ef… 0.0102 0.680 9.78e-01 rand boot (sm… ok
#> Classes 132/154[============= 85% ======== ] Estimated Time Left: 15s
Median Regr ~. median ef… NA NA NA jackknife nonest
#> Classes 133/154[============= 86% ======== ] Estimated Time Left: 14s
Median Regr ~. median ef… 0.0102 0.680 8.06e-01 bayes boot (%… ok
#> Classes 134/154[============= 87% ======== ] Estimated Time Left: 14s
Median Regr ~. median ef… 0.0102 0.680 NA bayes boot (b… ok
#> Classes 135/154[============= 87% ========= ] Estimated Time Left: 13s
Median Regr ~. median ef… 0.0102 0.680 9.91e-01 bayes boot (w… ok
#> Classes 136/154[============= 88% ========= ] Estimated Time Left: 12s
Median Regr ~. median ef… 0.0102 0.680 9.68e-01 bayes boot (s… ok
#> Classes 137/154[============= 88% ========= ] Estimated Time Left: 12s
Median Regr ~. median ef… 0.0102 0.680 9.86e-01 bayes boot (t) ok
#> Classes 138/154[============= 89% ========= ] Estimated Time Left: 11s
Median Regr ~. median ef… NA NA NA bayes boot (b… nonest
#> Classes 139/154[============= 90% ========= ] Estimated Time Left: 10s
Median Regr ~. median ef… 0.0102 0.680 9.64e-01 bayes boot (s… ok
#> Classes 140/154[============= 90% ========== ] Estimated Time Left: 10s
Median Regr ~. median ef… 0.0102 0.680 7.94e-01 boot (%ile) ok
#> Classes 141/154[============= 91% =========== ] Estimated Time Left: 9s
Median Regr ~. median ef… 0.0102 0.680 NA boot (basic) ok
#> Classes 142/154[============= 92% =========== ] Estimated Time Left: 8s
Median Regr ~. median ef… 0.0102 0.680 9.66e-01 boot (stud) ok
#> Classes 143/154[============= 92% =========== ] Estimated Time Left: 8s
Median Regr ~. median ef… 0.0102 0.680 9.68e-01 boot (t) ok
#> Classes 144/154[============= 93% =========== ] Estimated Time Left: 8s
Median Regr ~. median ef… 0.0102 0.680 NA boot (symm t) ok
#> Classes 145/154[============= 94% ============ ] Estimated Time Left: 7s
Median Regr ~. median ef… 0.0102 0.680 7.95e-01 boot (bca) ok
#> Classes 146/154[============= 94% ============ ] Estimated Time Left: 6s
Median Regr ~. median ef… 0.0102 0.680 NA boot (prepiv) ok
#> Classes 147/154[============= 95% ============ ] Estimated Time Left: 5s
Median Regr ~. median ef… 0.0102 0.680 NA boot (dbl-boo… ok
#> Classes 148/154[============= 96% ============ ] Estimated Time Left: 5s
Median Regr ~. median ef… 0.0102 0.680 NA boot (calib) ok
#> Classes 149/154[============= 96% ============ ] Estimated Time Left: 4s
Median Regr ~. median ef… 0.0102 0.680 NA boot (smth) ok
#> Classes 150/154[============= 97% ============= ] Estimated Time Left: 3s
Median Regr ~. median ef… 0.0102 0.680 1.00e+00 boot (symm) ok
#> Classes 151/154[============= 98% ============= ] Estimated Time Left: 2s
Robust Regr ~. mean Δ -0.0504 0.476 9.17e-01 wald ok
#> Classes 152/154[============= 98% ============= ] Estimated Time Left: 2s
Robust Regr ~. mean Δ -0.0504 0.476 9.54e-01 rand ok
#> Classes 153/154[============= 99% ============= ] Estimated Time Left: 1s
Robust Regr ~. mean Δ -0.0504 0.476 9.49e-01 jackknife ok
#> Classes 154/154[============= 100% =============] Estimated Time Left: 0s
-------------------------------------------------------------------------------------------
#> Status: Completed in 1min 54s.
#>
#> Estimand: HL shift (10 inferences) : p = 0.950
#> Estimand: mean Δ (83 inferences) : p = 0.781
#> Estimand: median effect (16 inferences): p = 0.998
#>
#> Combined evidence against the sharp null across 3 estimands
#> (109 inferences, weighting = uniform within estimand):
#> p = 0.996
results$results_table
#> inference_class method
#> 1 InferenceAllSimpleWilcox wald
#> 2 InferenceAllSimpleWilcox rand
#> 3 InferenceAllSimpleWilcox rand_bootstrap
#> 4 InferenceAllSimpleWilcox rand_bootstrap
#> 5 InferenceAllSimpleWilcox rand_bootstrap
#> 6 InferenceAllSimpleWilcox bootstrap
#> 7 InferenceAllSimpleWilcox bootstrap
#> 8 InferenceAllSimpleWilcox bootstrap
#> 9 InferenceAllSimpleWilcox bootstrap
#> 10 InferenceAllSimpleWilcox bootstrap
#> 11 InferenceAllSimpleWilcox bootstrap
#> 12 InferenceAllSimpleWilcox bootstrap
#> 13 InferenceAllSimpleWilcox bootstrap
#> 14 InferenceAllSimpleWilcox bootstrap
#> 15 InferenceAllSimpleWilcox bootstrap
#> 16 InferenceAllSimpleWilcox bootstrap
#> 17 InferenceAllSimpleAverageDiff wald
#> 18 InferenceAllSimpleAverageDiff rand
#> 19 InferenceAllSimpleAverageDiff rand_bootstrap
#> 20 InferenceAllSimpleAverageDiff rand_bootstrap
#> 21 InferenceAllSimpleAverageDiff rand_bootstrap
#> 22 InferenceAllSimpleAverageDiff rand_bootstrap
#> 23 InferenceAllSimpleAverageDiff jackknife
#> 24 InferenceAllSimpleAverageDiff bayes_boot
#> 25 InferenceAllSimpleAverageDiff bayes_boot
#> 26 InferenceAllSimpleAverageDiff bayes_boot
#> 27 InferenceAllSimpleAverageDiff bayes_boot
#> 28 InferenceAllSimpleAverageDiff bayes_boot
#> 29 InferenceAllSimpleAverageDiff bayes_boot
#> 30 InferenceAllSimpleAverageDiff bayes_boot
#> 31 InferenceAllSimpleAverageDiff bootstrap
#> 32 InferenceAllSimpleAverageDiff bootstrap
#> 33 InferenceAllSimpleAverageDiff bootstrap
#> 34 InferenceAllSimpleAverageDiff bootstrap
#> 35 InferenceAllSimpleAverageDiff bootstrap
#> 36 InferenceAllSimpleAverageDiff bootstrap
#> 37 InferenceAllSimpleAverageDiff bootstrap
#> 38 InferenceAllSimpleAverageDiff bootstrap
#> 39 InferenceAllSimpleAverageDiff bootstrap
#> 40 InferenceAllSimpleMeanDiffPooledVar wald
#> 41 InferenceAllSimpleMeanDiffPooledVar rand
#> 42 InferenceAllSimpleMeanDiffPooledVar rand_bootstrap
#> 43 InferenceAllSimpleMeanDiffPooledVar rand_bootstrap
#> 44 InferenceAllSimpleMeanDiffPooledVar rand_bootstrap
#> 45 InferenceAllSimpleMeanDiffPooledVar rand_bootstrap
#> 46 InferenceAllSimpleMeanDiffPooledVar jackknife
#> 47 InferenceAllSimpleMeanDiffPooledVar bayes_boot
#> 48 InferenceAllSimpleMeanDiffPooledVar bayes_boot
#> 49 InferenceAllSimpleMeanDiffPooledVar bayes_boot
#> 50 InferenceAllSimpleMeanDiffPooledVar bayes_boot
#> 51 InferenceAllSimpleMeanDiffPooledVar bayes_boot
#> 52 InferenceAllSimpleMeanDiffPooledVar bayes_boot
#> 53 InferenceAllSimpleMeanDiffPooledVar bayes_boot
#> 54 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 55 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 56 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 57 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 58 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 59 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 60 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 61 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 62 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 63 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 64 InferenceAllSimpleMeanDiffPooledVar bootstrap
#> 65 InferenceContinLin wald
#> 66 InferenceContinLin rand
#> 67 InferenceContinLin rand_bootstrap
#> 68 InferenceContinLin rand_bootstrap
#> 69 InferenceContinLin rand_bootstrap
#> 70 InferenceContinLin rand_bootstrap
#> 71 InferenceContinLin jackknife
#> 72 InferenceContinLin score
#> 73 InferenceContinLin lik_ratio
#> 74 InferenceContinLin gradient
#> 75 InferenceContinLin lik_ratio_bartlett_approx
#> 76 InferenceContinLin param_boot
#> 77 InferenceContinLin param_boot_direct
#> 78 InferenceContinLin bayes_boot
#> 79 InferenceContinLin bayes_boot
#> 80 InferenceContinLin bayes_boot
#> 81 InferenceContinLin bayes_boot
#> 82 InferenceContinLin bayes_boot
#> 83 InferenceContinLin bayes_boot
#> 84 InferenceContinLin bayes_boot
#> 85 InferenceContinLin bootstrap
#> 86 InferenceContinLin bootstrap
#> 87 InferenceContinLin bootstrap
#> 88 InferenceContinLin bootstrap
#> 89 InferenceContinLin bootstrap
#> 90 InferenceContinLin bootstrap
#> 91 InferenceContinLin bootstrap
#> 92 InferenceContinLin bootstrap
#> 93 InferenceContinLin bootstrap
#> 94 InferenceContinLin bootstrap
#> 95 InferenceContinLin bootstrap
#> 96 InferenceContinOLS wald
#> 97 InferenceContinOLS rand
#> 98 InferenceContinOLS rand_bootstrap
#> 99 InferenceContinOLS rand_bootstrap
#> 100 InferenceContinOLS rand_bootstrap
#> 101 InferenceContinOLS rand_bootstrap
#> 102 InferenceContinOLS jackknife
#> 103 InferenceContinOLS score
#> 104 InferenceContinOLS lik_ratio
#> 105 InferenceContinOLS gradient
#> 106 InferenceContinOLS lik_ratio_bartlett_approx
#> 107 InferenceContinOLS param_boot
#> 108 InferenceContinOLS param_boot_direct
#> 109 InferenceContinOLS bayes_boot
#> 110 InferenceContinOLS bayes_boot
#> 111 InferenceContinOLS bayes_boot
#> 112 InferenceContinOLS bayes_boot
#> 113 InferenceContinOLS bayes_boot
#> 114 InferenceContinOLS bayes_boot
#> 115 InferenceContinOLS bayes_boot
#> 116 InferenceContinOLS bootstrap
#> 117 InferenceContinOLS bootstrap
#> 118 InferenceContinOLS bootstrap
#> 119 InferenceContinOLS bootstrap
#> 120 InferenceContinOLS bootstrap
#> 121 InferenceContinOLS bootstrap
#> 122 InferenceContinOLS bootstrap
#> 123 InferenceContinOLS bootstrap
#> 124 InferenceContinOLS bootstrap
#> 125 InferenceContinOLS bootstrap
#> 126 InferenceContinOLS bootstrap
#> 127 InferenceContinRobustRegr wald
#> 128 InferenceContinRobustRegr rand
#> 129 InferenceContinRobustRegr jackknife
#> 130 InferenceContinQuantileRegr wald
#> 131 InferenceContinQuantileRegr rand
#> 132 InferenceContinQuantileRegr rand_bootstrap
#> 133 InferenceContinQuantileRegr rand_bootstrap
#> 134 InferenceContinQuantileRegr rand_bootstrap
#> 135 InferenceContinQuantileRegr rand_bootstrap
#> 136 InferenceContinQuantileRegr jackknife
#> 137 InferenceContinQuantileRegr bayes_boot
#> 138 InferenceContinQuantileRegr bayes_boot
#> 139 InferenceContinQuantileRegr bayes_boot
#> 140 InferenceContinQuantileRegr bayes_boot
#> 141 InferenceContinQuantileRegr bayes_boot
#> 142 InferenceContinQuantileRegr bayes_boot
#> 143 InferenceContinQuantileRegr bayes_boot
#> 144 InferenceContinQuantileRegr bootstrap
#> 145 InferenceContinQuantileRegr bootstrap
#> 146 InferenceContinQuantileRegr bootstrap
#> 147 InferenceContinQuantileRegr bootstrap
#> 148 InferenceContinQuantileRegr bootstrap
#> 149 InferenceContinQuantileRegr bootstrap
#> 150 InferenceContinQuantileRegr bootstrap
#> 151 InferenceContinQuantileRegr bootstrap
#> 152 InferenceContinQuantileRegr bootstrap
#> 153 InferenceContinQuantileRegr bootstrap
#> 154 InferenceContinQuantileRegr bootstrap
#> type cov_model response_type design_family
#> 1 <NA> <NA> continuous iid
#> 2 <NA> <NA> continuous iid
#> 3 percentile <NA> continuous iid
#> 4 studentized <NA> continuous iid
#> 5 symmetric-percentile-t <NA> continuous iid
#> 6 percentile <NA> continuous iid
#> 7 basic <NA> continuous iid
#> 8 studentized <NA> continuous iid
#> 9 bootstrap-t <NA> continuous iid
#> 10 symmetric-percentile-t <NA> continuous iid
#> 11 bca <NA> continuous iid
#> 12 prepivoted <NA> continuous iid
#> 13 double-bootstrap <NA> continuous iid
#> 14 calibrated <NA> continuous iid
#> 15 smoothed <NA> continuous iid
#> 16 symmetric <NA> continuous iid
#> 17 <NA> <NA> continuous iid
#> 18 <NA> <NA> continuous iid
#> 19 percentile <NA> continuous iid
#> 20 studentized <NA> continuous iid
#> 21 symmetric-percentile-t <NA> continuous iid
#> 22 smoothed <NA> continuous iid
#> 23 <NA> <NA> continuous iid
#> 24 percentile <NA> continuous iid
#> 25 basic <NA> continuous iid
#> 26 wald <NA> continuous iid
#> 27 studentized <NA> continuous iid
#> 28 bootstrap-t <NA> continuous iid
#> 29 bca <NA> continuous iid
#> 30 symmetric <NA> continuous iid
#> 31 percentile <NA> continuous iid
#> 32 basic <NA> continuous iid
#> 33 symmetric-percentile-t <NA> continuous iid
#> 34 bca <NA> continuous iid
#> 35 prepivoted <NA> continuous iid
#> 36 double-bootstrap <NA> continuous iid
#> 37 calibrated <NA> continuous iid
#> 38 smoothed <NA> continuous iid
#> 39 symmetric <NA> continuous iid
#> 40 <NA> <NA> continuous iid
#> 41 <NA> <NA> continuous iid
#> 42 percentile <NA> continuous iid
#> 43 studentized <NA> continuous iid
#> 44 symmetric-percentile-t <NA> continuous iid
#> 45 smoothed <NA> continuous iid
#> 46 <NA> <NA> continuous iid
#> 47 percentile <NA> continuous iid
#> 48 basic <NA> continuous iid
#> 49 wald <NA> continuous iid
#> 50 studentized <NA> continuous iid
#> 51 bootstrap-t <NA> continuous iid
#> 52 bca <NA> continuous iid
#> 53 symmetric <NA> continuous iid
#> 54 percentile <NA> continuous iid
#> 55 basic <NA> continuous iid
#> 56 studentized <NA> continuous iid
#> 57 bootstrap-t <NA> continuous iid
#> 58 symmetric-percentile-t <NA> continuous iid
#> 59 bca <NA> continuous iid
#> 60 prepivoted <NA> continuous iid
#> 61 double-bootstrap <NA> continuous iid
#> 62 calibrated <NA> continuous iid
#> 63 smoothed <NA> continuous iid
#> 64 symmetric <NA> continuous iid
#> 65 <NA> ~. continuous iid
#> 66 <NA> ~. continuous iid
#> 67 percentile ~. continuous iid
#> 68 studentized ~. continuous iid
#> 69 symmetric-percentile-t ~. continuous iid
#> 70 smoothed ~. continuous iid
#> 71 <NA> ~. continuous iid
#> 72 <NA> ~. continuous iid
#> 73 <NA> ~. continuous iid
#> 74 <NA> ~. continuous iid
#> 75 <NA> ~. continuous iid
#> 76 <NA> ~. continuous iid
#> 77 <NA> ~. continuous iid
#> 78 percentile ~. continuous iid
#> 79 basic ~. continuous iid
#> 80 wald ~. continuous iid
#> 81 studentized ~. continuous iid
#> 82 bootstrap-t ~. continuous iid
#> 83 bca ~. continuous iid
#> 84 symmetric ~. continuous iid
#> 85 percentile ~. continuous iid
#> 86 basic ~. continuous iid
#> 87 studentized ~. continuous iid
#> 88 bootstrap-t ~. continuous iid
#> 89 symmetric-percentile-t ~. continuous iid
#> 90 bca ~. continuous iid
#> 91 prepivoted ~. continuous iid
#> 92 double-bootstrap ~. continuous iid
#> 93 calibrated ~. continuous iid
#> 94 smoothed ~. continuous iid
#> 95 symmetric ~. continuous iid
#> 96 <NA> ~. continuous iid
#> 97 <NA> ~. continuous iid
#> 98 percentile ~. continuous iid
#> 99 studentized ~. continuous iid
#> 100 symmetric-percentile-t ~. continuous iid
#> 101 smoothed ~. continuous iid
#> 102 <NA> ~. continuous iid
#> 103 <NA> ~. continuous iid
#> 104 <NA> ~. continuous iid
#> 105 <NA> ~. continuous iid
#> 106 <NA> ~. continuous iid
#> 107 <NA> ~. continuous iid
#> 108 <NA> ~. continuous iid
#> 109 percentile ~. continuous iid
#> 110 basic ~. continuous iid
#> 111 wald ~. continuous iid
#> 112 studentized ~. continuous iid
#> 113 bootstrap-t ~. continuous iid
#> 114 bca ~. continuous iid
#> 115 symmetric ~. continuous iid
#> 116 percentile ~. continuous iid
#> 117 basic ~. continuous iid
#> 118 studentized ~. continuous iid
#> 119 bootstrap-t ~. continuous iid
#> 120 symmetric-percentile-t ~. continuous iid
#> 121 bca ~. continuous iid
#> 122 prepivoted ~. continuous iid
#> 123 double-bootstrap ~. continuous iid
#> 124 calibrated ~. continuous iid
#> 125 smoothed ~. continuous iid
#> 126 symmetric ~. continuous iid
#> 127 <NA> ~. continuous iid
#> 128 <NA> ~. continuous iid
#> 129 <NA> ~. continuous iid
#> 130 <NA> ~. continuous iid
#> 131 <NA> ~. continuous iid
#> 132 percentile ~. continuous iid
#> 133 studentized ~. continuous iid
#> 134 symmetric-percentile-t ~. continuous iid
#> 135 smoothed ~. continuous iid
#> 136 <NA> ~. continuous iid
#> 137 percentile ~. continuous iid
#> 138 basic ~. continuous iid
#> 139 wald ~. continuous iid
#> 140 studentized ~. continuous iid
#> 141 bootstrap-t ~. continuous iid
#> 142 bca ~. continuous iid
#> 143 symmetric ~. continuous iid
#> 144 percentile ~. continuous iid
#> 145 basic ~. continuous iid
#> 146 studentized ~. continuous iid
#> 147 bootstrap-t ~. continuous iid
#> 148 symmetric-percentile-t ~. continuous iid
#> 149 bca ~. continuous iid
#> 150 prepivoted ~. continuous iid
#> 151 double-bootstrap ~. continuous iid
#> 152 calibrated ~. continuous iid
#> 153 smoothed ~. continuous iid
#> 154 symmetric ~. continuous iid
#> likelihood_tier estimate se ci_a ci_b ci_method
#> 1 none -0.03783456 0.6362701 NA NA wald
#> 2 none -0.03783456 0.6362701 NA NA rand
#> 3 none -0.03783456 0.6362701 NA NA rand_bootstrap
#> 4 none -0.03783456 0.6362701 NA NA rand_bootstrap
#> 5 none -0.03783456 0.6362701 NA NA rand_bootstrap
#> 6 none -0.03783456 0.6362701 NA NA bootstrap
#> 7 none -0.03783456 0.6362701 NA NA bootstrap
#> 8 none -0.03783456 0.6362701 NA NA bootstrap
#> 9 none -0.03783456 0.6362701 NA NA bootstrap
#> 10 none -0.03783456 0.6362701 NA NA bootstrap
#> 11 none -0.03783456 0.6362701 NA NA bootstrap
#> 12 none -0.03783456 0.6362701 NA NA bootstrap
#> 13 none -0.03783456 0.6362701 NA NA bootstrap
#> 14 none -0.03783456 0.6362701 NA NA bootstrap
#> 15 none -0.03783456 0.6362701 NA NA bootstrap
#> 16 none -0.03783456 0.6362701 NA NA bootstrap
#> 17 none -0.15250605 0.5687377 NA NA wald
#> 18 none -0.15250605 0.5687377 NA NA rand
#> 19 none -0.15250605 0.5687377 NA NA rand_bootstrap
#> 20 none -0.15250605 0.5687377 NA NA rand_bootstrap
#> 21 none -0.15250605 0.5687377 NA NA rand_bootstrap
#> 22 none -0.15250605 0.5687377 NA NA rand_bootstrap
#> 23 none -0.15250605 0.5687377 NA NA jackknife
#> 24 none -0.15250605 0.5687377 NA NA bayes_boot
#> 25 none -0.15250605 0.5687377 NA NA bayes_boot
#> 26 none -0.15250605 0.5687377 NA NA bayes_boot
#> 27 none -0.15250605 0.5687377 NA NA bayes_boot
#> 28 none -0.15250605 0.5687377 NA NA bayes_boot
#> 29 none -0.15250605 0.5687377 NA NA bayes_boot
#> 30 none -0.15250605 0.5687377 NA NA bayes_boot
#> 31 none -0.15250605 0.5687377 NA NA bootstrap
#> 32 none -0.15250605 0.5687377 NA NA bootstrap
#> 33 none -0.15250605 0.5687377 NA NA bootstrap
#> 34 none -0.15250605 0.5687377 NA NA bootstrap
#> 35 none -0.15250605 0.5687377 NA NA bootstrap
#> 36 none -0.15250605 0.5687377 NA NA bootstrap
#> 37 none -0.15250605 0.5687377 NA NA bootstrap
#> 38 none -0.15250605 0.5687377 NA NA bootstrap
#> 39 none -0.15250605 0.5687377 NA NA bootstrap
#> 40 none -0.15250605 0.4830524 NA NA wald
#> 41 none -0.15250605 0.4830524 NA NA rand
#> 42 none -0.15250605 0.4830524 NA NA rand_bootstrap
#> 43 none -0.15250605 0.4830524 NA NA rand_bootstrap
#> 44 none -0.15250605 0.4830524 NA NA rand_bootstrap
#> 45 none -0.15250605 0.4830524 NA NA rand_bootstrap
#> 46 none -0.15250605 0.4830524 NA NA jackknife
#> 47 none -0.15250605 0.4830524 NA NA bayes_boot
#> 48 none -0.15250605 0.4830524 NA NA bayes_boot
#> 49 none -0.15250605 0.4830524 NA NA bayes_boot
#> 50 none -0.15250605 0.4830524 NA NA bayes_boot
#> 51 none -0.15250605 0.4830524 NA NA bayes_boot
#> 52 none -0.15250605 0.4830524 NA NA bayes_boot
#> 53 none -0.15250605 0.4830524 NA NA bayes_boot
#> 54 none -0.15250605 0.4830524 NA NA bootstrap
#> 55 none -0.15250605 0.4830524 NA NA bootstrap
#> 56 none -0.15250605 0.4830524 NA NA bootstrap
#> 57 none -0.15250605 0.4830524 NA NA bootstrap
#> 58 none -0.15250605 0.4830524 NA NA bootstrap
#> 59 none -0.15250605 0.4830524 NA NA bootstrap
#> 60 none -0.15250605 0.4830524 NA NA bootstrap
#> 61 none -0.15250605 0.4830524 NA NA bootstrap
#> 62 none -0.15250605 0.4830524 NA NA bootstrap
#> 63 none -0.15250605 0.4830524 NA NA bootstrap
#> 64 none -0.15250605 0.4830524 NA NA bootstrap
#> 65 full -0.17443153 0.6025512 NA NA wald
#> 66 full -0.17443153 0.6025512 NA NA rand
#> 67 full -0.17443153 0.6025512 NA NA rand_bootstrap
#> 68 full -0.17443153 0.6025512 NA NA rand_bootstrap
#> 69 full -0.17443153 0.6025512 NA NA rand_bootstrap
#> 70 full -0.17443153 0.6025512 NA NA rand_bootstrap
#> 71 full -0.17443153 0.6025512 NA NA jackknife
#> 72 full -0.17443153 0.6025512 NA NA score
#> 73 full -0.17443153 0.6025512 NA NA lik_ratio
#> 74 full -0.17443153 0.6025512 NA NA gradient
#> 75 full -0.17443153 0.6025512 NA NA lik_ratio_bartlett_approx
#> 76 full -0.17443153 0.6025512 NA NA param_boot
#> 77 full -0.17443153 0.6025512 NA NA param_boot_direct
#> 78 full -0.17443153 0.6025512 NA NA bayes_boot
#> 79 full -0.17443153 0.6025512 NA NA bayes_boot
#> 80 full -0.17443153 0.6025512 NA NA bayes_boot
#> 81 full -0.17443153 0.6025512 NA NA bayes_boot
#> 82 full -0.17443153 0.6025512 NA NA bayes_boot
#> 83 full -0.17443153 0.6025512 NA NA bayes_boot
#> 84 full -0.17443153 0.6025512 NA NA bayes_boot
#> 85 full -0.17443153 0.6025512 NA NA bootstrap
#> 86 full -0.17443153 0.6025512 NA NA bootstrap
#> 87 full -0.17443153 0.6025512 NA NA bootstrap
#> 88 full -0.17443153 0.6025512 NA NA bootstrap
#> 89 full -0.17443153 0.6025512 NA NA bootstrap
#> 90 full -0.17443153 0.6025512 NA NA bootstrap
#> 91 full -0.17443153 0.6025512 NA NA bootstrap
#> 92 full -0.17443153 0.6025512 NA NA bootstrap
#> 93 full -0.17443153 0.6025512 NA NA bootstrap
#> 94 full -0.17443153 0.6025512 NA NA bootstrap
#> 95 full -0.17443153 0.6025512 NA NA bootstrap
#> 96 full -0.17290752 0.4955827 NA NA wald
#> 97 full -0.17290752 0.4955827 NA NA rand
#> 98 full -0.17290752 0.4955827 NA NA rand_bootstrap
#> 99 full -0.17290752 0.4955827 NA NA rand_bootstrap
#> 100 full -0.17290752 0.4955827 NA NA rand_bootstrap
#> 101 full -0.17290752 0.4955827 NA NA rand_bootstrap
#> 102 full -0.17290752 0.4955827 NA NA jackknife
#> 103 full -0.17290752 0.4955827 NA NA score
#> 104 full -0.17290752 0.4955827 NA NA lik_ratio
#> 105 full -0.17290752 0.4955827 NA NA gradient
#> 106 full -0.17290752 0.4955827 NA NA lik_ratio_bartlett_approx
#> 107 full -0.17290752 0.4955827 NA NA param_boot
#> 108 full -0.17290752 0.4955827 NA NA param_boot_direct
#> 109 full -0.17290752 0.4955827 NA NA bayes_boot
#> 110 full -0.17290752 0.4955827 NA NA bayes_boot
#> 111 full -0.17290752 0.4955827 NA NA bayes_boot
#> 112 full -0.17290752 0.4955827 NA NA bayes_boot
#> 113 full -0.17290752 0.4955827 NA NA bayes_boot
#> 114 full -0.17290752 0.4955827 NA NA bayes_boot
#> 115 full -0.17290752 0.4955827 NA NA bayes_boot
#> 116 full -0.17290752 0.4955827 NA NA bootstrap
#> 117 full -0.17290752 0.4955827 NA NA bootstrap
#> 118 full -0.17290752 0.4955827 NA NA bootstrap
#> 119 full -0.17290752 0.4955827 NA NA bootstrap
#> 120 full -0.17290752 0.4955827 NA NA bootstrap
#> 121 full -0.17290752 0.4955827 NA NA bootstrap
#> 122 full -0.17290752 0.4955827 NA NA bootstrap
#> 123 full -0.17290752 0.4955827 NA NA bootstrap
#> 124 full -0.17290752 0.4955827 NA NA bootstrap
#> 125 full -0.17290752 0.4955827 NA NA bootstrap
#> 126 full -0.17290752 0.4955827 NA NA bootstrap
#> 127 quasi -0.05038058 0.4762767 NA NA wald
#> 128 quasi -0.05038058 0.4762767 NA NA rand
#> 129 quasi -0.05038058 0.4762767 NA NA jackknife
#> 130 none 0.01022966 0.6799695 NA NA wald
#> 131 none 0.01022966 0.6799695 NA NA rand
#> 132 none 0.01022966 0.6799695 NA NA rand_bootstrap
#> 133 none 0.01022966 0.6799695 NA NA rand_bootstrap
#> 134 none 0.01022966 0.6799695 NA NA rand_bootstrap
#> 135 none 0.01022966 0.6799695 NA NA rand_bootstrap
#> 136 none NA NA NA NA jackknife
#> 137 none 0.01022966 0.6799695 NA NA bayes_boot
#> 138 none 0.01022966 0.6799695 NA NA bayes_boot
#> 139 none 0.01022966 0.6799695 NA NA bayes_boot
#> 140 none 0.01022966 0.6799695 NA NA bayes_boot
#> 141 none 0.01022966 0.6799695 NA NA bayes_boot
#> 142 none NA NA NA NA bayes_boot
#> 143 none 0.01022966 0.6799695 NA NA bayes_boot
#> 144 none 0.01022966 0.6799695 NA NA bootstrap
#> 145 none 0.01022966 0.6799695 NA NA bootstrap
#> 146 none 0.01022966 0.6799695 NA NA bootstrap
#> 147 none 0.01022966 0.6799695 NA NA bootstrap
#> 148 none 0.01022966 0.6799695 NA NA bootstrap
#> 149 none 0.01022966 0.6799695 NA NA bootstrap
#> 150 none 0.01022966 0.6799695 NA NA bootstrap
#> 151 none 0.01022966 0.6799695 NA NA bootstrap
#> 152 none 0.01022966 0.6799695 NA NA bootstrap
#> 153 none 0.01022966 0.6799695 NA NA bootstrap
#> 154 none 0.01022966 0.6799695 NA NA bootstrap
#> pval pval_method estimand tau
#> 1 0.9692286 wald hodges_lehmann_shift NA
#> 2 0.9261477 rand hodges_lehmann_shift NA
#> 3 0.7584830 rand_bootstrap hodges_lehmann_shift NA
#> 4 0.9061876 rand_bootstrap hodges_lehmann_shift NA
#> 5 0.7984032 rand_bootstrap hodges_lehmann_shift NA
#> 6 0.9021956 bootstrap hodges_lehmann_shift NA
#> 7 NA bootstrap hodges_lehmann_shift NA
#> 8 0.9640719 bootstrap hodges_lehmann_shift NA
#> 9 0.9540918 bootstrap hodges_lehmann_shift NA
#> 10 NA bootstrap hodges_lehmann_shift NA
#> 11 0.9739557 bootstrap hodges_lehmann_shift NA
#> 12 NA bootstrap hodges_lehmann_shift NA
#> 13 NA bootstrap hodges_lehmann_shift NA
#> 14 NA bootstrap hodges_lehmann_shift NA
#> 15 NA bootstrap hodges_lehmann_shift NA
#> 16 0.9720559 bootstrap hodges_lehmann_shift NA
#> 17 0.7951406 wald mean_difference NA
#> 18 0.7425150 rand mean_difference NA
#> 19 0.7025948 rand_bootstrap mean_difference NA
#> 20 0.2810811 rand_bootstrap mean_difference NA
#> 21 0.9482289 rand_bootstrap mean_difference NA
#> 22 0.7544910 rand_bootstrap mean_difference NA
#> 23 0.7965349 jackknife mean_difference NA
#> 24 0.7784431 bayes_boot mean_difference NA
#> 25 NA bayes_boot mean_difference NA
#> 26 0.7667280 bayes_boot mean_difference NA
#> 27 0.7345309 bayes_boot mean_difference NA
#> 28 0.6786427 bayes_boot mean_difference NA
#> 29 0.7748984 bayes_boot mean_difference NA
#> 30 0.7624750 bayes_boot mean_difference NA
#> 31 0.7465070 bootstrap mean_difference NA
#> 32 NA bootstrap mean_difference NA
#> 33 NA bootstrap mean_difference NA
#> 34 0.7222896 bootstrap mean_difference NA
#> 35 NA bootstrap mean_difference NA
#> 36 NA bootstrap mean_difference NA
#> 37 NA bootstrap mean_difference NA
#> 38 NA bootstrap mean_difference NA
#> 39 0.8083832 bootstrap mean_difference NA
#> 40 0.7558514 wald mean_difference NA
#> 41 0.7425150 rand mean_difference NA
#> 42 0.6906188 rand_bootstrap mean_difference NA
#> 43 0.2339833 rand_bootstrap mean_difference NA
#> 44 0.9438503 rand_bootstrap mean_difference NA
#> 45 0.7265469 rand_bootstrap mean_difference NA
#> 46 0.7965349 jackknife mean_difference NA
#> 47 0.7305389 bayes_boot mean_difference NA
#> 48 NA bayes_boot mean_difference NA
#> 49 0.7609028 bayes_boot mean_difference NA
#> 50 0.7025948 bayes_boot mean_difference NA
#> 51 0.6806387 bayes_boot mean_difference NA
#> 52 0.8056509 bayes_boot mean_difference NA
#> 53 0.7365269 bayes_boot mean_difference NA
#> 54 0.7385230 bootstrap mean_difference NA
#> 55 NA bootstrap mean_difference NA
#> 56 0.7900000 bootstrap mean_difference NA
#> 57 0.8103792 bootstrap mean_difference NA
#> 58 NA bootstrap mean_difference NA
#> 59 0.7938770 bootstrap mean_difference NA
#> 60 NA bootstrap mean_difference NA
#> 61 NA bootstrap mean_difference NA
#> 62 NA bootstrap mean_difference NA
#> 63 NA bootstrap mean_difference NA
#> 64 0.8023952 bootstrap mean_difference NA
#> 65 0.7759254 wald mean_difference NA
#> 66 0.7265469 rand mean_difference NA
#> 67 0.7105788 rand_bootstrap mean_difference NA
#> 68 0.7704591 rand_bootstrap mean_difference NA
#> 69 0.7644711 rand_bootstrap mean_difference NA
#> 70 0.7624750 rand_bootstrap mean_difference NA
#> 71 0.7847208 jackknife mean_difference NA
#> 72 0.7327307 score mean_difference NA
#> 73 0.7327307 lik_ratio mean_difference NA
#> 74 0.7327307 gradient mean_difference NA
#> 75 0.7068113 lik_ratio_bartlett_approx mean_difference NA
#> 76 0.7200000 param_boot mean_difference NA
#> 77 NA param_boot_direct mean_difference NA
#> 78 0.6626747 bayes_boot mean_difference NA
#> 79 NA bayes_boot mean_difference NA
#> 80 0.7239869 bayes_boot mean_difference NA
#> 81 0.7984032 bayes_boot mean_difference NA
#> 82 0.7564870 bayes_boot mean_difference NA
#> 83 0.6810912 bayes_boot mean_difference NA
#> 84 0.7205589 bayes_boot mean_difference NA
#> 85 0.7465070 bootstrap mean_difference NA
#> 86 NA bootstrap mean_difference NA
#> 87 0.7904192 bootstrap mean_difference NA
#> 88 0.7604790 bootstrap mean_difference NA
#> 89 NA bootstrap mean_difference NA
#> 90 0.7988539 bootstrap mean_difference NA
#> 91 NA bootstrap mean_difference NA
#> 92 NA bootstrap mean_difference NA
#> 93 NA bootstrap mean_difference NA
#> 94 NA bootstrap mean_difference NA
#> 95 0.7684631 bootstrap mean_difference NA
#> 96 0.7314502 wald mean_difference NA
#> 97 0.7225549 rand mean_difference NA
#> 98 0.7265469 rand_bootstrap mean_difference NA
#> 99 0.7504990 rand_bootstrap mean_difference NA
#> 100 0.7345309 rand_bootstrap mean_difference NA
#> 101 0.7105788 rand_bootstrap mean_difference NA
#> 102 0.7796570 jackknife mean_difference NA
#> 103 0.7271663 score mean_difference NA
#> 104 0.7271663 lik_ratio mean_difference NA
#> 105 0.7271663 gradient mean_difference NA
#> 106 0.7456134 lik_ratio_bartlett_approx mean_difference NA
#> 107 0.6500000 param_boot mean_difference NA
#> 108 NA param_boot_direct mean_difference NA
#> 109 0.6626747 bayes_boot mean_difference NA
#> 110 NA bayes_boot mean_difference NA
#> 111 0.7140896 bayes_boot mean_difference NA
#> 112 0.7504990 bayes_boot mean_difference NA
#> 113 0.7065868 bayes_boot mean_difference NA
#> 114 0.6856427 bayes_boot mean_difference NA
#> 115 0.7265469 bayes_boot mean_difference NA
#> 116 0.7305389 bootstrap mean_difference NA
#> 117 NA bootstrap mean_difference NA
#> 118 0.7664671 bootstrap mean_difference NA
#> 119 0.7385230 bootstrap mean_difference NA
#> 120 NA bootstrap mean_difference NA
#> 121 0.6981422 bootstrap mean_difference NA
#> 122 NA bootstrap mean_difference NA
#> 123 NA bootstrap mean_difference NA
#> 124 NA bootstrap mean_difference NA
#> 125 NA bootstrap mean_difference NA
#> 126 0.7504990 bootstrap mean_difference NA
#> 127 0.9169950 wald mean_difference NA
#> 128 0.9540918 rand mean_difference NA
#> 129 0.9494190 jackknife mean_difference NA
#> 130 0.9881720 wald quantile_regression_effect 0.5
#> 131 0.8822355 rand quantile_regression_effect 0.5
#> 132 0.8263473 rand_bootstrap quantile_regression_effect 0.5
#> 133 0.8183633 rand_bootstrap quantile_regression_effect 0.5
#> 134 0.7824351 rand_bootstrap quantile_regression_effect 0.5
#> 135 0.9780439 rand_bootstrap quantile_regression_effect 0.5
#> 136 NA jackknife quantile_regression_effect 0.5
#> 137 0.8063872 bayes_boot quantile_regression_effect 0.5
#> 138 NA bayes_boot quantile_regression_effect 0.5
#> 139 0.9911529 bayes_boot quantile_regression_effect 0.5
#> 140 0.9680639 bayes_boot quantile_regression_effect 0.5
#> 141 0.9860279 bayes_boot quantile_regression_effect 0.5
#> 142 NA bayes_boot quantile_regression_effect 0.5
#> 143 0.9640719 bayes_boot quantile_regression_effect 0.5
#> 144 0.7944112 bootstrap quantile_regression_effect 0.5
#> 145 NA bootstrap quantile_regression_effect 0.5
#> 146 0.9660679 bootstrap quantile_regression_effect 0.5
#> 147 0.9680639 bootstrap quantile_regression_effect 0.5
#> 148 NA bootstrap quantile_regression_effect 0.5
#> 149 0.7947846 bootstrap quantile_regression_effect 0.5
#> 150 NA bootstrap quantile_regression_effect 0.5
#> 151 NA bootstrap quantile_regression_effect 0.5
#> 152 NA bootstrap quantile_regression_effect 0.5
#> 153 NA bootstrap quantile_regression_effect 0.5
#> 154 1.0000000 bootstrap quantile_regression_effect 0.5
#> fit_secs warnings status
#> 1 0.059902191 <NA> ok
#> 2 0.118165731 <NA> ok
#> 3 0.138579130 <NA> ok
#> 4 5.188518286 <NA> ok
#> 5 5.200045347 <NA> ok
#> 6 0.210901022 <NA> ok
#> 7 0.012223244 <NA> ok
#> 8 11.962827682 <NA> ok
#> 9 11.371673584 <NA> ok
#> 10 0.011763811 <NA> ok
#> 11 0.182598591 <NA> ok
#> 12 0.012079954 <NA> ok
#> 13 0.011754513 <NA> ok
#> 14 0.011879444 <NA> ok
#> 15 0.012142420 <NA> ok
#> 16 0.177815914 <NA> ok
#> 17 0.020965338 <NA> ok
#> 18 0.019849300 <NA> ok
#> 19 0.134555578 <NA> ok
#> 20 0.129106283 <NA> ok
#> 21 0.135946035 <NA> ok
#> 22 0.247006893 <NA> ok
#> 23 0.030326605 <NA> ok
#> 24 0.146484137 <NA> ok
#> 25 0.003058910 <NA> ok
#> 26 0.091918468 <NA> ok
#> 27 0.083066940 <NA> ok
#> 28 0.082567215 <NA> ok
#> 29 0.137579679 <NA> ok
#> 30 0.089218378 <NA> ok
#> 31 0.151295900 <NA> ok
#> 32 0.004204750 <NA> ok
#> 33 0.003700972 <NA> ok
#> 34 0.184689045 <NA> ok
#> 35 0.004196882 <NA> ok
#> 36 0.003724337 <NA> ok
#> 37 0.003794193 <NA> ok
#> 38 0.003998995 <NA> ok
#> 39 0.195251226 <NA> ok
#> 40 0.021564960 <NA> ok
#> 41 0.003508091 <NA> ok
#> 42 0.124040842 <NA> ok
#> 43 0.125257254 <NA> ok
#> 44 0.126347542 <NA> ok
#> 45 0.210476160 <NA> ok
#> 46 0.057380676 <NA> ok
#> 47 0.090238571 <NA> ok
#> 48 0.003121853 <NA> ok
#> 49 0.090002298 <NA> ok
#> 50 0.081347942 <NA> ok
#> 51 0.079780579 <NA> ok
#> 52 0.109174013 <NA> ok
#> 53 0.101216078 <NA> ok
#> 54 0.177995443 <NA> ok
#> 55 0.004067421 <NA> ok
#> 56 5.994385004 <NA> ok
#> 57 6.711049080 <NA> ok
#> 58 0.004271269 <NA> ok
#> 59 0.193300009 <NA> ok
#> 60 0.004453182 <NA> ok
#> 61 0.004233122 <NA> ok
#> 62 0.004309177 <NA> ok
#> 63 0.004039049 <NA> ok
#> 64 0.239356756 <NA> ok
#> 65 0.038105965 <NA> ok
#> 66 0.156400442 <NA> ok
#> 67 0.450530291 <NA> ok
#> 68 0.455615520 <NA> ok
#> 69 0.470707655 <NA> ok
#> 70 0.324370146 <NA> ok
#> 71 0.039142847 <NA> ok
#> 72 0.003986597 <NA> ok
#> 73 0.003698349 <NA> ok
#> 74 0.003715992 <NA> ok
#> 75 0.063429594 <NA> ok
#> 76 0.169975519 <NA> ok
#> 77 0.004399776 <NA> ok
#> 78 0.172269106 <NA> ok
#> 79 0.003543854 <NA> ok
#> 80 0.145790577 <NA> ok
#> 81 0.182634115 <NA> ok
#> 82 0.148845911 <NA> ok
#> 83 0.160722971 <NA> ok
#> 84 0.155814409 <NA> ok
#> 85 0.274686813 <NA> ok
#> 86 0.004955530 <NA> ok
#> 87 6.697855711 <NA> ok
#> 88 7.194753408 <NA> ok
#> 89 0.005559921 <NA> ok
#> 90 0.310842752 <NA> ok
#> 91 0.004763365 <NA> ok
#> 92 0.004461288 <NA> ok
#> 93 0.004386425 <NA> ok
#> 94 0.004328489 <NA> ok
#> 95 0.257985353 <NA> ok
#> 96 0.003168106 <NA> ok
#> 97 0.204759121 <NA> ok
#> 98 0.181384087 <NA> ok
#> 99 0.527695656 <NA> ok
#> 100 0.469374895 <NA> ok
#> 101 0.317084789 <NA> ok
#> 102 0.037158012 <NA> ok
#> 103 0.003544569 <NA> ok
#> 104 0.003224373 <NA> ok
#> 105 0.003051281 <NA> ok
#> 106 0.077447891 <NA> ok
#> 107 0.235266685 <NA> ok
#> 108 0.006395817 <NA> ok
#> 109 0.164518833 <NA> ok
#> 110 0.004008770 <NA> ok
#> 111 0.133242607 <NA> ok
#> 112 0.139883995 <NA> ok
#> 113 0.141478539 <NA> ok
#> 114 0.144209862 <NA> ok
#> 115 0.128180265 <NA> ok
#> 116 0.270155430 <NA> ok
#> 117 0.004810572 <NA> ok
#> 118 7.219660282 <NA> ok
#> 119 7.045979977 <NA> ok
#> 120 0.007349491 <NA> ok
#> 121 0.375242472 <NA> ok
#> 122 0.005179644 <NA> ok
#> 123 0.005267620 <NA> ok
#> 124 0.005106926 <NA> ok
#> 125 0.004756451 <NA> ok
#> 126 0.319339275 <NA> ok
#> 127 0.045727730 <NA> ok
#> 128 0.153317928 <NA> ok
#> 129 0.038629293 <NA> ok
#> 130 0.035399914 <NA> ok
#> 131 0.268756390 <NA> ok
#> 132 0.629224539 <NA> ok
#> 133 1.585719347 <NA> ok
#> 134 1.325687647 <NA> ok
#> 135 0.569822550 <NA> ok
#> 136 0.062953711 <NA> nonest
#> 137 0.626703501 <NA> ok
#> 138 0.004979372 <NA> ok
#> 139 0.675166368 <NA> ok
#> 140 1.011821985 <NA> ok
#> 141 1.033550739 <NA> ok
#> 142 0.666625261 <NA> nonest
#> 143 0.648057699 <NA> ok
#> 144 0.532153845 <NA> ok
#> 145 0.008710623 <NA> ok
#> 146 8.097377539 <NA> ok
#> 147 8.038814068 <NA> ok
#> 148 0.008051872 <NA> ok
#> 149 0.518912792 <NA> ok
#> 150 0.006154060 <NA> ok
#> 151 0.005468845 <NA> ok
#> 152 0.005347729 <NA> ok
#> 153 0.005474329 <NA> ok
#> 154 0.387069464 <NA> ok
#> message
#> 1 <NA>
#> 2 <NA>
#> 3 <NA>
#> 4 <NA>
#> 5 <NA>
#> 6 <NA>
#> 7 <NA>
#> 8 <NA>
#> 9 <NA>
#> 10 <NA>
#> 11 <NA>
#> 12 <NA>
#> 13 <NA>
#> 14 <NA>
#> 15 <NA>
#> 16 <NA>
#> 17 <NA>
#> 18 <NA>
#> 19 <NA>
#> 20 <NA>
#> 21 <NA>
#> 22 <NA>
#> 23 <NA>
#> 24 <NA>
#> 25 <NA>
#> 26 <NA>
#> 27 <NA>
#> 28 <NA>
#> 29 <NA>
#> 30 <NA>
#> 31 <NA>
#> 32 <NA>
#> 33 <NA>
#> 34 <NA>
#> 35 <NA>
#> 36 <NA>
#> 37 <NA>
#> 38 <NA>
#> 39 <NA>
#> 40 <NA>
#> 41 <NA>
#> 42 <NA>
#> 43 <NA>
#> 44 <NA>
#> 45 <NA>
#> 46 <NA>
#> 47 <NA>
#> 48 <NA>
#> 49 <NA>
#> 50 <NA>
#> 51 <NA>
#> 52 <NA>
#> 53 <NA>
#> 54 <NA>
#> 55 <NA>
#> 56 <NA>
#> 57 <NA>
#> 58 <NA>
#> 59 <NA>
#> 60 <NA>
#> 61 <NA>
#> 62 <NA>
#> 63 <NA>
#> 64 <NA>
#> 65 <NA>
#> 66 <NA>
#> 67 <NA>
#> 68 <NA>
#> 69 <NA>
#> 70 <NA>
#> 71 <NA>
#> 72 <NA>
#> 73 <NA>
#> 74 <NA>
#> 75 <NA>
#> 76 <NA>
#> 77 p-value (param_boot_direct) failed: argument "delta" is missing, with no default
#> 78 <NA>
#> 79 <NA>
#> 80 <NA>
#> 81 <NA>
#> 82 <NA>
#> 83 <NA>
#> 84 <NA>
#> 85 <NA>
#> 86 <NA>
#> 87 <NA>
#> 88 <NA>
#> 89 <NA>
#> 90 <NA>
#> 91 <NA>
#> 92 <NA>
#> 93 <NA>
#> 94 <NA>
#> 95 <NA>
#> 96 <NA>
#> 97 <NA>
#> 98 <NA>
#> 99 <NA>
#> 100 <NA>
#> 101 <NA>
#> 102 <NA>
#> 103 <NA>
#> 104 <NA>
#> 105 <NA>
#> 106 <NA>
#> 107 <NA>
#> 108 p-value (param_boot_direct) failed: argument "delta" is missing, with no default
#> 109 <NA>
#> 110 <NA>
#> 111 <NA>
#> 112 <NA>
#> 113 <NA>
#> 114 <NA>
#> 115 <NA>
#> 116 <NA>
#> 117 <NA>
#> 118 <NA>
#> 119 <NA>
#> 120 <NA>
#> 121 <NA>
#> 122 <NA>
#> 123 <NA>
#> 124 <NA>
#> 125 <NA>
#> 126 <NA>
#> 127 <NA>
#> 128 <NA>
#> 129 <NA>
#> 130 <NA>
#> 131 <NA>
#> 132 <NA>
#> 133 <NA>
#> 134 <NA>
#> 135 <NA>
#> 136 jackknife_estimate_unavailable
#> 137 <NA>
#> 138 <NA>
#> 139 <NA>
#> 140 <NA>
#> 141 <NA>
#> 142 bayesian_bootstrap_bca_adjustment_on_boundary
#> 143 <NA>
#> 144 <NA>
#> 145 <NA>
#> 146 <NA>
#> 147 <NA>
#> 148 <NA>
#> 149 <NA>
#> 150 <NA>
#> 151 <NA>
#> 152 <NA>
#> 153 <NA>
#> 154 <NA>
#> weight
#> 1 0.033333333
#> 2 0.033333333
#> 3 0.033333333
#> 4 0.033333333
#> 5 0.033333333
#> 6 0.033333333
#> 7 NA
#> 8 0.033333333
#> 9 0.033333333
#> 10 NA
#> 11 0.033333333
#> 12 NA
#> 13 NA
#> 14 NA
#> 15 NA
#> 16 0.033333333
#> 17 0.004016064
#> 18 0.004016064
#> 19 0.004016064
#> 20 0.004016064
#> 21 0.004016064
#> 22 0.004016064
#> 23 0.004016064
#> 24 0.004016064
#> 25 NA
#> 26 0.004016064
#> 27 0.004016064
#> 28 0.004016064
#> 29 0.004016064
#> 30 0.004016064
#> 31 0.004016064
#> 32 NA
#> 33 NA
#> 34 0.004016064
#> 35 NA
#> 36 NA
#> 37 NA
#> 38 NA
#> 39 0.004016064
#> 40 0.004016064
#> 41 0.004016064
#> 42 0.004016064
#> 43 0.004016064
#> 44 0.004016064
#> 45 0.004016064
#> 46 0.004016064
#> 47 0.004016064
#> 48 NA
#> 49 0.004016064
#> 50 0.004016064
#> 51 0.004016064
#> 52 0.004016064
#> 53 0.004016064
#> 54 0.004016064
#> 55 NA
#> 56 0.004016064
#> 57 0.004016064
#> 58 NA
#> 59 0.004016064
#> 60 NA
#> 61 NA
#> 62 NA
#> 63 NA
#> 64 0.004016064
#> 65 0.004016064
#> 66 0.004016064
#> 67 0.004016064
#> 68 0.004016064
#> 69 0.004016064
#> 70 0.004016064
#> 71 0.004016064
#> 72 0.004016064
#> 73 0.004016064
#> 74 0.004016064
#> 75 0.004016064
#> 76 0.004016064
#> 77 NA
#> 78 0.004016064
#> 79 NA
#> 80 0.004016064
#> 81 0.004016064
#> 82 0.004016064
#> 83 0.004016064
#> 84 0.004016064
#> 85 0.004016064
#> 86 NA
#> 87 0.004016064
#> 88 0.004016064
#> 89 NA
#> 90 0.004016064
#> 91 NA
#> 92 NA
#> 93 NA
#> 94 NA
#> 95 0.004016064
#> 96 0.004016064
#> 97 0.004016064
#> 98 0.004016064
#> 99 0.004016064
#> 100 0.004016064
#> 101 0.004016064
#> 102 0.004016064
#> 103 0.004016064
#> 104 0.004016064
#> 105 0.004016064
#> 106 0.004016064
#> 107 0.004016064
#> 108 NA
#> 109 0.004016064
#> 110 NA
#> 111 0.004016064
#> 112 0.004016064
#> 113 0.004016064
#> 114 0.004016064
#> 115 0.004016064
#> 116 0.004016064
#> 117 NA
#> 118 0.004016064
#> 119 0.004016064
#> 120 NA
#> 121 0.004016064
#> 122 NA
#> 123 NA
#> 124 NA
#> 125 NA
#> 126 0.004016064
#> 127 0.004016064
#> 128 0.004016064
#> 129 0.004016064
#> 130 0.020833333
#> 131 0.020833333
#> 132 0.020833333
#> 133 0.020833333
#> 134 0.020833333
#> 135 0.020833333
#> 136 NA
#> 137 0.020833333
#> 138 NA
#> 139 0.020833333
#> 140 0.020833333
#> 141 0.020833333
#> 142 NA
#> 143 0.020833333
#> 144 0.020833333
#> 145 NA
#> 146 0.020833333
#> 147 0.020833333
#> 148 NA
#> 149 0.020833333
#> 150 NA
#> 151 NA
#> 152 NA
#> 153 NA
#> 154 0.020833333
# }