Informative-hypothesis Bayes factors test competing order and equality constraints on group means — 'A < B < C' against 'A = B = C' — reporting fractional Bayes factors and posterior hypothesis probabilities.
Classical ANOVA answers 'are the means equal?'; researchers usually believe something more specific — a direction or an ordering. Informative-hypothesis testing evaluates those theories directly: each constraint set gets a Bayes factor against the unconstrained model (adjusted fractional BF, O'Hagan/Mulder), and the competing hypotheses plus an automatic complement are ranked by posterior probability.
Coefficients are the group means (dummy coding without intercept), named by group level so hypotheses read naturally. The exploratory table additionally tests each mean for = 0 / < 0 / > 0.
Response (numeric) + group column + optional numeric covariates + semicolon-separated hypotheses on level names (empty = exploratory only).
Per-hypothesis BF vs unconstrained + posterior probabilities (best in bold), exploratory =0/<0/>0 table per coefficient.
PHP 0.95 for 'A < B < C' means: among your stated theories (plus 'anything else'), the ordering has 95% posterior support. It is relative evidence — a poorly specified competitor set inflates it.