assess.variance {bayescount} | R Documentation |
Function used by bayescount.single() and bayescount() to determine if the variance estimate output by a model is sufficiently vague to warrant a warning. Uncertain estimates are often heavily prior-driven, so re-running the model using a different prior and comparing the results is recommended
assess.variance(model, alt.prior, l.95, u.95)
model |
Which model was used to generate the estimate? (string: 'ZI WP/GP/LP') |
alt.prior |
Was the model run with the alternative prior? (logical) |
l.95 |
the lower 95 percent confidence interval for variance (numeric) |
u.95 |
the upper 95 percent confidence interval for variance (numeric) |
A logical result; 'TRUE' if the variance estimate is considered vague or 'FALSE' otherwise.
Matthew Denwood m.denwood@vet.gla.ac.uk funded as part of the DEFRA VTRI project 0101.