Skip to contents

Print and summarize a TAM fit

Source: R/class.R
tam_fit_summary.Rd

Summarize parameter estimates, uncertainty, convergence information, and terminal population quantities from a fitted TAM model. Estimates and confidence limits are shown with their standard errors and the scale on which those standard errors are reported.

Usage

# S3 method for class 'tam_fit'
print(x, ...)

# S3 method for class 'tam_fit'
summary(object, ...)

# S3 method for class 'summary_tam_fit'
print(x, ...)

Arguments

x, object

A fitted TAM object, or its summary for print.summary_tam_fit().

...

Additional arguments passed to printing functions; unused by summary().

Value

summary() returns a summary_tam_fit list. Its coefficients and terminal_vals components are data frames containing numeric estimates, confidence limits, standard errors, and a character SE scale column. Printing methods return their input invisibly. The confidence level is taken from the tidy tables (95% by default).

Details

Standard errors may be reported on different scales depending on the parameterization. For parameters estimated on a log scale, a small SE approximately corresponds to a coefficient of variation (CV); for example, an SE of 0.10 is approximately 10% relative uncertainty. Logit-scale SEs should not be interpreted as CVs. An SE scale of reported means that the SE is on the same scale as the displayed estimate. This includes quantities such as dq, which are fitted directly as increments on the log-q scale.

Numerical, structural and advisory findings come from check_tam(). Both raw and bound-adjusted gradients are displayed. A parameter at a bound can have a nonzero raw gradient at a valid constrained optimum. Advisory findings do not change numerical convergence and do not establish biological suitability.

Terminal summaries exclude projection years.

See also

tidy_par(), tidy_pop(), check_tam()