Skip to contents

About rtprep

rtprep rtprep-package
rtprep: Screening, Trimming, and Aggregating Response Time Data
rtprep-glossary glossary
Terms used in rtprep

Screening

One engine, one return shape, whichever rule is applied.

rt_screen()
Screen response times with any rule
rt_keep()
Keep vector from a screening rule
screen_fits()
Per-group fit diagnostics from a screening rule
rule_cutoff() rule_sd() rule_mad() rule_iqr() rule_recursive() rule_ewma() rule_mixture() rule_none() rule_oracle() print(<rtprep_rule>)
Screening rules
format(<rtprep_screen>) print(<rtprep_screen>) as.data.frame(<rtprep_screen>) `[`(<rtprep_screen>)
Methods for a screening result

Combining rules

No single conventional rule reaches both the slow tail and the leading edge; these put two together.

rule_all() rule_any() rule_then()
Combine screening rules
rule_hierarchical()
Hierarchical screening

Reporting

The Methods paragraph a screen owes a reader, written from the screen rather than from memory.

Aggregation and estimation

From surviving trials to EZ-diffusion inputs and parameters.

rt_summary()
Aggregate response times into EZ-diffusion summary statistics
adjust_accuracy()
Correct accuracy counts for contamination
ez_ddm()
Invert summary statistics into diffusion parameters

Comparison and diagnostics

What a different choice would have removed, and whether the removed trials were what you thought.

screen_compare() print(<rtprep_comparison>) summary(<rtprep_comparison>) print(<rtprep_comparison_summary>) plot(<rtprep_comparison>)
Compare what several screening rules would remove
check_guessing()
Test whether the fast trials a rule removed were really guesses

Ground truth

Generate data whose contaminants are labelled, and remove exactly those, so a screen can be scored and read against perfect exclusion.

Extending rtprep

Adding a screening rule of your own, from a function or from a method.

new_rule() apply_rule() apply_rule_grouped()
Add a screening rule
rule_custom()
A screening rule from a function, in one call

Data

rt_example
Simulated response times from four participants, with ground truth