post_hoc_estimation: Post hoc estimations

In klausfrieler/HALT: Headphone and Loudspeaker Test

Post hoc estimations

Description

This function provides probabilistic statements about the composition of a sample after application of a certain test procedure within a screening strategy. For this purpose a Binomial distribution is used.

Usage

post_hoc_estimation(
  screening_strat,
  combination_method,
  A_threshold,
  B_threshold,
  C_threshold,
  baserate_hp = NA,
  devices,
  samplesize = NA,
  target_selfreported = NA,
  target_tested = NA,
  switch_to_target = NA,
  min_number = NULL,
  min_prob = NULL
)

Arguments

screening_strat	three-letter lower-case abbreviation of the screening strategy. See screening_strategies
combination_method	Number (1 to 18) corresponding to the test method.
A_threshold	(scalar integer) Threshold for Test A (1 to 6).
B_threshold	(scalar integer) Threshold for Test B (1 to 6).
C_threshold	(scalar integer) Threshold for Test C (1 to 6).
baserate_hp	Sets the (estimated) prevalence of headphones in the target population as a number between 0 and 1. Defaults to the unbiased prevalence B of 0.1767 from \insertCiteHALT_2;textualHALT.
devices	Sets the desired playback device. Possible settings are "HP" for headphones or "LS" for loudspeakers.
samplesize	Number of participants who got a test result indicating the use of the target device and who are therefore in the final sample. samplesize is needed when screening_strat is "fwr" or "far".
target_selfreported	Number of participants who reported the use of the target device. target_selfreported is needed when screening_strat is "scc".
target_tested	Number of participants who reported the use of a playback device other than the target device and got a test result indicating the use of the target device. target_tested is needed when screening_strat is "scc".
switch_to_target	Sets the (estimated) switching prevalence. The switching prevalence describes the probability that a participant who indicates the use of a device other than the target device actually switches to the target device after being prompted to do so.
min_number	minimum number of participants k who are in the final sample and should have used the correct device. When you use min_number you cannot use min_prob, i.e. min_prob = NULL.
min_prob	(greater than 0, less than 1) for the event that at least an unknown number of participants k who are in the final sample used the correct device. When you use min_prob you cannot use min_number, i.e. min_number = NULL.

Details

The event that at least k participants in the final sample used the target device is considered. The function determines either the probability of this event or for a given probability the number k.

Depending on the screening strategy either samplesize and baserate_hp (for "fwr" or "far") or target_selfreported, target_tested, and switch_to_target (for "scc") are needed.

For details, see \insertCiteHALT_2;textualHALT.

Note

Only one of the arguments min_number and min_prob can be used.

References

\insertAllCited

klausfrieler/HALT documentation built on March 17, 2023, 6:18 a.m.