DPA.orig: Divergence Point Analysis

In matsukik/RTsurvival: Survival Analaysis and Divergence Point Estimation Methods for Reaction Time Data

Description

Perform Divergence Point Anlysis of Reaction Time Survival Curves as described by Reingold, Reichle, Glaholt, and Sheridan (2012) and by Reingold and Sheridan (2014)

Usage

DPA.orig(subject, latency, condition, binsize = 1, window = 600,
  n.boot = 10000, r.crit = 5, ci.probs = c(0.001, 0.999), quiet = FALSE,
  reorder = TRUE)

DPA.ip(subject, latency, condition, binsize = 1, window = 600,
  n.boot = 1000, num_vbin = 1200, r.crit = 100, ci.probs = c(0.025,
  0.975), quiet = FALSE, reorder = TRUE)

DPA.ci(subject, latency, condition, binsize = 1, window = 600,
  n.boot = 1000, e.crit = 0.015, r.crit = 5, ci.probs = c(0.025, 0.975),
  quiet = FALSE, reorder = TRUE)

Arguments

subject	a vector speficying participant number or ID
latency	a numeric vector containing the latency measures such as reaction time or fixation in millisecond.
condition	a vector or factor specifying the experimental conditions. The conditions should be ordered such that the one that is expected to have shorter latencies (i.e., faster condition) comes first.
binsize	numeric. The size of each timing bins in millisecond. Default to 1.
window	numeric. The maximum window of the timing bin in millisecond. Default to 600.
n.boot	The number of bootstrapping to perform. Default to 10000 for DPA.orig and 1000 for DPA.ip and DPA.ci.
r.crit	The number of bins in a row that needs be significant when determinging the divergence point.
ci.probs	A numeric vector numeric vector of probabilities with values in [0,1], with length(ci.probs) = 2. Default values differ depending on DPA methods.
quiet	logical. If FALSE (default), text progress bar will be displyed.
reorder	logical. If TRUE (default), values of condition is checked to make sure that the order of conditions are coded appropriately (the condition with expected shorter duration comes first) and reorders the condition when needed.
num_vbin	The number of survival bins (or number of data points) to be sampled with replacement. Used only for DPA.ip. Default to 1200.
e.crit	numeric. Divergence point empirical criterion for detecting the signigicant difference between condition on survival proportion. Used only for DPA.ci. Default to 0.015.

Value

A list of class DPA containing:

type	Type of DPA performed (one of 'Original', 'CI', and 'IP')
dp	The divergence point estimate.
ci	The confidence interval of the divergence point.
binsize	The same value supplied in the argument binsize
window	The same value supplied in the argument window
n.boot	The same value supplied in the argument n.boot

Items specific to DPA.orig:

dp.max	The maximum of the significant divergence points.
dp.vec	A vector containing all the significant divergence points.

The fucntion DPA.ip additionally returns a data.frame dp_matrix that contins individual participant specific estimates. The dara.frame has following columns:

subject	Participant IDs corresponding to the values specified in the argument subjects
dpcount	The number of bootstrap samples from which divergence points were observed.
median_dp_point	the median of the divergence point estimates in the unit of tthe number of survival bins
median_duration	the median of the divergence point estimates in duration.
ci.lower	lower confidece interval of the divergence point estimate (in duration)
ci.upper	upper confidece interval of the divergence point estimate (in duration)

The function DPA.ci additionally returns a data.frame DV_strap_dat with following columns:

divergence_point_strap	A value for each bootstrap samples corresponding to the posiiton of the bin at which the divergence was found.
percentage_below_strap	corresponds to the percentage of the data points that 'died off' before the observed divergence point.
area_measure_strap	corresponds to the area (in percentage) formed by the difference in two survival curve proportional to the area below the curve for 'slower' condition.

References

Reingold, E. M. & Sheridan, H. (2014). Estimating the divergence point: A novel distributional analysis procedure for determining the onset of the influence of experimental variables. Frontiers in Psychology. doi: 10.3389/fpsyg.2014.01432.

Reingold, E. M., Reichle, E. D., Glaholt, M. G., & Sheridan, H. (2012). Direct lexical control of eye movements in reading: Evidence from a survival analysis of fixation durations. Cognitive Psychology, 65, 177-206.

Examples

data(DPAsample)
## n.boot is set to 100 here just so it runs and
## finishes within a reasonal amount of time
o.dpa <- DPA.orig(DPAsample$subject, DPAsample$duration, DPAsample$condition, n.boot = 100)

msc1 <- surv.mean(DPAsample$subject, DPAsample$duration, DPAsample$condition)
plot(msc1, dp.point = o.dpa, add.arrows = TRUE)

ci.dpa <- DPA.ci(DPAsample$subject, DPAsample$duration, DPAsample$condition)
plot(msc1, dp.point = ci.dpa, add.arrows = TRUE)

matsukik/RTsurvival documentation built on May 21, 2019, 12:57 p.m.