delay: Experiment 1 of Dunn, Newell, & Kalish (2012)

In monotonicity/stacmr: Conjoint Monotonic Regression approach to State-Trace Analysis

Description

Data from 130 participants with two between-participants factors (delay and structure) with two levels each and one within-participants factor (block) with four levels. The dependent variable is proportion correct pc.

Usage

delay

Format

An object of class data.frame with 520 rows and 5 columns.

Details

Participants in this study completed one of two category-learning tasks defined according to the category structure that participants learned. For the rule-based (RB) group, the category structure was defined by a simple rule. For the information-integration (II) group, the category structure was more complex and could not be defined by a simple rule.

The experiment consisted of four 80-trial blocks for each participant. Within each block, all 80 stimuli were presented in a random order (with different orders for each participant). Participants were told to learn which of four categories (labeled, 1, 2, 3, and 4) each stimulus belonged to. On each trial, a stimulus was presented, and participants terminated the display by pressing one of the keys labeled 1-4 on the computer keyboard corresponding to Categories 1-4, respectively. Following the response, a mask appeared. The length with which the mask was shown defined the delay condition. Either 0.5-s (No Delay condition) or 5-s (Delay condition).

Following presentation of the mask, feedback appeared on the computer screen for 0.75 s. If the response was correct, the word "Correct" was presented; otherwise, the word "Incorrect" was presented. Following presentation of feedback, a blank screen followed the duration of which was again defined the delay condition. The screen was blank for either 5 s (No Delay condition) or 0.5 s (Delay condition) before the next trial commenced. The sequence and timing of these events were same as those used by Maddox and Ing (2005).

Source

Dunn, J. C., Newell, B. R., & Kalish, M. L. (2012). The effect of feedback delay and feedback type on perceptual category learning: The limits of multiple systems. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38(4), 840-859. https://doi.org/10.1037/a0027867

References

Maddox, W. T., & Ing, A. D. (2005). Delayed Feedback Disrupts the Procedural-Learning System but Not the Hypothesis-Testing System in Perceptual Category Learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 31(1), 100-107. https://doi.org/10.1037/0278-7393.31.1.100

Examples

library("stacmr")

## load data from Exp. 1 of Dunn, Newell, & Kalish (2012)
data(delay)
str(delay, width = 78, strict.width = "cut")
# 'data.frame':	520 obs. of  5 variables:
#  $ participant: Factor w/ 130 levels "1","2","3","4",..: 1 2 3 4 5 6 7 8 9 1..
#  $ delay      : Factor w/ 2 levels "delay","no delay": 1 2 2 1 1 2 2 1 1 1 ...
#  $ structure  : Factor w/ 2 levels "rule-based","information-integration": 1..
#  $ block      : Factor w/ 4 levels "B1","B2","B3",..: 1 1 1 1 1 1 1 1 1 1 ...
#  $ pc         : num  0.338 0.287 0.525 0.35 0.237 ...

stats <- sta_stats(
  data = delay, 
  col_value = "pc", 
  col_participant = "participant",
  col_dv = "structure", 
  col_within = "block", 
  col_between = "delay"
)
stats
str(stats)
summary(stats)

### cmr() fits conjoint monotonic regression for state-trace analysis
## Fit and test CMR State-Trace Analysis Model
st_d1 <- cmr(
  data = delay, 
  col_value = "pc", 
  col_participant = "participant",
  col_dv = "structure", 
  col_within = "block", 
  col_between = "delay", 
  nsample = 1e4
)
st_d1  ## basic information about conjoint-monotonic model

summary(st_d1)  ## basic information plus estimated cell means

# state_trace(st_d1) ## produces state trace plot

# plot_null(st_d1)  ## produces histogram of empirical (boot-strapped) null distribution

str(st_d1, 1, give.attr = FALSE) ## overview of information in fitted object

## same model with approximate method gives same result here
st_d2 <- cmr(
  data = delay, 
  col_value = "pc", 
  col_participant = "participant",
  col_dv = "structure", 
  col_within = "block", 
  col_between = "delay", 
  approx = TRUE, 
  nsample = 1e4
)
summary(st_d2)


### mr() fits monotonic regression with specified partial order

## for delay data: order of factor-levels corresponds to expected partial order.
## Therefore, 'partial = "auto"' can be used to enforce this order.
mr_d1 <- mr(
  data = delay, 
  col_value = "pc", 
  col_participant = "participant",
  col_dv = "structure", 
  col_within = "block", 
  col_between = "delay", 
  nsample = 1e4, 
  partial = "auto"
)
mr_d1

## Alternatively, partial order can be specified symbolically:
mr_d2 <- mr(
  data = delay, 
  col_value = "pc", 
  col_participant = "participant",
  col_dv = "structure", 
  col_within = "block", 
  col_between = "delay", 
  nsample = 1e4, 
  partial = list(
    delay = "delay < `no delay`",
    block = "B1 < B2 < B3 < B4"
  )
)
mr_d2

## Partial order can also be specified partially symbolically:
mr_d3 <- mr(
  data = delay, 
  col_value = "pc", 
  col_participant = "participant",
  col_dv = "structure", 
  col_within = "block", 
  col_between = "delay", 
  nsample = 1e4, 
  partial = list(
    delay = "auto",
    block = "B1 < B2 < B3 < B4"
  )
)
mr_d3


### cmr() also accepts partial order. 
## CMR model is tested against MR model with partial order 
st_d3 <- cmr(
  data = delay, 
  col_value = "pc", 
  col_participant = "participant",
  col_dv = "structure", 
  col_within = "block", 
  col_between = "delay", 
  partial = "auto"
)
st_d3
summary(st_d3)


## p-value now changes somewhat with approximate method:
st_d4 <- cmr(
  data = delay, 
  col_value = "pc", 
  col_participant = "participant",
  col_dv = "structure", 
  col_within = "block", 
  col_between = "delay", 
  partial = "auto",
  approx = TRUE, 
  nsample = 1e4
)
st_d4

monotonicity/stacmr documentation built on Jan. 28, 2020, 3:29 a.m.