The guess
argument of quickpsy
allows you to set the leftward asymptote of the psychometric function.
The default value is 0
, but it can be changed:
library(quickpsy)
fit <- quickpsy(qpdat, phase, resp, grouping = c("participant", "cond"),
guess = 0.05, bootstrap = "none")
plot(fit, color = cond)
The leftward asymptote can also be free as a parameter (it wouldn’t be a guess rate in this case)
fit <- quickpsy(qpdat, phase, resp, grouping = c("participant", "cond"),
guess = TRUE, B = 10) # only 10 bootstrap samples
plot(fit, color = cond)
The lapses
argument of quickpsy
works as the guess
argument.
We can combined guesses and lapses. For example, for a task with a fixed guess rate of 50% and a variable lapse rate parameter we can do
library(ggplot2)
n <- 100
x <- c(.2, .4, .6, .8, 1)
k <- c(59, 56, 69, 92, 94)
y <- k / n
dat <- data.frame(x, k, y)
fit <- quickpsy(dat, x, k, n, guess = 0.5, lapses = TRUE, bootstrap = "none")
plot(fit) + ylim(0.25, 1)
Observe that the lapse rate parameter is now the p3
parameter (p1
and p2
correspond to the position and scale parameters respectively)
fit$par
#> # A tibble: 3 x 2
#> parn par
#> <chr> <dbl>
#> 1 p1 0.610
#> 2 p2 0.136
#> 3 p3 0.0545
If both the leftward and the rightward asymptotes are free, they will appear as the p3
and p4
parameters respectively.