Abstract
quickpsy is a package to parametrically fit psychometric functions. In comparison with previous R packages, quickpsy was build to easily fit and plot data for multiple groups. Here, we describe the standard parametric model used to fit psychometric functions and the standard estimation of its parameters using maximum likelihood. We also provide examples of usage of quickpsy, including how allowing the lapse rate to vary can sometimes eliminate the bias in parameter estimation, but not in general. Finally, we describe some implementation details, such as how to avoid the problems associated to round-off errors in the maximisation of the likelihood or the use of closures and non-standard evaluation functions.
Highlights
F is the probability mass function of the model or the likelihood when considered as a function of the parameters;
– θ = (α, β, γ, λ) is the vector of parameters that define the parametric family of probability mass functions of the model. α and β are the position and scale parameters. γ and λ are the parameters corresponding to the leftward and rightward asymptote of ψ
The data from the experiment is available in MPDiR, a package that includes material from the book Modeling Psychophysical Data in R (Knoblauch and Maloney, 2012)
Summary
Abstract quickpsy is a package to parametrically fit psychometric functions. In comparison with previous R packages, quickpsy was built to fit and plot data for multiple groups. We describe the standard parametric model used to fit psychometric functions and the standard estimation of its parameters using maximum likelihood. We provide examples of usage of quickpsy, including how allowing the lapse rate to vary can sometimes eliminate the bias in parameter estimation, but not in general. We describe some implementation details, such as how to avoid the problems associated to round-off errors in the maximisation of the likelihood or the use of closures and non-standard evaluation functions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
