RTSYS: A DOS application for the analysis of reaction time data

Heathcote, Andrew · 1996 · Behavior Research Methods, Instruments, & Computers

DOI: 10.3758/bf03200523

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper introduces RTSYS, a menu-driven DOS application designed for the manipulation, analysis, and graphical display of reaction time (RT) data. The development of RTSYS was motivated by two primary challenges in RT research: the presence of fast and slow outliers caused by anticipation or distraction, and the inherent positive skew of RT distributions. These characteristics complicate statistical interpretation, as skew can affect means and medians differently, and outliers can confound results. Furthermore, standard parametric tests often assume normality, which RT data violate, while nonlinear transformations to achieve normality may obscure meaningful information about distribution shape. To address these issues, RTSYS employs the ex-Gaussian distribution, which models RT data as the sum of independent Gaussian and exponential random variables. The software fits this distribution to data using maximum likelihood estimation rather than the less efficient and outlier-sensitive method of moments. This approach provides robust parameter estimates ($\mu$, $\sigma$, and $\tau$) that characterize the location, scale, and asymmetry of the distribution. RTSYS also offers flexible censoring and rescaling options to handle outliers and utilizes Vincent averaging to preserve distribution shape when sample sizes are too small for precise parameter estimation. The application is written in Turbo Pascal and features context-sensitive help, allowing it to run in single-task DOS environments or under Windows. The software automatically processes multiple input files from experiments with arbitrary factorial and nonfactorial designs. It calculates descriptive statistics including the number of RTs, percent censored, median, mean, variance, and a nonparametric skew measure. Additionally, it computes percent error and mean error RT to address speed-accuracy tradeoffs. RTSYS generates formatted output files containing these statistics and ex-Gaussian parameters, which can be directly imported into graphical and inferential statistical packages. The program includes visual inspection tools, such as probability histograms with superimposed fitted ex-Gaussian distributions, to assess goodness of fit. The significance of RTSYS lies in its ability to provide a comprehensive and efficient solution for analyzing complex RT data. By quantifying distribution shape through ex-Gaussian fitting, it reveals structural information in RT data that conventional analyses might miss. The software’s robust estimation techniques and automated processing capabilities reduce the computational burden on researchers, encouraging the exploration of various analytical solutions to the problems of skew and outliers. This facilitates more accurate and interpretable statistical inference in psychological research involving reaction times.

Key finding

RTSYS provides a robust and efficient method for analyzing reaction time data by fitting the ex-Gaussian distribution through maximum likelihood estimation, thereby addressing issues of skew and outliers more effectively than traditional statistical methods.

Methodology

other

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via author_sweep_intake on 2026-05-28.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• perception reaction time