User`s guide

E-Prime User’s Guide
Chapter 3: Critical Timing
Page 84
In response to the Measuring, Reporting, and Debugging problem, E-Prime has a built in Time
Audit logging capabilities available for every object. When an object takes an action that could
impact a subject (e.g., display, response input), E-Prime logs the time of that event, and this
information may be output to the analysis stream for easy recording and analysis. For example,
in minutes it is possible to get a listing of the millisecond timestamps at which critical probe stimuli
occurred, as well as the identity of each stimulus. This virtually eliminates the need to write down
what stimuli happened and when. You can output this timing log to a spreadsheet such as
Microsoft Excel. You can then use Excel to create figures that profile the timing (as was done for
Figures 1-3) and use these figures to quickly detect anomalies in the data, hunt them down, and
correct them before running subjects for a research study. E-Prime includes sample paradigms
that illustrate how this can be done (see Appendix A, this volume).
Getting millisecond accuracy and precision out of modern computers requires an awareness of
the problems, tools that minimize and report those problems when they occur, and vigilance of
the researcher to monitor the precision of measurement as the data are collected.
3.3 Achieving Accurate Timing in E-Prime
E-Prime provides both millisecond precise timing and tools to allow for the accurate interpretation
and reporting of timing. In this section, we present the concepts needed to master the
understanding of how to maximize the accuracy and precision of timing in experiments. We also
present specific techniques to implement for various common classes of paradigms. Although
modern computers are very fast, they can still be too slow to perform your intended experiment
with the timing accuracy and precision that you want, unless you specify when the computer can
perform preparation activities. For example, if a computer is asked to present a series of 10
images every 100ms and this request is specified as a simple, sequential sequence of activities,
then rest assured the computer will not be able to perform this action correctly. The problem is
that this simple, sequential specification provides no time for the loading of images and setup of
the display. However, if you understand how to specify your experiment in a way that matches
the abilities of the computer, you can achieve the display timing that you require. For example, in
the experiment, write script to preload the 10 images before the trial begins, thus reducing the
time necessary to load the images during the trial. Also, during the period when the first display
is up, specify that extra time available be pre-released to prepare the next display. Thus, the
setup time for the second display is absorbed by the first display, which allows the second display
to be displayed at the required 100ms delay after the beginning of the first display.
E-Prime offers the tools to make critical timing specifications easy. However, you must
understand and choose the correct options to allow E-Prime to optimize your experiment in order
to get the timing desired. Since timing specifications are critical and somewhat difficult concepts
at first, it is important to analyze timing results of the experiment to debug the design and
accurately report data.
3.3.1 Basic timing techniques and implementations in
E-Prime
In this section, we describe the critical timing concepts needed to interpret and specify
experiments. The complexity of the experimental specification varies with the timing precision
demands of the experiments. If the experiments require only that a given display is presented
and that an accurate reaction time is collected from the beginning of the display, move on to
Paradigm 1 in section 3.4.2.3. However, if the timing needs demand that timing be considered
between displays, precisely presenting short duration displays, presenting displays at a fixed rate,