User`s guide

E-Prime User’s Guide
Chapter 3: Critical Timing
Page 93
hidden or filtered using the E-DataAid appliction, the user should consider the long-term benefits
of having the extra timing data available should the timing of the experiment ever be challenged
or require re-examination.
The table below is an example of the output of an analysis in E-DataAid (the data editing and
analysis application within E-Prime). This data is from an experiment presenting a TextDisplay
object named Probe, with Standard data logging enabled, Event timing mode in use, a Duration
of 100ms, 0ms PreRelease, and Onset Sync set to vertical blank. The data file would include the
following (note that the columns in the following tables labeled “Onset-Onset” were created post-
hoc using a spreadsheet in order to provide the onset-to-onset time from one stimulus to the
next
8
):
Example Timing Log for Event Timing Mode Without PreRelease
Stimulus Probe.DurationError Probe.OnsetDelay Probe.OnsetTime Onset-Onset
1 0 17 11778 117
2 0 17 11895 118
3 0 18 12013 117
Stimulus 1 was presented with an OnsetDelay of 17ms from the end of the last display and a
DurationError of 0ms starting at millisecond 11778 from the beginning of the experiment. The last
column (Onset-Onset) shows how long the stimulus was displayed (i.e., actual duration). Note
that with Event mode, the Duration error is 0, and there are onset delays, which are to be
expected. Typically, onset delays are up to the duration of one refresh cycle. The actual time
between probe onsets is the duration of the event (100ms) plus the Probe.OnsetDelay of the next
event. The next table shows the duration errors, onset delay, and onset-onset times for different
timing modes.
Event Timing Mode
No PreRelease
Event Timing Mode with
PreRelease
Cumulative Timing Mode with
PreRelease
Duration
Error
Onset
Delay
Onset-Onset Duration
Error
Onset
Delay
Onset-Onset Duration
Error
Onset
Delay
Onset-Onset
0 18 117 0 0 101 -3 3 101
0 17 118 0 1 100 -4 4 100
0 18 117 0 0 101 -4 4 101
In Event mode there are generally no duration errors, because the specified duration of the event
is maintained. When PreRelease was added to the Event mode, the OnsetDelay decreased from
17-8ms to 0-1ms. For Cumulative mode with PreRelease, the OnsetDelay and the DurationError
match, as the specified duration is shortened by E-Prime to compensate for the delay. The 1ms
variation of onset-to-onset is due to rounding operations in time calculations.
A simple plot of the onset-to-onset times produces timing graphs such as in Figures 1-2 shown
earlier in this chapter. Within a few minutes, you can plot the timing data, determine if there are
any problems in the timing, and verify that the timing matches the specifications. Figure 10 below
shows a plot of onset-to-onset timing for a sequence of displays in Cumulative and Event timing
8
The experimenter must determine what events define the critical onset to onset times. E-Prime logs timing
data on an object-by-object basis (e.g., when did a specific object’s stimulus onset occur). For the between-
object comparisons, a spreadsheet can be used to calculate onset-to-onset times. For example, in a
masking stimulus, the critical onset-to-onset time may be the Probe.OnsetTime to Mask.OnsetTime
difference. This duration can easily be calculated by simply subtracting the respective onset times
(Mask.OnsetTime – Probe.OnsetTime).