|
|
|
Forum Newbie
Group: Forum Members
Last Login: 9/22/2009 1:55:35 PM
Posts: 8,
Visits: 21
|
|
I've found what seems to be a pretty serious problem with running E-Prime 2 under Vista, and I've replicated it on 3 different machines in our lab.
I do various types of experiments where stimuli are presented on the screen, and participants respond by pressing buttons on the keyboard. I am interested in the speed of their reaction times. There are many trials, for example a current experiment collects about 1000 RTs for each participant.
Looking at the distribution of RTs revealed that it is very non-normal, with some tall peaks separated by low valleys. For example in a current experiment I collected a total of 28,642 RTs across all participants. According to the .edat file, on 1002 independent trials participants responded with an RT of 253ms, but participants responded with an RT of 254ms on only 5 trials. A table representing a subset of the data is below, with each ms RT on the left, and the number of trials in the experiment on which that RT was recorded on the right.
milliseconds frequency of RT
215 2
216 3
217 3
218 11
219 13
220 501
221 1822
222 217
223 5
224 6
225 2
226 4
227 2
228 4
229 1
231 2
233 2
234 4
235 97
236 186
237 4
238 1
239 1
240 2
241 2
242 2
243 1
244 3
246 3
247 3
248 1
249 3
250 16
251 129
252 1650
253 1002
254 5
255 5
256 2
There seems to be some serious binning of the RT data going on, because there's no way a group of humans actually produced such consistent RT responses. I've replicated this problem with a different script on all 3 Vista machines in our lab; the problem did not occur on any XP machine I've tested.
RefreshClockTest correctly identifies the Vista machines as unreliable. The only solution I've found is to format the computer and install XP on it, after which RT logging works quite well. Is PST aware of this problem? Is there a fix to it? For now it's not difficult to just use XP for running participants, but it is already a dated OS and becoming more so as Microsoft continues to put its resources toward Vista.
|
|
|
|
|
Forum Newbie
Group: Forum Members
Last Login: 5/6/2009 4:18:10 PM
Posts: 1,
Visits: 3
|
|
| We have seen the exact same issue on our Vista machine. Is there a way to configure Vista so that Eprime2 does not exhibit these problems? If not, we are looking at the same solution - reformat for XP. Any input on a fix, other than reformat, would be appreciated. Thanks.
|
|
|
|
|
Forum MVP
Group: Administrators
Last Login: Yesterday @ 9:44:24 PM
Posts: 698,
Visits: 1,481
|
|
| Anyone exhibiting this issue should create a web support ticket and include the following... .es2 file
.wav, etc or subset of audio files
.edat2 that has the issue A listing of machine type, service pack, and type of sound card used.
Any information if the error is specific to one machine or entire lab.
|
|
|
|
|
Forum Newbie
Group: Forum Members
Last Login: 7/2/2010 11:04:41 AM
Posts: 6,
Visits: 17
|
|
I was wondering what type of response you got about the binning RT issue both of you reported - was it clarified by PST? Was it an E-Prime 2 issue exclusively, or was the problem evident in E-Prime 1 too?
Thanks!
|
|
|
|
|
Forum Newbie
Group: Banned Members
Last Login: 6/13/2010 3:29:36 AM
Posts: 2,
Visits: 10
|
|
|
|
|
