Read out certain levels only
David McFarlane
mcfarla9 at msu.edu
Tue May 11 14:33:51 UTC 2010
Vera,
Don't know why your code does not work, but just
a couple comments on your code fragment here...
First, almost every time we have resorted to
using List.SetWeight it turned out that we had a
poor design structure, and once we restructured
the design it eliminated the need for
List.SetWeight and solved a host of other lurking
problems as well. So I urge you to first take a
serious look at your design structure.
Next, when we did use List.SetWeight, we
absolutely had to execute a List.Reset, otherwise
the List ignored our .SetWeight changes. So if
your code does not work with the List.Reset
command in place, then your code has other problems.
And since you must do the List.Reset for either
branch of the If...Then...ElseIf, then you may as
well pull it out from within the branches and put
it after the whole If...Then...ElseIf block.
Finally, what happens if TriggerValue < 0 or
TriggerValue >= 7? As it stands, in this case
your code will simply leave the List intact. Is
this what you meant? Perhaps you think it
impossible for TriggerValue to ever take on
values outside of 0-6, but good coding practice
requires that you either include a final Else to
handle outliers or at least add a comment to
explain what you mean the code to do.
So just as an exercise, here is your code
fragement again (keeping your TriggerValue
variable because I do not want to get into that
issue, using "<=" instead of "<" just because I
think that makes the intent clearer, and adding a
Const to get rid of one "magic number"; hmm, on
further thought I restructured this to pull the
common inner loops out of the main loop, etc.):
Const TriggerValueMax as Integer = 6
Dim nLevel as Integer
Dim TriggerValue as Integer, TriggerValueToRun as Integer
' Assign TriggerValueToRun based on NewTriggerValue:
If (NewTriggerValue = 0) Then
TriggerValueToRun = 0
ElseIf ((1 <= NewTriggerValue) AND _
(NewTriggerValue <= TriggerValueMax)) Then
TriggerValueToRun = 1
Else MsgBox "NewTriggerValue " & NewTriggerValue _
& "out of range! Please contact programmer."
End If
' Now set List level weights according to match between TriggerValue &
' TriggerValueToRun:
For nLevel = 1 to DesignList.Size
TriggerValue = DesignList.GetAttrib(nLevel,"Trigger")
If TriggerValue = TriggerValueToRun Then
DesignList.SetWeight nLevel, 1
Else DesignList.SetWeight nLevel, 0
End If
Next nLevel
DesignList.Reset
Or just for fun, let's do this with a Select Case:
Const TriggerValueMax as Integer = 6
Dim nLevel as Integer
Dim TriggerValue as Integer, TriggerValueToRun as Integer
' Assign TriggerValueToRun based on NewTriggerValue:
Select Case NewTriggerValue
Case 0
TriggerValueToRun = 0
Case 1 to 6
TriggerValueToRun = 1
Case Else
MsgBox "NewTriggerValue " & NewTriggerValue _
& "out of range! Please contact programmer."
End Select
' Now set List level weights according to match between TriggerValue &
' TriggerValueToRun:
For nLevel = 1 to DesignList.Size
TriggerValue = DesignList.GetAttrib(nLevel,"Trigger")
If TriggerValue = TriggerValueToRun Then
DesignList.SetWeight nLevel, 1
Else DesignList.SetWeight nLevel, 0
End If
Next nLevel
DesignList.Reset
-- David McFarlane, Professional Faultfinder
At 5/7/2010 01:11 PM Friday, you wrote:
>:-(
>
>I seem to have cheered too early. :-( Some filteringin E-DataAid
>showed that E-Prime is still exactly doing as it likes.
>
>I also had some mistakes in the previous code, here it is again:
>
>Dim nLevel as Integer
>Dim TriggerValue as Integer
>
>If (NewTriggerValue > 0) AND (NewTriggerValue < 7) Then
> For nLevel = 1 to DesignList.Size
> TriggerValue = DesignList.GetAttrib(nLevel,"Trigger")
> If TriggerValue = 1 Then
> DesignList.SetWeight nLevel, 1
> Else DesignList.SetWeight nLevel, 0
> End If
> Next nLevel
>' DesignList.Reset
>ElseIf NewTriggerValue = 0 Then
> For nLevel = 1 to DesignList.Size
> TriggerValue = DesignList.GetAttrib(nLevel,"Trigger")
> If TriggerValue = 0 Then
> DesignList.SetWeight nLevel, 1
> Else DesignList.SetWeight nLevel, 0
> End If
> Next nLevel
>' DesignList.Reset
>End If
>
>A few remarks to this code:
>
>1. E-Prime is still doing what it likes, guaranteeing me no
>counterbalancing whatsoever.
>2. I had to add: TriggerValue = DesignList.GetAttrib(nLevel,"Trigger")
>- If TriggerValue = 0 Then and I know I could have been doing that
>more elegantly (like you did Michiel), but then E-Prime complains that
>there are "too many variables".
>3. If I add this piece of code, not only will E-Prime still not do
>what I expect it to do, but it will also generate blocks as it likes.
>Sometimes 3, sometimes 4... But never 7 as it is supposed to do!! In
>my last test run it even stopped in the middle of a trial going
>immediately to the block feedback! Anybody any idea why that could be?
>It seems to me that this piece of script has nothing to do with the
>number of blocks or anything.
>4. As you can see, DesignList.Reset has been commented out, because
>when I add this, E-Prime runs only one trials (instead of 36 in a
>normal block) and then goes immediately to the Block Feedback.
>
>I don't understand this program, it makes absolutely no sense to me. :-
>(
>
>Greetings, Vera
>
>
>
>
>On 7 Mai, 17:05, Vera <vera.d... at googlemail.com> wrote:
> > Ok, found a way to be at the experimental setup AND have internet
> > (whew!!!). ;-)
> >
> > So the thing I did was:
> >
> > If (NewTriggerValue > 0) AND (NewTriggerValue < 7) Then
> > For nLevel = 1 to DesignList.Size
> > If c.GetAttrib("Trigger") = "1" Then
> > DesignList.SetWeight nLevel, 1
> > Else DesignList.SetWeight nLevel, 0
> > End If
> > Next nLevel
> > ElseIf NewTriggerValue = 0 Then
> > For nLevel = 1 to DesignList.Size
> > If c.GetAttrib("Trigger") = "0" Then
> > DesignList.SetWeight nLevel, 1
> > Else DesignList.SetWeight nLevel, 0
> > End If
> > Next nLevel
> > End If
> >
> > And now that I could finally connect, I see that you proposed me the
> > same thing. :-)
> > And the best thing: it works!!!! Checked the database with E-DataAid
> > and it's all perfectly balanced. :-)
> >
> > Thanks a lot for giving me the idea!
> >
> > Have a great weekend!
> >
> > Vera
> >
> > On 7 Mai, 13:42, Vera <vera.d... at googlemail.com> wrote:
> >
> > > Ok, I now had a closer look at your solution (I wasn't on the right PC
> > > before, with the USB-stick-licence it's kind of a hassle sometimes,
> > > making a lot of kms in the institute! ;-) ) and I think I might try
> > > something with your solution.
> >
> > > So to get back at your question:
> >
> > > "If trigger 1 then PLEASE take only line 1"
> > > ...
> > > "You mean, of a list, right?"
> >
> > > Yes, I mean of a list (so a level of a list). Of course I simplified
> > > the whole a bit, pretending that I had only two levels in my list and
> > > two triggers. In reality I have 20 lines and two triggers (so the list
> > > kind of gets separated into two). :-)
> >
> > > So basically what I want to do is
> > > "If trigger = 1 then take any of the lines for which attribute.trigger
> > > = 1" (any of 10 in a random manner)
> >
> > > Oh and then don't worry about TMS, it's already all working fine. :-)
> > > I am getting the triggers and all, E-Prime just doesn't know what to
> > > do with them for the moment (ok if I am more precise: E-Prime knows
> > > what to do with the triggers, but I can't control the conditions,
> > > making sure that the design is completely perfectly counterbalanced
> > > over the two conditions). :-)
> >
> > > I am going back to my experimental setup again, seeing what I can do
> > > with your solution in mind. I definitely should get internet there,
> > > but that's the hassle with experimental computers: you want to keep
> > > them clean from anything that might slow them such as anti-virus and
> > > stuff. :-| But feel free to reply if you came up with another idea, I
> > > will check for input anyway. :-)
> >
> > > Greetings, Vera
> >
> > > On May 7, 12:25 pm, Michiel Spape <Michiel.Sp... at nottingham.ac.uk>
> > > wrote:
> >
> > > > Hi Vera & Group,
> > > > I like that "Sudoku-effect"! In fact, I
> think I might have misstated that nested-lists
> are untrustworthy - in general, people are much
> more prone to making errors than computers, and
> indeed, this to me is the main problem of
> nested lists. I do, however, use them all the
> time, myself, but usually not for my main 'triallist'.
> >
> > > > Anyway, I'm sorry for the lack of help
> I'm going to offer, I'm just not entirely clear
> on what you want to achieve. Do you want to run
> certain conditions based on a trigger readout
> from TMS (err, I never worked with TMS)? Does
> that mean certain stimuli? Certain different orderings?
> >
> > > > Lacking a clear idea on what exactly it
> is you want, I can help with certain specifics:
> >
> > > > "If trigger 1 then PLEASE take only line 1"
> > > > ...
> > > > You mean, of a list, right?
> >
> > > > This can be achieved in a quick and dirty
> way by adding a bit of inline that sets all
> weights of a list to 0 except line 1. From E-basic help, and adjusted
> >
> > > > Dim nLevel As Integer 'originally Long,
> but how many people really have lists that have more than 32768 levels?
> > > > For nLevel = 1 To List1.Size
> > > > If nLevel = 1 then
> List1.SetWeight nLevel, 1 else List1.SetWeight nLevel, 0
> > > > Next 'nLevel
> > > > List1.Reset
> >
> > > > You can even put this List1 in
> unreferenced objects (i.e. your garbage can) and run it on request (List1.Run).
> > > > Hope any of this helps (I might be way
> off in understanding your problem),
> > > > Mich
> >
> > > > Michiel Spapé
> > > > Research Fellow
> > > > Perception & Action group
> > > > University of Nottingham
> > > > School of Psychology
> >
> > > > -----Original Message-----
> > > > From: e-prime at googlegroups.com
> [mailto:e-prime at googlegroups.com] On Behalf Of Vera
> > > > Sent: 07 May 2010 08:24
> > > > To: E-Prime
> > > > Subject: Re: Read out certain levels only
> >
> > > > Ok guys, just to let you know that I am really trying here, a small
> > > > update:
> >
> > > > Yesterday I tried to solve this problem with nested lists (knowing
> > > > that nested lists very quickly get very complicated and taking into
> > > > account that Michiel pointed out in his E-Prime doc that the random
> > > > doesn't seem to be completely "trustable") but the problem is that I
> > > > get to the same thing: I would basically have two lines, 1 pointing to
> > > > the condition for trigger 1 and another one pointing to trigger 2. But
> > > > then, how to access this line??
> > > > I need to figure out some way to tell E-Prime:
> >
> > > > If trigger 1 then PLEASE take only line 1
> >
> > > > (Or in the old solution, which I very much preferred, because it was
> > > > much cleaner - nested lists do have a high "sudoku-effect"):
> >
> > > > If trigger 1 PLEASE take any of the lines for which attribute.trigger1
> > > > = 1
> >
> > > > Down to the beginning again, I don't know how to solve this. :-( Any
> > > > ideas are very welcome. :-)
> >
> > > > Greetings, Vera
> >
> > > > On May 6, 2:31 pm, Vera <vera.d... at googlemail.com> wrote:
> > > > > Ok, got a bit further (but it's still not working fine):
> >
> > > > > I added this inline:
> >
> > > > > If (NewTriggerValue > 0) AND (NewTriggerValue < 7) Then
> > > > > Set DesignList.Deletion =
> PickOne(c.GetAttrib("TriggerPos") = "1")
> > > > > ElseIf NewTriggerValue = 0 Then
> > > > > Set DesignList.Deletion =
> PickOne(c.GetAttrib("TriggerPos") = "0")
> > > > > End If
> >
> > > > > It's based on something David once wrote concerning counterbalancing
> > > > > (it would have been so much easier if E-Prime could just
> > > > > counterbalance for any attribute and not just Group, Subject and
> > > > > Session). I am actually hoping to tell E-Prime with this, that it
> > > > > needs to take any of the 18 levels which has either 1 or 0 in the
> > > > > TriggerPos attribute. It doesn't seem to work though. :-(
> >
> > > > > E-Prime now only runs 1 trials (instead of 36). Any hints on that?
> >
> > > > > Grrr, I know I am not very far from a solution...
> >
> > > > > Greetings, Vera
> >
> > > > > On May 6, 11:17 am, Vera <vera.d... at googlemail.com> wrote:
> >
> > > > > > Dear forum members,
> >
> > > > > > I already got some very useful help
> here and as I am now running into
> > > > > > a new problem with E-Prime, I will
> see if you can help me out (again).
> > > > > > I did all my homework: yes I looked at the reference guide, getting
> > > > > > started guide, all other documents concerning E-Prime, yes I combed
> > > > > > through this forum, yes I combed through the PST-forum. :-)
> >
> > > > > > So here's the problem:
> >
> > > > > > I created a list of conditions (and all that has been working fine
> > > > > > until now) and as we are moving on to
> another experimental setup, it
> > > > > > is now very important that half of
> the conditions appear on one type
> > > > > > of stimulus and the other half of
> conditions appear on another type of
> > > > > > stimulus.
> >
> > > > > > To be more clear:
> >
> > > > > > On Trigger1 I want only a certain
> type of the 18 conditions I have to
> > > > > > be executed.
> > > > > > On Trigger2 I want the other 18 conditions (another type) to be
> > > > > > executed.
> >
> > > > > > I know I could be able to do this
> over nested lists (probably), but it
> > > > > > would take a lot of time too, as then
> "big surgery" has to be done on
> > > > > > my program. ;-)
> >
> > > > > > So I figured that there should be
> some way to add a few lines of code,
> > > > > > which tell E-Prime that it should only use a certain level in a
> > > > > > certain condition.
> >
> > > > > > I wrote:
> >
> > > > > > If (NewTriggerValue > 0) AND (NewTriggerValue < 7) Then
> > > > > > c.getAttrib("TriggerPos") = 1
> > > > > > ElseIf NewTriggerValue = 0 Then
> > > > > > c.getAttrib("TriggerPos") = 0
> > > > > > End If
> >
> > > > > > TriggerPos = the type of trigger (1 or 0) which I added as an
> > > > > > attribute in my designlist.
> > > > > > When I try to execute, Eprime tells me "Can't Assign to Constant"
> >
> > > > > > Ok. But what do I need to add then in
> order to make it work? Basically
> > > > > > I want Eprime in certain conditions
> only to use the levels that have
> > > > > > "1" in the attribute "TriggerPos".
> >
> > > > > > I know that I am not very far from a
> solution. Any help would be very
> > > > > > much appreciated. :-)
> >
> > > > > > Vera
--
You received this message because you are subscribed to the Google Groups "E-Prime" group.
To post to this group, send email to e-prime at googlegroups.com.
To unsubscribe from this group, send email to e-prime+unsubscribe at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/e-prime?hl=en.
More information about the Eprime
mailing list