[solved] Trouble Creating Progressive Experiment using Run if

glawton

I want to build a collection of phases in an experiment, in which each successive test sequence is only run if the response is below a certain threshold (2000 ms). I am finding the sequence never runs when I set the "run_if" panel for a successive sequence as [avg_rt]<2000 and always runs when I set [avg_rt]>2000.

I am not sure if this is a consequence of the way I am setting and getting avg_rt or if this a bug in the program. The feedback part seems to get the number right, but then the conditional logic part does not work.

Thanks,
George

Here is the code from the sequence in question.

reset_feedback "always"

run block_loop "always"

run loop "[avg_rt]<2000"

run feedback "always"

Here is the whole code:
[gist:3842684]

I also tried toying with the affect orientation experiment example, because I noticed the feedback loop would report using the [avg_rt] variable. However, I still could not get a follow-on sequence to trigger when run if [avg_rt]<2000 was specified.

[gist:3842650]

Any help or coaching of better approaches to make this work would be greatly appreciated.

--George

lvanderlinden

EDIT: The new variable should not be called "run", because a built-in variable named "run" already exists. (See posts below.)

Hi George,

Thanks for your post! You discovered a bug which Sebastiaan plans to fix for the next version of OpenSesame. The reason your experiment does not work as expected is that values that are entered in the interface are always interpreted as being a string. Therefore, comparisons like [avg_rt] > 2000 do not work (yet).

For now, the best way to get around this issue is by using a bit of Python inline code to make a string variable to indicate whether or not to run a certain sequence. This can be achieved like so:

if self.get("avg_rt") < 2000:

    exp.set("run","yes")

else:

    exp.set("run", "no")

If you place this piece of code in an inline_script item just before the loop where you need the information (i.e. your loop called "loop") you can get the desired result by setting the run-if statement to [run] = "yes". Note also that in the interface such statements should always have spaces in between the words (so [run]="yes", without spaces, won't work).

I hope this helps! Please feel free to ask any further questions!!

Best wishes,

Lotje

glawton

Thanks for the coaching. I tried it with the modifications you suggested, and still ran into the same problem. I also tried running the old code on the beta version (.27) , as I read that there had been some work to fix this bug.

Here is the updated code:
[gist: 3853813]

lvanderlinden

Hi George,

Did you place the suggested Python code in the "Run phase" tab of your inline_script item? The variable avg_rt is calculated during the run phase of a sequence, so if the code is placed in the "Prepare phase" tab it won't work.

Another possibility is that avg_rt itself isn't calculated correctly. You could easily check this by either printing its value to the debug window in your inline_script item:

print exp.get("avg_rt")

or by making a temporary feedback item to show the variable (simply use the text tool to show [avg_rt]).

Please let us know if you still encounter the same problem. (It seems like something went wrong with uploading your experiment, but if you add the updated version we can have a closer look at the code if necessary.)

Best wishes,

Lotje

Edwin

Hi all!

The problem might arise from the fact that the runif variable is called 'run'. I've tested it on OpenSesame Portable 0.26 (should be the same as running 0.26 from source on Windows) and using a variable called run just skips all of the elements, whereas a variable called runn (or anything similar to 'run' that isn't acutually 'run') does work. (I suspect there already is a variable called 'run', probably created somewhere during the running of a sequence, but I can't find the actual creating nor addressing this variable in the source code)

On the spaces: do they really have to be in there? Since my previous experiments (0.25 and 0.26 on WinXP and Win7) all work without spaces in Run if statements (e.g. '[runn] = yes" works just as well as "[run]=yes]"). Is this something specific to 0.27 or to a different OS?

Thanks!

Edwin

sebastiaan

Ah, right, of course...

What exp.set() essentially does is add a variable as an attribute to self.experiment. Since self.experiment already has an attribute named run (namely the function that handles the running!) doing exp.set('run', ...) causes OpenSesame to malfunction. The user should be protected from this, so I created an issue:

https://github.com/smathot/OpenSesame/issues/105

Regarding the spaces in Run if statements: In some cases it may work, but sometimes it may not, so I would recommend using spaces for now. For 0.27, I plan to improve the parsing of the Run if statements (but in a way that is backwards-compatible, of course).

Cheers!

glawton

Problem Kludged:

I changed the variable to go, and now the code exhibits the intended conditional execution behavior.

if self.get("avg_rt") < 2000:

        exp.set("go","yes")

else:

        exp.set("go", "no")

Thanks for the support!