Skip-To + Responses.xls combination occasionally failing to work (case 3235)

I'm having an interesting problem combining skipto logic with the responses.xls file.

The general flow of this stage of my experiment is as follows.

[FONT=Calibri]1)[/FONT]Participant sees a word pair and is asked to guess what one of them is and types in his/her response.

[FONT=Calibri]2) [/FONT]If his/her response is correct, the response.xls file updates to forward them to the “correct response screen” ("none" as the skipto option since it's right after the response) or to the wrong response screen +1 since the wrong response screen is after the correct response screen) which gives them feedback on what the correct response is. In both screens participants are asked to indicate if they remember seeing that pair before.

[FONT=Calibri]3) [/FONT]The correct response screen then forwards them past the wrong response item to the confidence rating screen (using a +1 skipto option). If they had made a wrong response they get forwarded to that screen anyway given the item sequence.

Now the problem is that, on a few occasions, when a correct response is made and participants are forwarded to the correct response screen, after responding the skipto does not kick in (I'm assuming this is the skipto from the experiment file and not the responses.xls file though I could be wrong) and participants are still shown the wrong response screen before making their judgments. There isn't a particular pattern to this (i.e. on some trials items 2 & 140 fail whereas on other trials it's items 4 & 50).

What could possibly be the reason for this? I've attached this portion of my experiment in a zip file along with data from two runs with the problem trials highlighted.

Best

John

Just to outline steps I've tried :

1) Using specific skip-tos doesn't work either

2) Putting all the "wrong feedback items" at the end of the questionaire and creating a logic structure that skips to them when participants get the answer wrong and then back up to the confidence rating items also doesn't seem to work.

And this method works fine for the earlier stage of the experiment where the only difference is that instead of being multiple choice items, both the wrong and right screens are of the instruction type. Unfortunately in this stage, a yes no response is required of the participant so I can't use the same type here.

Ah well...I do need to get data collection going. I found an alternative way to get this to sorta kinda work for my paradigm which will just have to do.

Thread closed.