Random Without Replacement Not Behaving Correctly


Author
Message
AnthonyFCollinsSussex
AnthonyFCollinsSussex
Associate Member (100 reputation)Associate Member (100 reputation)Associate Member (100 reputation)Associate Member (100 reputation)Associate Member (100 reputation)Associate Member (100 reputation)Associate Member (100 reputation)Associate Member (100 reputation)Associate Member (100 reputation)
Group: Forum Members
Posts: 39, Visits: 183
Hello

I thought I would flag this here as I do think it is an important issue, and one I am struggling to resolve now for several weeks. I am unsure whether this is an inherent bug or not.

I have 4 conditions. So, for each set (or 'array') of 4 participants, the condition order should be randomised:

<batch>
/ subjects = (1 of 4)
/ groupassignment = groupnumber
/ file = "Stroop.iqx"
/ file = "Recognition Task A.iqx"
/ file = "Recognition Task B.iqx"
/ file = "N Back.iqx"
</batch>

<batch>
/ subjects = (2 of 4)
/ groupassignment = groupnumber
/ file = "Stroop.iqx"
/ file = "Recognition Task B.iqx"
/ file = "Recognition Task A.iqx"
/ file = "N back.iqx"
</batch>

<batch>
/ subjects = (3 of 4)
/ groupassignment = groupnumber
/ file = "N Back.iqx"
/ file = "Recognition Task A.iqx"
/ file = "Recognition Task B.iqx"
/ file = "Stroop.iqx"
</batch>

<batch>
/ subjects = (4 of 4)
/ groupassignment = groupnumber
/ file = "N Back.iqx"
/ file = "Recognition Task B.iqx"
/ file = "Recognition Task A.iqx"
/ file = "Stroop.iqx"
</batch>

<defaults>
/ screencolor = white
/ txcolor = black
/ txbgcolor = white
/ fontstyle = ("Arial", 2.8%)
</defaults>

<data>
/ separatefiles = true
</data>

Moreover, I have set the Participant & Group ID webpage registration to random selection without replacement. Similarly, number of conditions is set to 4. Yet, after a series of test runs (documenting groupid), this is what each set of 4 looks like:

1, 3, 2, 4

2, 3, 1, 4

3, 1, 2, 4

1, 3, 2, 4

1, 2, 3, 4

1, 2, 3, 4

1, 2, 3, 4

1, 3, 2, 4

1, 2, 3, 4

3, 2, 1, 4

2, 1, 3, 4

3, 1, 2, 4

1, 3, 2, 4

1, 2, 3, 4

1, 2, 3, 4

1, 2, 3, 4

1, 3, 2, 4

1, 2, 3, 4

3, 2, 1, 4

2, 1, 3, 4

As you can see, condition 4 is never allocated anywhere else other than the forth and final position. This is systematic. It would be nice to have a fully randomised ordering of all 4 conditions as (what should be) implied by the code - it only seems to randomize the first three conditions, and this is very bizarre. Is there something I am missing here? Or failing to see? I have tried using subjectID and random in the groupassignment in batch, but it seems not to make any difference - condition 4 always seems to fall last for every set of 4. (though 'random' groupassignment ironically produces a sequence of 1,2,3,4).

Just so you know, these tests were conducted ensuring that there is no extraneous refresh of the start-up screen, nor re-visiting the start up-screen, to rule this out as a potential factor.

Many Thanks & Yours Sincerely

GO

Merge Selected

Merge into selected topic...



Merge into merge target...



Merge into a specific topic ID...




Reading This Topic

Explore
Messages
Mentions
Search