Millisecond Forums

Setting up separate a separate <trial> for each geometric form, each number, and each building seems inefficient.

You only need four <trial> elements:
(1)
<trial form>
/ stimulusframes = [1=formen]
...
</trial>

<picture formen>
/ items = ("form1.jpg", "form2.jpg", "form3.jpg", ...., "form10.jpg")
...
</picture>

(2)
<trial zahl>
/ stimulusframes = [1=zahlen]
...
</trial>

<text zahlen>
/ items = ("1", "2", "3", ...., "10")
...
</text>

(3)
<trial haus>
/ stimulusframes = [1=haeuser]
...
</trial>

<picture haeuser>
/ items = ("haus1.jpg", "haus2.jpg", "haus3.jpg", ...., "haus10.jpg")
...
</picture>

and

(4)
<trial SetA9Y1>
/ stimulusframes = [1=seta9y1pic]
...
</trial>

<picture seta9y1pic>
/ items = ("seta9y1.jpg")
...
</picture>

Then you can simply do:

<block primingSetA9Y1>
     /trials = [1-40=noreplacenorepeat(zahl, form, haus, SetA9Y1)]
    </block>

Hi Dave,

Thank you for your quick reply.
I have two more questions:

1. If one trial contains 10 items and another trial contains only one item and I randomise those, is it guaranteed that the single item of the latter trial is represented as often as each item of the first trial?

2. I always thaught that the randomization of trials in a block leads to a randomization across all trials, but not across the items. In the end, I want the items to be displayed like this for example: Form5, SetA9Y1, Haus7, SetA9Y1, From4, Zahl10, Haus3, Zahl2, SetA9Y1... I do not want a randimization only within the trials, which would look like this Form10, Form2, Form8..., Haus3, Haus9, Haus10...., Zahl2, Zahl5, Zahl1....etc. And that is what I would suspect to happen, if I use your suggestion. Am I wrong? And if not, how can I display the items the way I want them to be displayed?

Thank you very much!

M.

"1. If one trial contains 10 items and another trial contains only one item and I randomise those, is it guaranteed that the single item of the latter trial is represented as often as each item of the first trial?

2. I always thaught that the randomization of trials in a block leads to a randomization across all trials, but not across the items. In the end, I want the items to be displayed like this for example: Form5, SetA9Y1, Haus7, SetA9Y1, From4, Zahl10, Haus3, Zahl2, SetA9Y1... I do not want a randimization only within the trials, which would look like this Form10, Form2, Form8..., Haus3, Haus9, Haus10...., Zahl2, Zahl5, Zahl1....etc. And that is what I would suspect to happen, if I use your suggestion. Am I wrong? And if not, how can I display the items the way I want them to be displayed?"

You are mixing two things up.

<block primingSetA9Y1>
 /trials = [1-40=noreplacenorepeat(zahl, form, haus, SetA9Y1)]
</block>

The block's /trials attribute determines *how trial elements are sampled*. The above simply means "sample the 4 trial elements zahl, form, haus, and seta9y1 in *equal* proportions", randomly without replacement and avoiding repetitions ("noreplacenorepeat()").

I.e., you are guaranteed to end up with 10 zahl-trials, 10 form-trials, 10 haus-trials, and 10 seta9y1-trials.

How any stimuli *presented* by those trials sample *items* is an entirely different matter. That is determined by the respective *stimulus* element's /select attribute. The default for any stimulus element is to sample its items *randomly without replacement*:

If

<picture formen>
/ items = ("form1.jpg", "form2.jpg", "form3.jpg", ...., "form10.jpg")
/ select = noreplace
</picture>

as presented by

<trial form>
/ stimulusframes = [1=formen]
...
</trial>

has 10 items, and you run the trial 10 times in the block (as per its /trials attribute), each item will be presented once.

If there is only one item as in

<picture seta9y1pic>
/ items = ("seta9y1.jpg")
...
</picture>

as presented by

<trial SetA9Y1>
/ stimulusframes = [1=seta9y1pic]
...
</trial>

and you run that trial 10 times in the block, you'll see that item 10 times.

Given the above, you are indeed wrong regarding

"I do not want a randimization only within the trials, which would look like this Form10, Form2, Form8..., Haus3, Haus9, Haus10...., Zahl2, Zahl5, Zahl1....etc. And that is what I would suspect to happen, if I use your suggestion. Am I wrong?"

This is *not* what you'll get with my suggestion. You will get a random trial-sequence that will look like this

Trial:   Form    SetA9Y1   Haus    SetA9Y1  Form    Zahl      Haus    Zahl... (i.e., no trial-repetitions like Zahl - Zahl or SetA9Y1 - SetA9Y1)
Item:   Form5  SetA9Y1   Haus7 SetA9Y1  Form4  Zahl10 Haus3  Zahl2... (i.e., items sampled randomly w/o replacement)

where each instance of <trial Form> will display a different *item* of <picture formen> (sampled randomly without replacement), and so forth for the other <trial> elements and the respective stimulus elements.

Based on my reading of

"In the end, I want the items to be displayed like this for example: Form5, SetA9Y1, Haus7, SetA9Y1, From4, Zahl10, Haus3, Zahl2, SetA9Y1... "

this should give you exactly what you want.

Thanks a lot!!! 
You are absolutely right, I was a bit slow there.. :)

Hi again,

now, I am facing the problem the other way around:

I have two trials that I want to be displayed randomly in one block. This time I do not want the items to be displayed equally.
I need the items of the first trial to be randomly displayed 18 times and in-between, I need the stimuli of the second trial to be randomly displayed three times in total. How can I do that?

Thank you in advance!
M.

You enter the trials into the selection pool in the desired quantities (this is covered in the documentation for the /trials attribute).

"I need the items of the first trial to be randomly displayed 18 times and in-between, I need the stimuli of the second trial to be randomly displayed three times in total."

means you have a 6:1 ratio. Thus you do:

<block someblock>
/ trials = [1-21 = noreplace(a,a,a,a,a,a,b)]
</block>

This will end up giving you a total of 18 a-trials and 3 b-trials in random order.