how to present each case in a case structure a certain percentage of the iterations

[psloan21]

I am programming a task for use in a psychology experiment. There are two options for the stimulus, the test and the catch stimuli. I want to present the test stimulus on 90% of the trials and the catch stimulus on 10% of the trials. Right now, I have the test and catch stimulus as two cases in a case structure and I have that case structure in a for loop so I can control the number of trials. I cannot figure out how to make the test stimulus run on 90% of the iterations. 

[tst · Accepted solution]

Ignoring any functions for randomizing data which probably exist in LV, here are a couple of options:

1. The "sort of" one - you don't know how many iterations you will have and you want to choose between two options at a probablity of 10% and 90%. Take the random number generator and check if the result is larger than 0.9. It will be 10% of the time (at least close enough for something involving human interaction, which presumably has a limited number of runs).
2. You know how many iterations you have and you want an exact number - generate an array of booleans of size N initialized to F. Divide N by 10 to get your 10 percent and then use a shift register to replace specific elements in the array. The elements you replace should be a random number multiplied by N and rounded and you should also have a while loop in the for loop to regenerate a number if there is one you already picked.

[jcarmody]

3. Generate an array with the needed quantity of T and F and run it through Riffle.vi

[psloan21]

Where is the Riffle VI? 

[crossrulz]

Signal Processing->Signal Operation

[GerdW]

Hi psloan,

you could use Quickdrop to find it - or search the LabVIEW help for "Riffle"…

[jcarmody]

You can't use Riffle to riffle a boolean array after all, but you can pretend.  This doesn't provide fractional True values, so if you wanted 45% True you'd actually get 50%. 😛

[crossrulz]

We are making things too difficult here.  Let's pull out a trick from the bag of Altenbach.  Use the Bernoulli Noise VI (found in the Signal Processing->Signal Generation palette).

[jcarmody]

Maybe the probability of a "1" is valid for each element, but I don't think this is what the OP wanted.  This example should have five Trues, no?  I couldn't get it to fill it with 15 Trues, but 14 happened pretty quickly.

[altenbach]

---

@crossrulz wrote:

We are making things too difficult here.  Let's pull out a trick from the bag of Altenbach.  Use the Bernoulli Noise VI (found in the Signal Processing->Signal Generation palette).

---

I was just about to post that... 😄