# Applying the Weibull equation to fit data to a psychometric curve

#### prud49

##### New Member
Hello,

I would be very grateful if someone could tell me if my application of the Weibull function to my data is correct or not. (The question is very simple, but requires a bit of background information, so please bear with me!)

- I am conducting a psychometric study on individual differences in predictive salience perception, and have a group of 9 subjects. It is a within-subject design, with 2 conditions: a passive condition, and an active one.

- Participants hear vocal excerpts of vowels /a/ and /o/, that have each been morphed (with another person's voice) to a certain degree (increments of 10%), from 0% morphed to 100% morphed. Both conditions require that the participants indicate whether the stimulus sounds "more like them" or "more like other" (binary choice). So for each vowel (/a/ and /o/) there exists 11 morphed versions (0%, 10%, 20%, 30%, .... , 100%).

- In the passive block, participants hear the sound and are asked to respond. In the active block, a visual cue appears instructing participants to press a button to initiate the sound. After the sound, participants provide a response.

- Each of the 11 morphs of /a/ and the 11 of /o/ are presented 10 times to the participants (in a counterbalanced order), resulting in 110 stimuli in total for /a/ and 110 for /o/. So, in the passive block, 220 stimuli (of /a/ and /o/) are used. The same applies for the active block.

Concerning the data:

- I am plotting curves with the proportion of "other" responses plotted on the y-axis, and the degree of morph (0% to 100%) plotted on the x-axis. My goal is to end up with a neat psychometric curve that looks similar to this:

Unfortunately, the curves I am obtaining look nothing like this.

I have tried using the Weibull equation to fit the data to a psychometric function. I used the Microsoft Excel formula (Weibull.dist). I would like to know if I have plugged in the correct data. Please find an example below (using fake data). The following example concerns only the /a/ vowel in the active condition for participant X:

When I plot the curve using the now Weibull-transformed data, I end up with something that looks much more like a psychometric function. I just want to be sure that I have plugged the correct values in the equation (the general equation is =WEIBULL.DIST(x, alpha, beta, cumulative)).

Many thanks in advance for taking the time to answer my question!