![]() ![]() ![]() The experiment difficulties were designed to have two easier difficulties that should be relatively easy do determine by most people, the middle three should be ambiguous (which is where we expect people's threshold to lie), and the top two should be difficult. We are looking to find the threshold in the middle, the 'hardest' point at which someone is reliably able to tell the difference between signals. We are running a signal detection paradigm, and based on previous research, the shape we are expecting to see is a traditional psychophysics curve - at the easy and hard difficulties, responses are relatively predictable (high accuracy at low difficulties, and approximately chance accuracy at high difficulties). You're right, and that was a clumsy choice of words on my part. (I flaired it as a homework question, but it's not really homework, I'm just trying to get it working so I can start fitting curves to my data for an experiment I'm running). Also, I've flipped the results to get the curve to go upwards, but the real data has a negative slope (low x-axis values are high (easy difficulty = near 100% accuracy) and high x-values are low (hard difficulty are about 50% chance), but would love to find a way to make it so the curve starts high and drops off as it goes further to the right. I really want it to be flat at the bottom and the top, and the bulk of the slope to be in the middle.Īlso, does anyone have any good resources for being able to learn the curve fitting tool? Any help would be greatly appreciated. When I click onto fit options, I have no idea what any of that means, so I've left it all as default. We only have 7 data points (the average accuracy over 7 difficulty levels), and I'm looking for a general function that can be applied to multiple participants with thresholds at different levels along the x-axis. I have tried using the equation f(x) = a/(1+exp(-b*x)), and I was hoping that the curve would look something like this. ![]() I'm a second year psychology student doing a placement within cognitive neuroscience, and while I'm pretty comfortable with data and statistics, this is all at a level that is just above where I'm at. She has suggested that I'm looking to get a sigmoid curve, or cumulative gaussian or something along those lines (psychophysics curves). I've had a play around in the curve-fitting tool, but can't seem to get anything to work properly in the way that I'd expect it to. Hi all, I've been using Matlab for about a month now (and am still incredibly amateur!) and my supervisor has asked me to have a go at fitting some of our test data to a curve. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |