liontamer
(1000+ posts)
Send PM |
Profile |
Ignore
|
Fri Nov-19-04 01:44 PM
Original message |
Any tips on analyzing reaction time data? |
|
Specifically algorithms for removing outliers?
Keep in mind that I can only code in Matlab :)
|
Dogmudgeon
(1000+ posts)
Send PM |
Profile |
Ignore
|
Fri Nov-19-04 01:51 PM
Response to Original message |
|
As a normal curve.
In your data section, you generally only have to specify the mean and the 1SD. At that point, skew and kurtosis would allow you to study the non-normal parameters (the "outliers").
You can also then just count 1SD or 2SD in either direction, as you find useful.
Sorry, but I have no algorithms. I learned the math in the late 1970s and promptly forgot how to do it.
--bkl
|
liontamer
(1000+ posts)
Send PM |
Profile |
Ignore
|
Fri Nov-19-04 02:13 PM
Response to Reply #1 |
2. thanks, that was along the lines of what I was looking for |
|
but I was hoping for some code to boot. alas I guess I can't expect the web to do my work for me :)
|
LibLabUK
(1000+ posts)
Send PM |
Profile |
Ignore
|
Fri Nov-19-04 02:36 PM
Response to Reply #2 |
|
You can do it in Excel, you use the standard deviation function (stdev) and then the normal distritbution function (normdist), then plot them on a graph.
|
DU
AdBot (1000+ posts) |
Thu Apr 25th 2024, 10:59 PM
Response to Original message |