Hi Everybody,
I have just been introduced to R and have used the ltm package for
calibrating items under IRT.
I have been able to get the item parameters using the fit function under
different models.
I am using the 2PL model. I have got the parameters for some 300 items
from 15 different tests using the function fit.
Now I need to automatically generate tests such that they have atleast a
certain level of information.
Thus I have created a data frame with item id, and information at 20
different points/ability levels.
I have to pick questions from the calibrated item set.
Right now I am just picking up 30 questions at random from the bank and
adding the information at these 20 points.
But I also need to plot the information function.
Till now I have been using the plot function defined in the package ltm
but it needs an object which inherits from ltm.
However I don't know how to create this object as I don't have any
response pattern for the set of 30 questions randomly picked from the bank
as it hasn't been used in any test yet.
How can I do this?
Also how can I plot the ICC for any item if I just have the item
parameters, difficulty and discrimination but not the fit object?
Is there someway to plot the curve using the formula given under 2 PL IRT
model and the values for a and b?
Thanks and regards,
--
Megha
Student,
Chennai Mathematical Institute
Chennai (India)
