com.itemanalysis.psychometrics.irt.estimation

Class StartingValues

java.lang.Object

com.itemanalysis.psychometrics.irt.estimation.StartingValues

public class StartingValues
extends java.lang.Object

Computes starting values for binary item response models. It begins with PROX estimates of item difficulty and person ability (theta). It then maximizes the marginal likelihood for each item using observed counts obtained from the PROX estimates. This class is based on the Start3PL.h class in Brad Hanson's ETIRM library.

Constructor Summary

Constructors

Constructor and Description
StartingValues(ItemResponseVector[] responseVector, ItemResponseModel[] irm) Default constructor takes an array of item response vectors and an array of item response models.

Method Summary

Modifier and Type	Method and Description
void	addEMStatusListener(EMStatusListener listener)
ItemResponseModel[]	computeStartingValues() The public method for computing start values.
void	fireEMStatusEvent(int iteration, double delta, double loglikelihood)
void	fireEMStatusEvent(java.lang.String message)
java.lang.String	printItemParameters()
void	removeEMStatusListener(EMStatusListener listener)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

StartingValues

public StartingValues(ItemResponseVector[] responseVector,
                      ItemResponseModel[] irm)

Default constructor takes an array of item response vectors and an array of item response models.

Parameters:

responseVector -

irm -

Method Detail

computeStartingValues

public ItemResponseModel[] computeStartingValues()

The public method for computing start values. Starting values for item discrimination are constrained to the interval 0.3 to 3.0. Lower asymptote (guessing) values are constrained to the interval 0.05 to 1.0. No constraints are applied to item difficulty values. This method first computes PROX estimates for item difficulty and person ability. It then finds the item parameters that maximize teh marginal likelihood using observed counts.

Returns:

printItemParameters

public java.lang.String printItemParameters()

addEMStatusListener

public void addEMStatusListener(EMStatusListener listener)

removeEMStatusListener

public void removeEMStatusListener(EMStatusListener listener)

fireEMStatusEvent

public void fireEMStatusEvent(java.lang.String message)

fireEMStatusEvent

public void fireEMStatusEvent(int iteration,
                              double delta,
                              double loglikelihood)