info_func {lazy.irt} | R Documentation |

This function calculates the information functions associated with
the set of given weights,

and two types of locally optimam weights.

info_func( param, theta = NULL, weight = NULL, npoints = 31, thmin = -4, thmax = 4, thdist = "NORMAL", thmean = 0, thstd = 1, legend = 1, maxinfo = 0, numderiv = 0, smallP = 1e-09, print = 1, plot = 0, debug = 0 )

`param` |
Item Parameter Data Frame. |

`theta` |
Discrete theta values |

`weight` |
Weight data frame. |

`npoints` |
# of discrete points for theta. |

`thmin` |
Minimum value of discrete thata value. |

`thmax` |
Maximum value of discrete thata value. |

`thdist` |
= "NORMAL" or "UNIFORM" |

`thmean` |
mean of theta distribution |

`thstd` |
std of theta distribution |

`legend` |
= 0 to skip printing legend. |

`maxinfo` |
= The maximum value of item information function for plot. |

`numderiv` |
= 1 to use numerical derivatives. |

`smallP` |
Minimum value of probability in irf and dirf functions. |

`print` |
= 1 to print the summary |

`plot` |
= 1 to plot test information functions |

`debug` |
= 1 to print intemediate result. |

Note that, given category and item weights, information function is defined
as

( slope of TRF at theta )^2 / (variance of x at theta)

where TRF and x is calculated with the given set of weights.

In general. the best weights depends on the value of theta.
Therefore, the name locally best weight.

The best item weight given category weights is called as
locally best item weight, or LOW.

When the category weights themselves are optimized it is called as
the locally best weight, or, LO.

This weight is equivalent to the basic function of Samejima(1969) .

The information function with LO is defined as

* ∑_j ∑_k (P'_{kj}(θ))^2 / P_{kj}(θ) *

where *P_{kj}(θ)* is the item category response function,

and *P'_{kj}(θ)* is its derivative.

The information function with LOW is defined as

* ∑_j (P_j^{*'}(θ))^2 / var(U_j^{*} | θ) *

where *U_j^{*} = ∑_k v_{kj} U_{kj}* is
the weighted item score,

and *P_j^{*'}(θ)* is the derivative of the expected value of
*U_j^{*}* at theta.

The Basic Function is defined as * P`(theta) / P(theta) *.

The Best Item Category Weight = Basic Function - Basic Function
of the 0-th category.

Item Info is the sum of * P`(theta)^2 / P(theta) *.

A list of

theta: theta points

thd: theta distribution (to calculate average info)

info : information function defind as (slope_TRF)^2 / (stdx_t)^2

fromP, toP: location of each item category in item info

TRF: test response function (tcc)

TRF_LO: test response function (tcc) with locally best weight

icrf: item category response functions

dicrf: derivative of icrf

info_LOW: information function with locally best item weight

info_LO: information function with locally best category weight

info_item_LOW: item information function
with locally best item weight

info_item_LO: item information function
with locally best category weight

stdx_t: SEM at theta (conditional standard deviation of X given theta)

w_LO: the locally best item weight

v_LO: the locally best item category weight
scaled so that v[0,j]=0, j=1,2,...,nitems.

A: The Basic Function

a_stdx_t: square root of the average of stdx_t^2

a_info: average of info

a_info_LOW: average of info_LOW

a_info_LO: average of info_LO

a_info_item_LOW: average of info_item_LOW (for each item)

a_info_item_LO: average of info_item_LO (for each item)

a_w_LO: average of w_LO (for each item)

a_v_LO: average of v_LO (for each item category)

Birnbaum, A.(1968) Some Latent Traint Models.
In F. M. Lord and M. R. Novick, Statistical Theories of Mental Test Scores.
Reading, Mass.: Addison-Wesley.

Samejima, F. (1969). Estimation of a latent ability using a response pattern of graded scores. Psychometrika Monographs, 34 (Suppl. 4).

resInfo <- info_func( paramS2, plot=2, print=1 ) resInfo <- info_func( paramS2, weight=weightS21, plot=1, print=0 ) resInfo <- info_func( paramS2, weight=weightS22, plot=1, print=0 )

[Package *lazy.irt* version 0.1.3 Index]