/****************************************************************/
/* STUIRT */
/* A Computer Program for Scale Transformation under */
/* Unidimensional Item Response Theory Models */
/* Version 1.0 */
/* */
/* Seonghoon Kim and Michael J. Kolen */
/* */
/* Iowa Testing Programs */
/* University of Iowa */
/****************************************************************/
THE FOLLOWINGS WERE READ FROM THE INPUT FILE:
< Test_2.stu >
OK 55
66 Z9 2 MC 3.40100 3.40100 0.5502 4.0322
67 Z9 2 MC 3.40100 3.40100 0.6070 4.0322
68 Z9 2 MC 3.40100 3.40100 0.8946 4.0322
69 Z9 2 MC 3.40100 3.40100 -2.4236 4.0322
70 Z9 2 MC 3.40100 3.40100 0.2261 4.0322
SK 91
66 Z9 2 MC 3.40100 3.40100 0.6690 4.0322
67 Z9 2 MC 3.40100 3.40100 0.3567 4.0322
68 Z9 2 MC 3.40100 3.40100 0.0184 4.0322
69 Z9 2 MC 3.40100 3.40100 -1.2204 4.0322
70 Z9 2 MC 3.40100 3.40100 0.9993 4.0322
CI 55 BK
OP
ND 41 SN 4.0322 3.40100 4.0000
OD 41 SN 4.0322 3.40100 4.0000
KO SL
BY
===[OPTIONS AND DEFAULTS]===
1. ST: [DEFAULT] slope = 6, intercept = 5
2. IT: [DEFAULT] maximum number of iterations =15
3. OD: number of theta values = 5
** OLD FORM: THETA VALUES AND WEIGHTS] **
#0001: -8.10484E+000 1.58347E-005
#0002: -6.94837E+000 3.14155E-005
#0003: -7.38954E+000 2.01854E-006
#0004: -9.48437E+000 5.48388E-006
#0005: -7.38474E+000 3.84373E-006
4. ND: number of theta values = 5
** NEW FORM: THETA VALUES AND WEIGHTS] **
#0001: -8.10484E+000 1.58347E-005
#0002: -6.94837E+000 3.14155E-005
#0003: -7.38954E+000 2.01854E-006
#0004: -9.48437E+000 5.48388E-006
#0005: -7.38474E+000 3.84373E-006
5. FS: [DEFAULT]
HAEBARA METHOD: do standardize loss functions
STOCKING & LORD METHOD: do standardize loss functions
6. SY: [DEFAULT]
HAEBARA METHOD: bidirection (f1 and f2)
STOCKING & LORD METHOD: bidirection (f1 and f2)
7. LM: [DEFAULT] NONE
8. OP: ** This option works here. *
9. KO: The Stocking-Lord method is applied to obtain a common scale.
Input files for Kolen's POLYEQUATE are provided.
===[Summary Statistics by Model]===
The three-parameter logistic (3PL) model:
Attempt1 Attempt2
-------------------------- --------------------------
a_j b_j c_j a_j b_j c_j
N.Obs. 10 10 10 10 10 10
Mean 6.77700 -7.6336 2.0345 2.57800 -9.1536 3.0322
Std.Dev. 4.0322 0.54568 4.0322 4.0322 0.9232 4.0322
===[Solutions for Moment Methods]===
*** When including the NR and/or MC models,
Attempt1 Attempt2
---------------- ----------------
a_jk b_jk a_jk b_jk
N.Obs. 10 10 10 10
Mean 3.40100 -6.5384 3.40100 -0.1536
Std.Dev. 4.0322 0.00011 4.0322 00003
Method Slope Intercept
------------------------------------
Mean/Mean 2.8375903 -3.333333
Mean/Sigma 4.847654 -1.111139
*** When excluding the NR and/or MC models,
Attempt1 Attempt2
---------------- ----------------
a_jk b_jk a_jk b_jk
N.Obs. 10 10 10 10
Mean 3.40100 -6.5384 3.40100 -0.1536
Std.Dev. 4.0322 0.00011 4.0322 00003
Method Slope Intercept
------------------------------------
Mean/Mean 2.8375903 -3.333333
Mean/Sigma 4.847654 -1.111139
===[First Solutions for Characteristic Curve Methods]===
Numb. Meth. 1st.S 1st.I Slope Intercept Iter# TermC Grad.S Grad.I Func.V
0000 HA 1.0 0.0 0.000087 -2.843657 2 1 -9.736328 4.032200 4.2847810
SL 1.0 0.0 2.897536 -18352874 3 1 -3.184737 4.032200 0.670331
===[Final Summary of Solutions by Method]===
Method Slope Intercept
------------------------------------
Mean/Mean 5.5555999 -1.5839547
Mean/Sigma 4.443838 -2.863526
Haebara 3.141414 -1.987654
Stocking-Lord 3.847377 -9.999999
As the documents in question contain confidential data, I manually changed the values. There is quite a bit more to the document that would take too much time to anonymize.
But I need a code that can pull out the tables and data, but ignore the header and options.
My first thought was that if i could get the code to scan the document for a certain string. (i.e. ** OLD FORM: THETA VALUES AND WEIGHTS] **, or ===[Solutions for Moment Methods]===) and then import the following table. Since the documents should all follow that similar layout and format.
Any help greatly appreciated. I'm quite stumped at the moment.
