Calibrating Attitude Scale with Negatively Worded Items Using PARELLA and Rating Scale Models [electronic resource] / Peter Lam and Yoke-Yeen Foong.

An important principle in constructing rating scales is to develop items that reflect various degrees of the "pro" (positive) and "contra" (negative) aspects of the trait being measured. Where both positive and negative items are pooled, they can be arranged in order along the tr...

Full description

Saved in:
Bibliographic Details
Online Access: Full Text (via ERIC)
Main Author: Lam, Peter
Other Authors: Foong, Yoke-Yeen
Format: Electronic eBook
Language:English
Published: [S.l.] : Distributed by ERIC Clearinghouse, 1996.
Subjects:

MARC

LEADER 00000nam a22000002u 4500
001 b6395516
003 CoU
005 20080220152052.0
006 m d f
007 cr un
008 960101s1996 xx |||| ot ||| | eng d
035 |a (ERIC)ed398270 
040 |a ericd  |c ericd  |d MvI 
099 |f ERIC DOC #  |a ED398270 
099 |f ERIC DOC #  |a ED398270 
100 1 |a Lam, Peter. 
245 1 0 |a Calibrating Attitude Scale with Negatively Worded Items Using PARELLA and Rating Scale Models  |h [electronic resource] /  |c Peter Lam and Yoke-Yeen Foong. 
260 |a [S.l.] :  |b Distributed by ERIC Clearinghouse,  |c 1996. 
300 |a 20 p. 
500 |a ERIC Document Number: ED398270. 
500 |a ERIC Note: Paper presented at the Annual Meeting of the American Educational Research Association (New York, NY, April 8-12, 1996).  |5 ericd. 
520 |a An important principle in constructing rating scales is to develop items that reflect various degrees of the "pro" (positive) and "contra" (negative) aspects of the trait being measured. Where both positive and negative items are pooled, they can be arranged in order along the trait continuum, but for classical and item response theory analysis, scores for negatively worded items will have to be reversed in the Likert tradition. The data for this study of calibrating item scales came from 350 first-year teacher interns after their first internship who responded to 10 career commitment statements based on a 5-point Likert scale. The dataset was fitted into the Rating Scale Model (D. Andrich, 1978), and after dichotomizing, fitted into the PARELLA model (H. Hoijtink, 1991). Results showed that the Rating Scale model was able to arrange items 7 through 10 of the scale in the order as intended, but there were problems with the negatively worded items. The PARELLA model, however, was able to align the items correctly. On the other hand, the step estimates from the Rating Scale model give additional information on the way in which respondents indicate their level of agreement with the statements. Although the binary conversion of the Likert scale may result in a loss of information, the PARELLA model can provide information on item location and person separation of the trait. It is recommended that separate analysis of positive and negative items on Likert scales can be made using the Rating Scale model with the PARELLA model as a complement in establishing proximity items with item scale order. (Contains 4 tables and 12 references.) (Author/SLD) 
650 1 7 |a Attitude Measures.  |2 ericd. 
650 0 7 |a Foreign Countries.  |2 ericd. 
650 0 7 |a Internship Programs.  |2 ericd. 
650 0 7 |a Item Response Theory.  |2 ericd. 
650 1 7 |a Likert Scales.  |2 ericd. 
650 1 7 |a Rating Scales.  |2 ericd. 
650 0 7 |a Scaling.  |2 ericd. 
650 1 7 |a Student Teachers.  |2 ericd. 
650 0 7 |a Test Construction.  |2 ericd. 
650 1 7 |a Test Items.  |2 ericd. 
650 0 7 |a Test Theory.  |2 ericd. 
700 1 |a Foong, Yoke-Yeen. 
856 4 0 |u http://files.eric.ed.gov/fulltext/ED398270.pdf  |z Full Text (via ERIC) 
907 |a .b63955167  |b 07-06-22  |c 10-16-10 
998 |a web  |b 10-26-12  |c f  |d m   |e -  |f eng  |g xx   |h 0  |i 1 
956 |a ERIC 
999 f f |i b21fc005-f3ce-54d4-803d-e5549583d090  |s a7ae4d30-e758-5148-964b-4cd8eeccaee7 
952 f f |p Can circulate  |a University of Colorado Boulder  |b Online  |c Online  |d Online  |e ED398270  |h Other scheme  |i web  |n 1