| Literature DB >> 25812787 |
Jong Cook Park1, Kwang Sig Kim2.
Abstract
PURPOSE: The reliability of test is determined by each items' characteristics. Item analysis is achieved by classical test theory and item response theory. The purpose of the study was to compare the discrimination indices with item response theory using the Rasch model.Entities:
Keywords: Discrimination; Fit statistics; Item response theory; Rasch model
Year: 2012 PMID: 25812787 PMCID: PMC8814536 DOI: 10.3946/kjme.2012.24.1.15
Source DB: PubMed Journal: Korean J Med Educ ISSN: 2005-727X
Fig. 1.Scatter Plot Point‐Biserial Correlation (Cpbs) vs. Other Discrimination Indices
Regression Equation between Point.Biserial Correlation (Cpbs) and Other Discrimination Indices
| Discrimination index | Equation of linear regression | r2 | SE | F‐statistic | Degree of freedom | SSreg | SSR |
|---|---|---|---|---|---|---|---|
| Ccit | y1=1.02×Cpbs | 1.00 | 0.00 | 3.74×10+30 | 23 | 0.74 | 0.00 |
| Cit | y3=1.04×Cpbs+0.10 | 0.99 | 0.02 | 351.20 | 23 | 0.87 | 0.06 |
| Cbs | y2=1.11×Cpbs+0.08 | 0.94 | 0.05 | 2,643.88 | 23 | 0.77 | 0.01 |
| D | y4=1.04×Cpbs+0.10 | 0.45 | 0.16 | 18.79 | 23 | 0.47 | 0.58 |
Ccit: Corrected item.total correlation, Cit: Item.total correlation, Cbs: Biserial correlation, D: Method of extreme group, SE: Standard error, SSreg: Sum of squared regressions, SSR: Sum of squared residuals.
Examinee Characteristics of the Rasch Model and Fit Statistics
| Examinee‘s ID | Score | Ability logit | SE | Outfit | Infit |
|---|---|---|---|---|---|
| S01 | 15 | 0.42 | 0.47 | 1.55 | 1.46 |
| S02 | 21 | 2.13 | 0.64 | 0.27 | 0.52 |
| S03 | 14 | 0.20 | 0.46 | 1.23 | 1.27 |
| S04 | 23 | 3.19 | 0.85 | 0.14 | 0.47 |
| S05 | 14 | 0.20 | 0.46 | 1.59 | 1.36 |
| S06 | 13 | -0.01 | 0.46 | 0.82 | 0.84 |
| S07 | 21 | 2.13 | 0.64 | 1.46 | 1.13 |
| S08 | 13 | -0.01 | 0.46 | 0.72 | 0.82 |
| S09 | 14 | 0.20 | 0.46 | 1.23 | 1.23 |
| S10 | 13 | -0.01 | 0.46 | 0.73 | 0.82 |
| S11 | 14 | 0.20 | 0.46 | 0.80 | 0.91 |
| S12 | 13 | -0.01 | 0.46 | 1.17 | 1.18 |
| S13 | 15 | 0.42 | 0.47 | 0.91 | 0.91 |
| S14 | 14 | 0.20 | 0.46 | 0.83 | 0.91 |
| S15 | 16 | 0.65 | 0.48 | 0.62 | 0.73 |
| S16 | 18 | 1.15 | 0.52 | 1.09 | 1.04 |
| S17 | 19 | 1.43 | 0.55 | 0.69 | 0.91 |
| S18 | 18 | 1.15 | 0.52 | 1.72 | 1.13 |
| S19 | 12 | -0.21 | 0.45 | 0.91 | 0.93 |
| S20 | 13 | -0.01 | 0.46 | 1.12 | 1.12 |
| S21 | 16 | 0.65 | 0.48 | 0.68 | 0.83 |
| S22 | 21 | 2.13 | 0.64 | 0.85 | 1.08 |
| S23 | 14 | 0.20 | 0.46 | 0.83 | 0.88 |
| S24 | 17 | 0.89 | 0.50 | 0.77 | 0.87 |
| S25 | 15 | 0.42 | 0.47 | 0.89 | 0.98 |
| S26 | 17 | 0.89 | 0.50 | 1.55 | 1.40 |
| S27 | 16 | 0.65 | 0.48 | 0.80 | 0.95 |
| S28 | 10 | -0.62 | 0.45 | 0.94 | 1.06 |
| S29 | 14 | 0.20 | 0.46 | 0.67 | 0.73 |
| S30 | 13 | -0.01 | 0.46 | 1.38 | 1.06 |
| S31 | 14 | 0.20 | 0.46 | 0.94 | 0.96 |
| S32 | 1 | -3.69 | 1.03 | 64.97 | 1.14 |
| S33 | 19 | 1.43 | 0.55 | 0.60 | 0.78 |
| S34 | 12 | -0.21 | 0.45 | 0.80 | 0.91 |
SE: Standard error.
Misfit (fit statistics >1.3).
Individual Item Difficulty and Discrimination Based on Point.Biserial Correlation (Cpbs), Item Characteristics of the Rasch Model and Fit Statistics
| Item ID | Difficulty | Cpoint‐biserial | Difficulty logit | SE | Outfit | Infit |
|---|---|---|---|---|---|---|
| Item01 | 0.53 | 0.38 | 0.59 | 0.37 | 0.80 | 0.88 |
| Item02 | 0.76 | 0.38 | -0.70 | 0.44 | 0.73 | 0.91 |
| Item03 | 0.82 | 0.32 | -0.78 | 0.49 | 0.84 | 0.96 |
| Item04 | 0.79 | 0.27 | -0.74 | 0.46 | 0.87 | 1.00 |
| Item05 | 0.50 | 0.36 | -0.38 | 0.37 | 0.83 | 0.88 |
| Item06 | 0.74 | 0.28 | 0.34 | 0.42 | 0.85 | 0.98 |
| Item07 | 0.62 | 0.39 | -0.51 | 0.38 | 0.77 | 0.88 |
| Item08 | 0.15 | 0.10 | -0.07 | 0.52 | 1.21 | 1.14 |
| Item09 | 0.06 | -0.41 | -0.02 | 0.76 | 48.19 | 1.24 |
| Item10 | 0.76 | 0.10 | 0.30 | 0.44 | 1.16 | 1.12 |
| Item11 | 0.82 | 0.42 | 0.22 | 0.49 | 0.68 | 0.88 |
| Item12 | 0.56 | 0.23 | 0.56 | 0.37 | 1.01 | 1.01 |
| Item13 | 0.65 | 0.21 | 0.45 | 0.39 | 0.96 | 1.03 |
| Item14 | 0.76 | 0.32 | 0.30 | 0.44 | 0.78 | 0.96 |
| Item15 | 0.44 | 0.23 | -0.31 | 0.37 | 0.94 | 1.01 |
| Item16 | 0.38 | 0.36 | -0.26 | 0.38 | 0.81 | 0.88 |
| Item17 | 0.85 | 0.22 | 0.18 | 0.53 | 1.07 | 1.02 |
| Item18 | 0.68 | 0.15 | 0.41 | 0.39 | 1.02 | 1.08 |
| Item19 | 0.56 | 0.23 | 0.56 | 0.37 | 1.03 | 1.02 |
| Item20 | 0.76 | 0.40 | 0.30 | 0.44 | 0.72 | 0.89 |
| Item21 | 0.26 | 0.17 | -0.15 | 0.42 | 1.03 | 1.10 |
| Item22 | 0.32 | 0.11 | 0.80 | 0.40 | 1.10 | 1.11 |
| Item23 | 0.79 | 0.07 | 0.26 | 0.46 | 1.65 | 1.10 |
| Item24 | 0.85 | 0.42 | -0.82 | 0.53 | 0.67 | 0.88 |
| Item25 | 0.62 | 0.20 | -0.51 | 0.38 | 1.06 | 1.03 |
SE: Standard error.
Misfit (fit statistics >1.3).
Fig. 2.Scatter Plot the Raw Score vs. Examinee's Ability with Standard Error
The horizontal line is in logit and the vertical line is in percent.
Fig. 3.Map of Item Difficulty and Examinee's Ability on Rasch Transformed Logit Scale
Right and side shows item difficulties and left had side shows the distribution of the examinee's abilities