A Study of Criteria Used in Equating

Harris, Deborah J.; Crouse, Jill D.

doi:10.1207/s15324818ame0603_3

Cited by 74 publications

(75 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this study, the identity method was concluded to be preferable only when test forms were nearly equivalent in difficulty. This finding is consistent with the results of the studies of Harris and Crouse (1993), Kolen and Brennan (2004), and Babcock et al (2012). Since the RMSE for the identity method increases rapidly with an increase in difference in test form difficulty, test forms should be equated.…”

Section: Evaluating the Main Effect Of Average Difference In Form Difsupporting

confidence: 93%

“…This finding is similar to the findings from Babcock et al (2012). Harris and Crouse (1993) and Kolen and Brennan (2004) pointed out that not equating the forms (identity method) produced a smaller equating error as the difference between form difficulty lessened. Similarly, Babcock et al (2012) also remarked that when the differences in form difficulty were similar, not equating the forms produced more accurate results.…”

Section: Evaluating the Main Effect Of Average Difference In Form Difmentioning

confidence: 99%

See 1 more Smart Citation

Investigating Test Equating Methods in Small Samples Through Various Factors

2016

EDUC SCI-THEOR PRACT

View full text Add to dashboard Cite

In this study, equating methods for random group design using small samples through factors such as sample size, difference in difficulty between forms, and guessing parameter was aimed for comparison.Moreover, which method gives better results under which conditions was also investigated. In this study, 5,000 dichotomous simulated data consistent with the three-parameter logistic model were produced for each form (X and Y) by manipulating the factors' levels for random group design. In order to equate two test forms, the random groups design was used in this study. Simulated test forms were equated using the equating methods of identity, mean, linear, circle-arc, and 2-and 3-moments pre-smoothed equipercentile for different sample sizes (10, 25, 50, 75, 100, 150, 200) through 100 replications. The results obtained from this simulation study were evaluated based on the criterion of equating error (RMSE). The findings indicated that in the case where the sample size was 50 or more and the difficulty difference between forms was 0.4, equating the forms was concluded to give better results than not equating. Moreover, the circle-arc and mean-equating methods produced lower equating errors than other equating methods for small-sample equating under most of the conditions studied.

show abstract

Section: Evaluating the Main Effect Of Average Difference In Form Difsupporting

confidence: 93%

Section: Evaluating the Main Effect Of Average Difference In Form Difmentioning

confidence: 99%

Investigating Test Equating Methods in Small Samples Through Various Factors

2016

EDUC SCI-THEOR PRACT

View full text Add to dashboard Cite

show abstract

“…To answer this question, equating results should be evaluated. Harris and Crouse (1993) provided an extensive review of equating criteria used in equating studies. Lord's (1980) equity property is one of criteria which have been used to evaluate equating results.…”

Section: Evaluating Equating Resultsmentioning

confidence: 99%

“…According to Lord's equity property (Kolen & Brennan, 2004, p. 10): …It must be a matter of indifference to each examinee whether Form X or Form Y is administered… In other words, if for examinees with a given true score, the distribution of equated scores on the new form is identical to the distribution of the scores on the old form, then Lord's equity property holds. This is not possible unless two forms are strictly parallel, in which case equating is unnecessary (Brennan, 2010;Harris, 1993). For this reason, Morris (1982) suggested a less strict definition of equity, which is called weak equity or first-order equity.…”

Section: Evaluating Equating Resultsmentioning

confidence: 99%

“…Evaluating equating results with using equity criteria is important because it is directly related with special case of Lord's equity definition of equating (Harris, 1993). Equity criteria allows to assess the degree to which a given examinee is advantaged or disadvantaged by being administered alternative test forms (Bolt, 1999).…”

Section: Evaluating Equating Resultsmentioning

confidence: 99%

See 1 more Smart Citation

The Impact of Test Dimensionality, Common-Item Set Format, and Scale Linking Methods on Mixed-Format Test Equating

2016

EDUC SCI-THEOR PRACT

View full text Add to dashboard Cite

The purpose of this study was to examine the impact of dimensionality, common-item set format, and different scale linking methods on preserving equity property with mixed-format test equating. Item response theory (IRT) true-score equating (TSE) and IRT observed-score equating (OSE) methods were used under common-item nonequivalent groups design. The equity property was evaluated based on firstorder equity (FOE) and second-order equity (SOE) properties. A simulation study was conducted based on actual item parameter estimates obtained from the TIMSS 2011 8th grade mathematics assessment.The results showed that: (i) The FOE and SOE properties were best preserved under the unidimensional condition, were poorly preserved when the degree of multidimensionality was severe. (ii) The TSE and OSE results, which were provided by using a mixed-format common-item set, preserved FOE better compared to equating results, which provided only a multiple-choice common item set. (iii) Under the unidimensional and multidimensional test structure, characteristic curve methods performed significantly better than moment scale linking methods in terms of preserving FOE and SOE properties.

show abstract

IRT Linking and Equating

Lee¹,

Lee²

2018

The Wiley Handbook of Psychometric Testing

View full text Add to dashboard Cite

A Study of Criteria Used in Equating

Cited by 74 publications

References 8 publications

Investigating Test Equating Methods in Small Samples Through Various Factors

Investigating Test Equating Methods in Small Samples Through Various Factors

The Impact of Test Dimensionality, Common-Item Set Format, and Scale Linking Methods on Mixed-Format Test Equating

IRT Linking and Equating

Contact Info

Product

Resources

About