Scores from the IPT Reading and Writing Tests
The questions on the IPT Reading test are grouped into tasks that assess different aspects of reading: vocabulary both in isolation and in the context of sentences; literary and informational texts; and language usage. All students attempt all questions. The scores from the IPT Reading tests are derived from raw scores (number correct scores), and are reported in the same standardized score units as the IPT Oral scores: scaled scores, percentile ranks, and normal curve equivalents. They are also reported as designations and proficiency levels. Since all students attempt all questions in the IPT Reading test, the relationship between all the score points with this test is direct.
The IPT Writing test is composed of one multiple-choice part that evaluates writing conventions, and two productive writing parts. The scores from the productive writing tasks are weighted when calculating all the reported scores besides the raw score, so that productive writing receives 2/3 weight in the scores. As with the other skills, the reported scores are scaled scores, percentile ranks, normal curve equivalents, designations, and proficiency levels. Each of the scores is described briefly below. View the IPT Reading and Writing designation tables here, and the complete score conversion tables for the IPT Reading and Writing tests here.
Raw Scores
A raw score is simply a tally of a participant’s correct answers on the test. In the IPT Reading test, this has a straightforward interpretation since all students attempt all items and all items are weighted equally. In the IPT Writing test, the raw score is a sum of the number-correct score from the multiple choice part and the points earned from ratings of the student’s productive writing. However, because the students’ productive writing scores are weighted when deriving all the reported scores besides the raw score, the conversion tables look a little more complex than they do for Reading. To convert scores by hand, you need to know the student’s scores in each Writing test part. Score conversion is automatic within the OnlineIPT.
Scaled Scores
Scaled scores are on a common scale for the IPT Reading and Writing Tests. A common scale is useful for making comparisons across different forms and levels. Once a raw score has been converted to a scaled score, it may be compared to scaled scores from a different form (for example, a scaled score from Form 2C may be compared to a scaled score on Form 2D). Additionally, a scaled score may be compared to scaled scores from other levels, such that changes in performance can be studied across time. The scaled scores are on a continuous vertical scale that links together the different proficiency levels measured by the IPT (Levels 1, 2 and 3) and allows growth to be determined across levels. The scaled scores were derived using the Rasch model.
Percentile Ranks
The percentile rank of a score indicates the percentage of scores from the norming sample that is equal to or less than that score. Percentiles are best for obtaining a straightforward indicator of the relative performance of a student. Percentiles must be interpreted relative to the group from which they were derived. For the IPT, the percentiles represent the rankings of the students in their grade or narrow grade band in the latest norming study. As a result, most of the conversion tables for the IPT Reading and Writing tests are grade-specific. It is important to check that the student’s grade matches the title of the table you are using to convert their scores.
Normal curve equivalents
Normal curve equivalents (NCEs) are derived from percentile ranks and are typically used for research. NCEs resemble percentile ranks in that they range from 1 to 99 and have a mean of 50, but unlike percentile ranks, they represent an equal interval scale. This means that, for example, a change of 10 points represents the same change in ability regardless of where it occurs on the scale. NCEs are suitable for computing averages and difference scores, neither of which are meaningful when computed using percentile ranks. NCEs can be useful for averaging a group’s scores or when computing gains made on the same test over time.
Designations
The designation categories are called Non-, Limited, and Competent Reading or Writing (depending on test). The cut scores for the designations are grade specific, which means that as students advance in grade, more language skills are expected of them to achieve a Limited or a Fluent designation than from students in lower grades. The derivation of the cuts for the designation categories scores is both criterion-referenced and norm-referenced. The criterion-referenced criteria are of two types: comparison with teacher opinion, and comparison against monolingual and proficient students’ performance on the test. The norm-referenced element informing the designation categories derives from the score distribution of the learner group on the test. This is reviewed at every renorming and publication of new test forms in order to maintain continuity in score interpretation.
Proficiency Levels
The Proficiency Levels are Beginning, Early Intermediate, Intermediate, Early Advanced, and Advanced, and they correspond to the proficiency levels used in Ballard & Tighe’s instructional product lines. These score categories were developed to provide greater detail on students’ oral proficiency within the “Limited” designation: The Beginning level corresponds to the Non- designation category, the Early Intermediate, Intermediate, and Early Advanced levels correspond to the Limited designation, and the Advanced level corresponds to the Competent designation. The cuts are grade-specific just like the designation categories. As described above, the cut for the Competent designation is derived in a criterion-referenced way from association with teacher evaluation and from the performance of monolingual, proficient students on the test. The Advanced cut is therefore strongly criterion-referenced. The division of the Limited category into the proficiency levels is based on the Limited learners’ frequency distribution across the scores in the Limited designation range, however, so the derivation of the middle proficiency levels is based on a mixture of norm- and criterion-referencing. The Beginning proficiency level corresponds to the Non- designation, and is again strongly criterion-referenced. The conversion from raw and scaled scores to proficiency levels are included in both sets of score conversion charts linked above.
