Performance. When looking at the performance data presented in Table 3, it should be noted that the passing rates for 1997 to 1999 are figured on the number of students achieving a score of at least 75% correct on the reading or math test. For eighth graders in 1996 (those graduating in 2000), the passing rate was only 70%. After 1996, the passing rate was raised for all students graduating after the year 2000. This difference in what constitutes a passing score may account for the drop in the passing rates of eighth grade students for both the math and the reading test between 1996 and 1997.

Table 3. Performance of 8th Graders 1996-99

Since 1997, the math test scores for all students have stayed about the same, with around 70% of the students passing the math test on the first try. The percentage of LEP students passing has not changed much either, increasing slightly from 21 percent in 1997 to 24 percent in 1999 (also see Figure 2). It should be noted, however, that the number of LEP students taking the test in eighth grade has increased dramatically over this time, from 876 students tested in math in 1997 to 1,890 students in 1999.

Figure 2. Percentage of 8th Graders Passing BSTs 1996-99

Like the percentage passing math, the percentage passing for the reading test dropped a little between 1996 and 1997. Since 1997, however, the percentage of all students passing the reading test has increased steadily from 59% to 74%. These gains have mirrored the gains of LEP students on the reading test. The passing rate for LEP eighth graders has risen from 8% in 1997 to 22% in 1999.

The percentage of LEP students passing is considerably lower on both the reading and the math tests compared to the percentage of all students who pass. The percentage of LEP students passing the math test has stayed fairly stable at around 25%. However, the percentage of LEP students passing the reading test is growing steadily. In 1999 for the first time, nearly the same percentage (22%-24%) of LEP 8th graders passed the reading test and the math test on their first attempt.

It is useful to look not only at the percent of students passing, but also at student scores in order to determine whether students are improving over time even if they are not passing the tests. Table 4 and Figure 3 show the mean percentages of items correct on the math BST over the four year period 1996-99. The analyses do not include the roughly 100 students with invalid test scores. The mean percentages correct for both LEP students and all students have remained relatively constant over the four years the test has been given (around 80% for all students and 57% for LEP students). The mean percentage of items correct for LEP students is consistently about 23% lower than that for all students on the math test.

Table 4. 1996-99 Mean BST Math Performance of 8th Graders

Figure 3. Mean Percentage of Math Items Correct

Table 5 and Figure 4 show similar data on the Basic Standards reading test. The mean percentage of items correct has increased slightly for all students in eighth grade over the four year period from 72% in 1996 to 81% in 1999. During the same period, the mean percentage of items correct on the reading test for LEP students increased from 46% in 1996 to 57% in 1999. Although LEP students typically start out with lower scores on the reading test, they are making larger amounts of gain than on the math test.

Table 5. 1996-99 Mean BST Reading Performance on 8th Graders

Figure 4. Mean Percentage of Reading Items Correct

Performance in grades 8 -11 on 1999 BSTs. Students who do not pass one or both of the BSTs have additional opportunities each year to take the tests, which they are required to pass in order to graduate from high school. In addition, new students moving into the state also need to pass the BSTs in order to be eligible to graduate from high school in Minnesota. For mainly these reasons, students beyond eighth grade also take the BSTs. Table 6 shows the results of all students and LEP students in grades 8-11 taking the BSTs in February 1999. It is important to keep in mind that the passing score for 11th graders (the class of 2000) is 70% and for other students in the table it is 75%. It is not clear from the data how many students represented in 9th through 11th grade are first time test takers and how many are taking the test for a 2nd, 3rd or 4th time because of not passing one or both portions previously.

Table 6. Performance of 8-11th Graders on 1999 BST

For all students in grades 9 through 11, about 50% of the students in any grade passed the reading test compared to about 25% of LEP students. For the math test, about one third of the students in grades 9 to 11 passed. LEP students taking the math tests have passing rates lower than the rates for all students in the same grades, ranging from 18 to 24 percent. Overall, LEP students in all grades have lower passing rates than students in general.

Accommodations. The way in which data are collected on the use of accommodations during the BSTs has changed over time. In 1999, data on what particular accommodations were used by each student were collected on a form that accompanied the test form. Table 7 summarizes the use of accommodations in relation to performance by LEP students in 8th grade in 1999. It is important to note that in these data, accommodations and translations of the test are not differentiated, so “accommodated” means that the student was provided an accommodation, a translation, or both. The data in Table 7 include all LEP students who were reported to have received at least one accommodation or translation. Some of these students may have also been students with disabilities and the accommodation received could have been an accommodation based on their LEP status, their disability status, or both. Whatever the case, very few LEP students were reported to have received accommodations or translations in this testing cycle.

Table 7. Accommodation Use and Performance for LEP 8th Grade Students

Of the 8th grade LEP students taking the tests, only 2.7% used an accommodation on the reading test and only 2.1% used an accommodation on the math test. Over 97% of LEP students participating in each test took the test with no accommodation.

In terms of performance, just less than 8% of the accommodated LEP students on the reading and math tests achieved a passing level. Of the unaccommodated LEP students, 21.9% passed the reading test and 24.5% passed the math test. While the passing rate is lower for the accommodated students, this should not be taken to mean that the accommodations decrease student performance on the test. Students receiving accommodations are usually students with lower proficiency levels in English, and thus have more difficulty passing the tests (Albus, Liu, Thurlow, & Bielinski, 2000; Anderson, Liu, Swierzbin, Thurlow, & Bielinski, 2000). Taking into consideration the low numbers of students taking a test with accommodations it is possible that some of the students in the unaccommodated group would have benefited from receiving accommodations, or did receive accommodations that were not reported.

These data are in contrast to the accommodation usage data from 1997 when 12% of LEP students were reported to use some form of accommodation. Of LEP students using accommodations in 1997, 83% passed the math test and 2% passed the reading test (Liu & Thurlow, 1999). Even though the numbers of LEP eighth graders taking the BSTs more than doubled between 1997 and 1999, the number of LEP students who were reported as using accommodations on the tests fell by more than 50% for each. These numbers indicate that either accommodation use by LEP students in Minnesota is drastically declining or that it is not being reported accurately.

Types of accommodations used. Test administrators also reported the specific accommodations students used when taking the BSTs. The data were complied by CFL and merged with the MARSS file on test participation. The accommodations used by LEP students are reported in Table 8. These data reflect the number of LEP students who were reported to use each type of accommodation and who also had a valid test score in the MARSS database.

Although we have questions about the validity of the accommodations, the accommodation data for LEP students who had valid scores for each test are reported in Table 8. Some of the students may have received multiple accommodations and are counted in more than one category in the table.

Table 8. Accommodations Used by LEP 8th Grade Students in 1999

* This accommodation is not offered for the reading test.
** This accommodation is for students with disabilities.

Like the data presented in Table 7, the numbers of students reported as having received accommodations is very low. Of the more than 1,800 eighth grade LEP students who took each test, the number of students reported as receiving the most popular accommodation was 23. The most frequently used accommodations for the reading test were short segment test booklets, extended time, and special setting accommodations. The most used accommodations for the math test were short segment test booklets, individual or small group administration of the test, extended time, and special setting accommodations. Although permitted, translations of tests were rarely offered for the math test, and no LEP students were reported as having received clarification or a translation of the test directions. More LEP students were reported as having received a large print version of the test, an accommodation specific to students with disabilities, than a translation of the math test.

Implications

Based on four years of Basic Standards Test data for LEP students, we have observed the following:

•  Participation for LEP students in the Minnesota BSTs remains high.

•  Performance for LEP students remains low. There is an achievement gap of approximately 20%-25% between mean scores for LEP students and mean scores for their native-English speaking peers on both the reading and the math tests. This gap does appear to be narrowing a bit, but the gap is not unexpected given the fact that these students are in the process of learning the academic English needed to take the tests.

•  Greater gains in mean percentages of items correct are being made by LEP students taking the reading tests than the math tests. This suggests that ESL teachers may be successful in teaching reading skills needed on the BST. More detailed examination of the types of remediation programs provided for students who do not pass the math test may give clues as to why the gains are smaller for the math BST.

Based on the very limited data in this report pertaining to students using accommodations and students retaking the BSTs in grades 9-11, several observations can be made.

If approximately 25% of the LEP students who retake the reading or math test each year in grades 9-11 pass it, there will be LEP students who do not pass the tests by 12th grade and therefore do not receive a high school diploma. Future BST analyses need to include those LEP students taking the tests in summer remediation programs to gain an accurate picture of how many LEP students are not receiving diplomas because they are not able to pass the tests.

Data on which accommodations are being used for the reading versus the math test are not yet accurate enough to use to draw conclusions. Although the specific accommodation data did allow for recording whether the accommodation was used for the reading or the math test, all of the students in the file were listed as using the same accommodations for both tests. This fact makes the data somewhat suspect considering that some accommodations (e.g., oral administration of the test in English and translation) are not allowed for the reading test. Most likely these accommodations were not used for the reading test but are a result of not differentiating between the reading and math tests on the accommodations form. In fact, some students who were not listed as having been tested for a particular test in the MARSS database were reported as having received an accommodation for the test. Clearly, if a student does not take a test, that student did not receive an accommodation.

At the present time, data on accommodations use are only collected for 8th graders. Test administrators are asked to fill out a separate sheet for the reading and math tests telling what accommodations an individual student uses. Anecdotal information from the test administrators suggests that the extra time and effort required to fill in the separate sheets for large numbers of LEP students is a burden. It appears that instead of two, only one sheet is filled out with information about reading and math accommodations combined. There may be a way to collect accommodations data so that it is less of a burden on test administrators and so that it is easier to analyze. In Missouri, for example, accommodations information is collected on the test answer sheet so that separate forms are not needed. If students using accommodations were tested in small groups, it would be easier for test administrators in this type of setting to accurately record accommodations information. Better accommodations data would help researchers and policymakers determine which accommodations are being used, and whether particular accommodations are having more benefit for LEP students than others.

Finally, the numbers of LEP students using accommodations seems low. Although accommodation usage varies among states with statewide accountability systems, the numbers tend to be higher than found in the Minnesota data. In a recent testing cycle in Missouri, about 10% of LEP students use accommodations on statewide tests (J. Bielinski, personal communication, June 20, 2000). Reports of accommodation use by students with disabilities in Rhode Island, another group for whom accommodations are allowed, have shown up to 53% of these students using accommodations on statewide math tests (Elliott, Bielinski, Thurlow, DeVito, & Hedlund, 1999). If the numbers of students using an accommodation on the BSTs are fairly accurate, more research is needed into why the numbers of accommodated LEP students are so low in comparison with states like Missouri that have a smaller LEP student population and lower than other groups of students who are allowed accommodations such as students with disabilities. Do educators and school staff who make testing decisions need more training regarding the purpose of accommodations and the benefit to LEP students? Are ESL and bilingual teachers included in making accommodations decisions for LEP students? Are there more useful accommodations than the ones currently offered? Do the students and their families not want to use accommodations? Is the school worried about the legal ramifications of offering a translation to one language group, but not offering it to another because of the financial cost and the availability of translators? Answering these questions would help to make the testing situation more equitable for LEP students.

References

Albus, D., Liu, K., Thurlow, M., & Bielinski, J. (2000). The effects of a simplified English dictionary accommodation for LEP students taking large-scale reading assessments. (Delaware Report). Minneapolis, MN: University of Minnesota, National Center on Educational Outcomes.

Anderson, M., Liu, K., Swierzbin, B., Thurlow, M., & Bielinski, J. (2000). Bilingual accommodations for limited English proficient students on statewide reading tests: Phase 2. (Minnesota Report 31). Minneapolis, MN: University of Minnesota, National Center on Educational Outcomes.

Kopriva, R. (2000). Ensuring accuracy in testing for English language learners. Washington, DC: Council of Chief State School Officers.

Elliott, J., Bielinski, J., Thurlow, M., DeVito, P., & Hedlund, E. (1999). Accommodations and the performance of all students on Rhode Island’s Performance Assessment. (Rhode Island Report 1). Minneapolis, MN: University of Minnesota, National Center on Educational Outcomes.

Lachat, M. (1999). What policymakers and school administrators need to know about assessment reform for English language learners. Providence, RI: Brown University, Northeast and Islands Regional Education Laboratory.

Liu, K., & Thurlow, M. (1999). Limited English proficient students’ participation and performance on Minnesota’s statewide assessments: Minnesota Basic Standards reading and math, 1996-1998 (Minnesota Report 19). Minneapolis, MN: University of Minnesota, National Center on Educational Outcomes.

Rivera, C., Stansfield, C., Scialdone, L. & Sharkey, M. (2000). An analysis of state policies for the inclusion and accommodation of English language learners in state assessment programs, 1998-99. Arlington, VA: Center for Equity and Excellence in Education, George Washington University.

Rothman, R., & Elmore, R. (1999). Testing, teaching and learning: A guide for states and school districts. Washington, D.C.: National Academy Press

Thurlow, M., Liu, K., Erickson, R., Spicuzza, R., & El Sawaf, H. (1996). Accommodations for students with limited English proficiency: Analysis of guidelines from states with graduation exams (Minnesota Report 6). Minneapolis, MN: University of Minnesota, National Center on Educational Outcomes.