MINNEAPOLIS PUBLIC SCHOOLS Districtwide Assessment Results

MINNEAPOLIS PUBLIC SCHOOLS Districtwide Assessment Results 2003-2004 Kindergarten Assessments Oral Reading Assessments Northwest Achievement Levels Tests Minnesota Comprehensive Assessments Minnesota Basic Skills Tests Test of Emerging Academic English Charter School Comparison Graduation and Drop-outs District Survey Results Research, Evaluation and Assessment December 2004

Contributors to this report include: Judy Allgood Joe Betts Alex Chan Donna Clark Carol Dordan Diane Dalluge Elizabeth Dolan Melody Jacobs-Cassuto Christina Sheran Susan Godlewski Juanita Lyons Mary Pickart Kathryn O Gorman Eric Vanden Berk David Heistad, Director Page 2Page 2

TABLE OF CONTENTS Executive Summary. 1 Introduction. 3 Glossary of Terms 4 Kindergarten Assessments 5 Oral Reading Assessments... 12 Northwest Achievement Levels Tests (NALT)... 16 Elementary Minnesota Comprehensive Assessments (MCA). 32 Minnesota Basic Skills Tests (MBST). 47 High School Minnesota Comprehensive Assessments (MCA)... 61 Graduation and Drop-out Rates... 70 Test of Emerging Academic English (TEAE). 72 Staff and Student Survey Results. 89 Page 3Page 3

MINNEAPOLIS PUBLIC SCHOOLS 2003-2004 District wide Assessment Results EXECUTIVE SUMMARY OF FINDINGS Overall, all of the students in the district made larger gains over the school year than in previous years on the Kindergarten literacy measures. All ethnic groups except White students made gains in Phonemic Awareness. Gaps between students of color and White students decreased. All ethnic groups made gains in the Alphabetic Principle domain of literacy. There were significant increases in the percent of students from all ethnic groups that were able to read more than 10 words per minute by end of kindergarten. Forty-one percent of first graders met or exceeded the district standard (Level 3) for reading words per minute and reading expression. This was a higher rate of reading fluency compared to the past three years of 1 st grade assessment. Average NALT scale score performance in 2004 compared to national norms was higher in math than reading. Average math performance was above the national average for grades 2-5 but below average for grades 6, 7 and 9; reading was below the national scale score average for all grades. Reading growth for continuously enrolled students from 2003 to 2004 was below the national norm for all grades except grade 4. Math growth for continuously enrolled students from 2003 to 2004 was above the national norm for grades 4 and 5 but below the national norm for the remaining grades. The math and reading rates of growth for White students exceeded the rates for Students of Color. In other words, the gap between these groups increased from grade 3 through grade 7. White students started and ended at considerably higher levels. Average MCA scores in Grade 3 Reading have increased every year. Average MCA scores in Grade 3 Math decreased seven scale score points last year after four years of consecutive gains. Average MCA scores in Grade 5 Reading have leveled off and closely mirror 2003 scores. Average MCA scores in Grade 5 Math have increased for the past five years. Performance gaps to the standards for students of color, students in poverty, students receiving ELL and Special Ed services continue to remain large. Comparisons to the State of Minnesota MCA totals show similar gaps at the state level. Hispanic and White performance in MPS was higher than State Page 1Page 1

EXECUTIVE SUMMARY OF FINDINGS (Continued from previous page) Key Findings From 2000 to 2004 the passing rate for 8 th grade reading declined 3% and math declined 6%. Seven additional items were added to the math test in 2004 possibly contributing to the lower passing rates. Of the 9 math content areas Measurement Concepts, Fractions, and Ratios/Percentages, had the lowest scores compared to state averages. In reading and math, passing rates for White Students were considerably higher than those for Students of Color. This gap was present but less pronounced for writing. All ethnic groups had lower math passing rates compared to 2003. Non-ELL students had much higher passing rates than ELL students. This was true for reading, math, and writing for all ethnic groups. For both math and reading, the largest decline in passing rates was noted for student receiving Special Education services. High Schools Readiness for graduation has continued to improve. As of October 2004, 39% of the graduating Class of 2004 had yet to pass one or more MBST area, up 1% from last year, and 26% were not on course with enough credits to graduate, down 9% from last year. About 55% of the Class of 2004 was credit ready and had passed all MBST areas as of October 2003, up 5% from last year. About half of Minneapolis grade 10 students scored at or above the benchmark of proficiency (Level 3 or above) on the high school MCA reading test, compared to 8 of students in the state. Forty-four percent of Minneapolis grade 11 students scored at Level 3 or above on the high school MCA mathematics test. This was about 25% below the percentage of students scoring at or above the proficiency level in the state (7). Ethnic groups varied in performance. White and Asian students had a higher percentage scoring at Level 3 or above on both tests, compared to Indian, African, and Hispanic students. The test performance of White students in MPS on the high school MCAs was comparable to the White students in the State, whereas the test performance of students of color in MPS was substantially behind the State. African students represent the largest ethnic group in Minneapolis Public Schools, and had the lowest percent of students scoring at Level 3 or above. Official graduation rates reported for NCLB increased from 73% to 78% for the 7 major High Schools. Total district graduation rates including all alternative schools increased from 51% to 57% Page 2Page 2

INTRODUCTION For the past several years, the Minneapolis Public Schools (MPS) Board of Education has reached consensus around a District Improvement Agenda (DIA) that prioritizes and guides the work of the district. The District Improvement Agenda for 2001-2004 included four main goals: a) accelerate achievement of all students, b) develop mutual trust and involvement with students, families, and the community, c) hold everyone accountable for student learning, and d) ensure efficient management of the district operations. This annual report on district wide assessment results focuses on the attainment of the first and broadest goal, that of accelerating the achievement of all students. To answer these objectives, results for most assessments are presented in at least two major ways: 1) current performance level, and 2) growth of continuously enrolled students. Current performance level summarizes 2003-2004 test results for all Minneapolis students and is reported on several variables, such as ethnicity, home languages and educational program. Performance in 2003-2004 is compared to previous years performance when available. The growth section reports achievement information on students who were continuously enrolled in the Minneapolis Public Schools across specified periods of time, usually across two testing periods. This section also reports results disaggregated by several variables. Page 3Page 3

GLOSSARY OF TERMS Cohort is defined as a group of students at a particular grade level. A cross-cohort analysis compares two different groups of students (for example, grade 3, 2004 vs. grade 3, 2003). The change in student performance of different cohorts is referred to as improvement rather than growth. Continuous enrollment is defined as those students who were enrolled for at least two testing periods, so that gain can be measured. Individual school gains were calculated using all students who were enrolled in the fall and the spring at the same school. District gains were calculated for students enrolled fall and spring in the district. English Language Learners (ELL) is the preferred term for students served in Limited English Proficiency programs. Most, but not all, English Language Learners have a non-english home language. Home language codes were used to differentiate the type of non-english language spoken in the home. General education students are those who receive neither Special Education nor Limited English Proficiency services. Median percentile is the national percentile corresponding to the score where half of the district s students score above and half score below a particular value. The national median percentile is 50. As a rough rule of thumb, districts that have a median percentile rank between 40 and 60 can be considered within the national median range. NCEs (Normal Curve Equivalents) place students percentile rankings on an equal-interval scale so that gains made at different points of the achievement distribution can be compared. NCEs, like percentiles, range from 1 to 99 with an average of 50. As a rough rule of thumb, districts that have an average NCE between 45 and 55 can be considered within the national average NCE range. Percentile rank indicates a student s relative standing. It is the percentage of students in the national norm group at the same grade that scored the same or lower than the student. For example a student with a percentile rank of 75 in reading exceeded 75% of the students in the norm group in reading. Rounding: Data in this report were rounded to the nearest whole number for percentages. Percentages of students by categories may not add to 10 due to rounding. Page 4Page 4

KINDERGARTEN ASSESSMENTS Key Findings: Overall, all of the students in the district made larger gains over the school year than in previous years on the literacy measures. All ethnic groups except White students made gains in Phonemic Awareness. Gaps between students of color and White students decreased. All ethnic groups made gains in the Alphabetic domain of literacy. There were significant increases in the percent of students from all ethnic groups that were able to read more than 10 words per minute by end of kindergarten. Students are coming into the district less well prepared than in past years. This is indicated by the lower percentages of students at the Level 2 predictor on all BKA measures. Lower initial skill level indicates that students may not be as well prepared for kindergarten as they have been in the past. Phonological Awareness Gains Decreased for Indian, African, Asian and Hispanic in 2002-03 vs. 2001-02. Beginning of Kindergarten Assessments All kindergarten students enrolled in the Minneapolis Public Schools were assessed in September and October. This assessment was individually administered by the Early Childhood Cadre, a group of testers trained on its standardized administration using a standard set of materials. The Beginning Kindergarten Assessments measures literacy and numeracy skills. Items were selected to align with MPS Beginning of Kindergarten Expectations in Reading, the MPS Early Childhood Standards Birth to Kindergarten, and the National Research Council. The literacy assessment was revised in August of 2001 to include measures of phonemic awareness (rhyming and alliteration) and vocabulary (picture naming). Dr. Scott McConnel and the late Dr. Mary McEvoy developed the items added in 2001 at the University of Minnesota. They are in the public domain and can be found at http://ggg.umn.edu. For further information, contact Mary Pickart (mary.pickart@ mpls.k12.mn.us). Basic Literacy Benchmarks Literacy Benchmarks have been established for the Beginning and End of Kindergarten Assessments. Benchmarks are grouped into two domains: the alphabetic principle and phonemic awareness. The Alphabetic Principle Domain is calculated by adding the letter sound and letter naming scores. The Phonemic Awareness Domain is calculated by adding the rhyming and alliteration scores. Two levels of benchmarks have been calculated based on validity studies where Grade One Oral Reading Assessment data were used to predict Grade Three Minnesota Comprehensive Assessment proficiency and where Beginning and End of Kindergarten Assessment scores were used to predict Grade One Oral Reading Assessment proficiency. Table 1 below shows the benchmarks for each domain for beginning and end of kindergarten. Page 5Page 5

Table 1. 2003-2004 Literacy Benchmarks Scores that predict reading well at the end of first grade (Level 2 = 48 words per minute; Level 3 = 60 words per minute) Phonemic Awareness Alphabetic Principle Beginning of Kindergarten End of Kindergarten Level 2 Level 3 Level 2 Level 3 12 21 27 36 15 24 63 81 The Level benchmarks correspond with reading proficiency at the end of first grade and with passing levels on the Grade Three Minnesota Comprehensive Assessment. The Level 2 benchmark corresponds with reading 48 words correctly at the end of first grade (MPS Oral Reading) and obtaining a score at level 2 on the MCA. Level Three benchmarks correspond with reading 60 words correctly at the end of first grade (MPS Oral Reading) and obtaining a score at level 3 on the MCA. Table 2 summarizes beginning kindergarten results on seven literacy and numeracy concepts. For each of the literacy items large differences by racial/ethnic group existed at the beginning of the year. For example, twice as many White students met the level two Phonemic Awareness benchmark as Hispanic and Asian students. In addition, three times as many White students met the level two Alphabetic Principle benchmark as Hispanic students. Table 2. 2003-2004 Beginning Kindergarten Assessment By Racial/Ethnic Group Literacy Concepts of Print Phonemic Awareness Alphabetic Principle N Level 2 Level 3 Level 2 Level 3 Level 2 Level 3 Indian 130 21% 13% 12% 11% 17% 25% African 1193 22% 13% 13% 22% 2 19% Asian 342 13% 9% 1 22% 16% 15% Hispanic 468 15% 6% 6% 11% 16% 18% White 912 31% 45% 19% 46% 22% 44% Numeracy Identifies Numbers Count to 35 1-1 Correspondence Extends Pattern Indian 18% 76% 82% 39% African 27% 73% 66% 38% Asian 17% 66% 73% 39% Hispanic 12% 61% 7 29% White 52% 93% 88% 71% Page 6Page 6

Figure 1. Percentage of Kindergarten Students at Level 2 Benchmark for Phonemic Awareness- Fall 2001-2004 2001 2002 2003 2004 10 Percent of Students 8 6 4 2 Indian African Asian Hispanic White All Students Racial/ethnic gaps in the literacy area of phonemic awareness changed very little from the fall of 2001 to the fall of 2004. While approximately 8 of White students scored at the level two benchmark, less than 4 of African and Indian students scored at the level two benchmark. In addition, approximately 2 of the Asian and Hispanic students met the level two benchmark. It should be noted that these skills are assessed in English and that native language phonemic awareness is likely much higher for English Language Learners. Figure 2. Percentage of Kindergarten Students at Level 2 Benchmark for Alphabetic Principle- Fall 2001-2004 2001 2002 2003 2004 10 Percent of Students 8 6 4 2 Indian African Asian Hispanic White All Students Page 7Page 7

Figure 2 depicts the racial/ethnic trends for the alphabetic. In this case the percentage of students at the level 2 benchmark for 2004 has decreased slightly for all groups as compared to the previous years. The four-year trend for total literacy is graphed in Figure 3. All literacy areas were combined into a total score. Asian and Hispanic student readiness for Kindergarten has continued to increase from the fall of 2001 to the fall of 2004. Figure 3. Percentage of Kindergarten Students at Level 2 Benchmark for Total Literacy- Fall 2001-2004 2001 2002 2003 2004 10 Percent of Students 8 6 4 2 Indian African Asian Hispanic White All Students Table 3 indicates the literacy and numeracy gaps at the beginning of kindergarten by home language group. Gaps in numeracy are smaller than literacy as is expected for students who speak a language other than English at home. In general, Hmong and Spanish-speaking students had the lowest percentage of students meeting level benchmarks in literacy and numeracy. Table 3. 2003-2004 Beginning Kindergarten Assessment By Home Language Literacy Phonemic Awareness Alphabetic Principle Concepts of Print N Level 2 Level 3 Level 2 Level 3 Level 2 Level 3 English 2185 27% 27% 15% 32% 21% 31% Hmong 231 7% 1% 8% 14% 13% 9% Spanish 384 12% 5% 6% 7% 13% 1 Somali 129 15% 12% 11% 29% 15% 16% Numeracy Count to 35 1-1 Correspondence Extends Pattern Identifies Numbers English 39% 84% 78% 54% Hmong 8% 59% 69% 31% Spanish 8% 56% 68% 25% Somali 24% 61% 61% 43% Page 8Page 8

End of Kindergarten Assessments The Early Childhood Cadre assessed all kindergarten students in April and May on literacy and numeracy skills. All of the measures administered in the fall were repeated in the spring. In addition, phoneme segmentation and passage reading were assessed. As with the beginning of Kindergarten measures, the Minneapolis Public Schools Grade Level Expectation guided the content of the assessment. Table 4 summarizes end of kindergarten results on eight literacy and numeracy concepts. While a greater percentage of White students are meeting level two and three benchmarks, on most literacy items in the spring of 2004, the gaps had narrowed significantly. In the area of Alphabetic Principle, a greater percentage of Asian students met the level three alphabetic principle benchmark than White students in the spring; African students also improved to levels much closer to White students. Table 4. 2003-2004 End-of-Kindergarten Assessments By Ethnic Group Literacy Phonemic Awareness Alphabetic Principle Concepts of Print N Level 2 Level 3 Level 2 Level 3 Level 2 Level 3 Indian 130 22% 12% 18% 21% 28% n/a African 1193 23% 13% 2 31% 3 n/a Asian 342 17% 12% 19% 44% 35% n/a Hispanic 468 16% 5% 13% 21% 26% n/a White 912 35% 31% 23% 39% 52% n/a Numeracy Count to 35 Extends Pattern Orders NumbersIdentifies Shapes Identifies Coins Indian 75% 78% 78% 75% 41% African 77% 71% 78% 73% 53% Asian 78% 82% 82% 78% 58% Hispanic 62% 78% 66% 53% 36% White 88% 9 9 86% 63% Literacy gains for kindergarten students enrolled the entire year are presented in Figures 4 and 5. Consistent gains across racial/ethnic groups are evident from these graphs. Page 9Page 9

Figure 4. Phonemic Awareness Gains 2001-2003 by Racial/Ethnic Group 01-02 02-03 03-04 15.0 Mean Raw Score Gain 12.0 9.0 6.0 3.0 0.0 Indian African Asian Hispanic White All Students Figure 5. Alphabetic Principle Gains 2001-2003 by Racial/Ethnic Group 01-02 02-03 03-04 70 Mean Raw Score Gain 60 50 40 30 20 10 0 Indian African Asian Hispanic White All Students End of Kindergarten reading fluency on a simple first grade passage increased significantly from spring 2003 to spring 2004. When viewed by racial/ethnic category all groups reading increased. These trends are depicted in Figure 6 as the percentage of students reading at least 10 words correctly per minute at the end of Kindergarten. Page 10Page 10

Figure 6. Percentage of Students Reading at Least 10 words per minute at the End of Kindergarten by Racial/Ethnic Category 01-02 02-03 03-04 10 Percent of Students 8 6 4 2 Indian African Asian Hispanic White All Students Page 11Page 11

ORAL READING ASSESSMENT Key Findings: Forty-one percent of first graders met or exceeded the district standard (Level 3) for reading words per minute and reading expression. This was a higher rate of reading fluency compared to the past three years of 1 st grade assessment. More White students than Students of Color met the district standards in the Oral Reading Assessments. Large gaps remain for students of color, especially for Hispanic students. All student groups continue to make gains in Oral Reading Fluency in the spring of 2004 compared to the spring of 2003. Students whose home language was English scored higher than students whose home language was not English. About the Assessments All students enrolled in first grade are assessed in the spring with the district-developed Oral Reading Assessment Students received scores on each of three reading dimensions: Standards Words per minute is the median number of words read correctly in one minute across three passages. The district standard is 60 words per minute for ending first graders. While efforts are made to keep the level of difficulty the same, the reading passages are not identical from year to year and may contribute to some fluctuation in scores. Reading expression is rated on a five-point scale with attention to phrasing, flow, and inflection. The three passages are averaged and the district standard is 3.3. Reading comprehension is based on student answers to literal and inferential questions asked after the students read each passage. Because the comprehension questions are not equivalent across passages or years, a district standard has not been set. Validity studies of the relationship between Oral Reading at the end of first grade and performance on the MCA reading test at the end of third grade found a high correlation between words correct in first grade and reading comprehension in third grade. Students who read 60 or more words correctly per minute at the end of first grade scored proficient (level 3 or better) on the MCA 8 of the time. Students who read 48 or more words correctly per minute at the end of first grades scored at level 2 or better on the MCA 9 of the time. Questions? Contact Mary Pickart; mary.pickart@mpls.k12.mn.us Page 12Page 12

Overall Performance on the Oral Reading Assessment Performance on the Spring Oral Reading Assessment in first grade was the highest in 2004 compared to every year since the assessment was begun in 1998. The percentage of students reading at least 48 words correctly per minute grew another 4% in 2004 (see figure 7). In the past two years the percentage of students projected to score at level 2 or above on the 3 rd grade Reading MCA has increased from 46% to 54%. Figure 7 Percent of Grade 1 Students Meeting the Level 2 Fluency Standard 2000-2004 Percent of Students 10 8 6 4 2 45% 45% 46% 5 54% 2000 n= 3881 2001 n=3601 2002 n=3401 2003 n=3108 2004 n=3003 Words Read Correctly 48 Figure 8. Percent of Grade 1 Students Meeting the Level 3 Fluency Standard 2000-2004 10 Percent of Students 8 6 4 2 33% 33% 34% 4 41% 2000 n= 3881 2001 n=3601 2002 n=3401 2003 n=3108 2004 n=3003 Words Read Correctly Figure 8 shows an increase in the percentage of students projected to score at Level 3, the proficiency standard for No Child Left Behind, on the 3 rd grade reading MCA test. In the past two years the percentage of students reading at least 60 words per minute has increased from 34% to 41%. Page 13Page 13

Figure 9. Percent of Grade 1 Students Meeting Level 2 Oral Reading Assessment Standard by Racial/Ethnic Group Percent of students 10 8 6 4 2 5 45% Indian n=132 African n=1164 74% 57% 36% Asian n=356 Hispanic n=460 White n=891 Words Read Correctly Large gaps remain among racial/ethnic groups on the end of Grade 1 Oral Reading Assessment. The proportion of White students at the level two standard was approximately double that of Hispanic students (i.e. 74% White compared to 36% Hispanic). English Language Learners (ELL) had lower rates of fluency (words read correctly) and reading expression compared to students who were not receiving ELL services (see figure 10). Figure 10 Percent of Grade 1 Students meeting the Oral Reading Assessment Standards by ELL and non-ell 2004 Percent of students 10 8 6 4 2 58% 41% 45% 3 Words Read Correctly Level 2 Words Read Correctly Level 3 ELL Non-ELL Page 14Page 14

Reading fluency at the end of grade 1 in 2004 was higher than 2003 for every racial/ethnic group. Improvements of 7 percentage points were observed for Asian students. African and White student s scores increased 1 percentage point. Indian students improved by 2 percentage points and Hispanic students by 4 percentage points. Percent of Students at District Standard Figure 11. Grade 1 Oral Reading Improvement by Racial/Ethnic Group Percent of Students at Level 3 Standard 8 6 4 2 African Asian Hispanic White 2002 24% 28% 35% 15% 55% 2003 31% 32% 38% 2 6 2004 33% 33% 45% 24% 61% Page 15Page 15

NORTHWEST ACHIEVEMENT LEVELS TESTS (NALT) Key Findings for Overall Performance: Average NALT scale score performance in 2004 compared to national norms was higher in math than reading. Average math performance was above the national average for grades 2-5 but below average for grades 6, 7 and 9; reading was below the national scale score average for all grades. Average reading scores were at or above the basic standard cut score (which projects passing the Minnesota Basic Skills Tests in 8 th grade) for all grades. Math scores were at or above the basic standard cut score except at grade 7 which was below. Average reading and math scores for Asian students was at or above the basic standard cut score for all grades. White students were considerably above the cut score at all grades for both reading and math. Reading performance as measured by scale scores and percentiles showed small positive increases from 2002 through 2004 for grades 2-4 and 7, 9. It remained steady at grades 5 and 6. Math performance increased from 2002 through 2004 for all grades Key Findings for Student Growth: Reading growth for continuously enrolled students from 2003 to 2004 was below the national norm for all grades except grade 4. Math growth for continuously enrolled students from 2003 to 2004 was above the national norm for grades 4 and 5 but below the national norm for the remaining grades. Reading growth from 2003 to 2004 was similar to the growth from 2002 to 2003 for grades 3 and 7, showed an increase in grade 4 and a decrease in grades 5 and 6. For math, growth increased for grades 3-5 but decreased for grades 6 and 7. Hispanic, White and Asian students made the most growth in reading and math. Continuously enrolled students remained steady in their standing with respect to the basic standard reading and were above the math cut scores (projects passing the Minnesota Basic Skills Tests in 8 th grade) as students progressed from grade 3 through grade 7. Continuously enrolled students were below the high standards for all grades except 7 for reading. In math, continuously enrolled students were below the high standards for grades 3-4 but at or above for grades 5-7. The math and reading rates of growth for White students exceeded the rates for Students of Color. In other words, the gap between these groups increased from grade 3 through grade 7. White students started and ended at considerably higher levels. Page 16Page 16

About the Assessments The Northwest Achievement Levels Tests (NALT) were given district wide in March to students in grades 2 through 7 and 9 in reading and math. The NALT measurement system serves multiple purposes: a) to measure students knowledge and skills with respect to the Minneapolis curriculum standards; b) to compare individual and aggregate student growth in learning to national user norms; and c) to generate information that can be used for accountability, policy making, and school improvement. The multiple levels of the NALT make it possible to give each student a test appropriate to a student s current achievement level. Because the test is linked from level-to-level and grade-to-grade, it is possible to get an accurate measure of the growth at the student, class, school, and district level. NALT test items were specifically selected by district personnel from a large item bank to best match the district curriculum standards and daily learning in the classroom. The NALT user group norms, collected in 2002, include scores from over 1,00000,000 students and provide the most comprehensive growth data ever assembled for a standardized multiple-choice test. In the fall of 2000, a new set of math tests was developed by Minneapolis staff and in the fall of 2002, a new set of reading tests was developed. Questions? Contact Melody Jacobs-Cassuto, mcassuto@mpls.k12.mn.us. For more information: the web site for the publisher of the NALT Northwest Evaluation Association, is http://www.nwea.org. Overall NALT Performance for all Test Takers Table 5 shows the average test scores on the NALT in terms of general education students (GE), English Language Learners (ELL), Special Education students (Sp.Ed.), and all students combined by grade level. The table includes all students tested in 2002, 2003, and 2004 but a different group of students is reported each year at each grade (cross cohort). The number of students assessed in April, 2004 by program appears in the first row for each grade in Table 5. The percentage column refers to the percent of all students enrolled at each grade level who were assessed with the NALT. Participation rates for grades 2-7 during the past three years have been excellent (92% or higher). Grade 9 participation rates are lower. They had shown an increase in 2003 but dropped in 2004. Below the average scale scores are NALT percentile ranks that shows the performance of the average Minneapolis student as compared to students in the national user group norms. As a rule of thumb, district and school median percentile ranks that range between the 40 th and 60 th percentile can be considered in the national average range. Math achievement was stronger than reading. The math percentiles clustered in the 40 th to 50 th percentile ranges for grades 2 through 5 with grades 6, 7 and 9 all below the national range. The reading percentiles, however, were in the 30 th to 40 th percentile ranges (below the national range) for all grades except Grade 2. In comparison to national norms, the average percentile ranks change with each new norm group, but the scale scores for the NALT mean exactly as much achievement on the curriculum scale as it did three years ago or ten years ago. Table 5 shows increases in both math and Page 17Page 17

reading scale scores across all grades except grade 6 reading which remained steady from 2002 to 2004 for the entire population of students. Table 5. Average NALT Scale Scores and Median Percentile Ranks by Program ** Grade and Year Reading Mean Scale Scores Math Mean Scale Scores % Tested All ELL Sp. Ed. GE* % Tested All ELL Sp. Ed. GE* Grade Two Percent 2999 803 265 1931 Percent 3004 805 268 1931 2002 96% 183 176 164 188 97% 192 189 180 194 2003 97% 184 177 169 188 98% 191 188 181 193 2004 94% 186 178 169 191 94% 194 189 182 197 NALT Percentile 2002 44 th 28 th 7 th 53 rd 56 th 47 th 20 th 66 th NALT Percentile 2004 47 th 31 st 12 th 58 th 62 nd 47 th 22 nd 47 th Grade Three Percent 2924 765 289 1870 Percent 2933 776 288 1869 2002 96% 193 187 177 198 96% 203 200 190 205 2003 97% 193 185 178 198 97% 201 198 190 204 2004 95% 194 188 177 199 95% 203 200 191 206 NALT Percentile 2002 38 th 25 th 12 th 47 th 60 th 48 th 20 th 69 th NALT Percentile 2004 36 th 27 th 10 th 49 th 60 th 49 th 20 th 36 th Grade Four Percent 3111 817 343 1951 Percent 3118 825 344 1949 2002 96% 200 192 185 205 96% 208 204 197 212 2003 97% 199 192 185 205 97% 208 205 196 212 2004 94% 201 193 186 207 95% 211 207 197 215 NALT Percentile 2002 35 th 22 nd 12 th 43 rd 49 th 40 th 17 th 56 th NALT Percentile 2004 36 th 22 nd 11 th 52 nd 53 rd 45 th 20 th 36 th Grade Five Percent 3113 723 444 1946 Percent 3117 730 443 1944 2002 97% 205 196 191 211 96% 215 209 202 220 2003 98% 206 197 192 212 98% 216 210 203 221 2004 94% 206 198 191 212 95% 218 213 204 223 NALT Percentile 2002 31 st 20 th 11 th 42 nd 43 rd 33 rd 15 th 53 rd NALT Percentile 2004 34 th 21 st 11 th 50 th 48 th 40 th 20 th 34 th Grade Six Percent 3135 672 404 2059 Percent 3146 675 416 2055 2002 95% 210 199 195 216 95% 218 210 203 224 2003 97% 210 200 193 216 97% 218 212 202 224 2004 92% 210 201 195 217 92% 219 212 202 225 NALT Percentile 2002 30 th 17 th 10 th 41 st 38 th 23 rd 11 th 49 th NALT Percentile 2004 33 rd 18 th 9 th 46 th 39 th 27 th 11 th 33 rd Grade Seven Percent 3109 657 442 2010 Percent 3100 656 441 2003 2002 93% 213 202 198 220 93% 222 213 205 229 2003 96% 213 202 197 220 96% 222 214 206 229 2004 92% 214 205 198 221 92% 224 217 206 230 NALT Percentile 2002 28 th 15 th 9 th 41 st 35 th 23 rd 11 th 49 th NALT Percentile 2004 33 rd 20 th 8 th 46 th 40 th 28 th 12 th 33 rd Grade Nine Percent 2224 420 250 1554 Percent 2153 410 247 1496 2002 73% 220 205 205 226 73% 231 218 212 238 2003 88% 218 203 203 226 87% 231 219 210 239 2004 59% 221 209 204 227 57% 234 223 213 240 NALT Percentile 2002 32 nd 13 th 10 th 48 th 25 th 13 th 6 th 35 th NALT Percentile 2004 33 rd 11 th 9 th 47 th 26 th 14 th 7 th 33 rd *GE signifies general education students who received neither Special Education nor English Language Learning support Page 18Page 18

Figures 13 and 14 depict the change in overall average NALT scale score for each grade level in reading and math based on the data in Table 5. Scale scores increase across grade level on a developmental scale similar to height and weight. As students get older, their skills in reading and math should increase. Also graphed is the MBST cut score; that is the score that predicts whether a student is likely to pass the eighth grade Minnesota Basic Skills Tests (MBST, see Table 7). Figure 13. Average Reading Scale Score on NALT 2002-2004 240 NALT Scale Score 220 200 180 160 140 120 100 Grade 2 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Grade 9 2002 183.4 193.3 199.8 205.4 209.9 213.4 219.8 2003 183.5 192.9 199.3 205.7 209.7 213.5 218.8 2004 185.8 194.1 201.2 206.0 210.4 214.2 219.4 MBST Cut 183 190 197 204 209 213 With respect to the cut score, both Minneapolis reading and math 2002 averages were above the basic skills cut score except for math at grade 7 and increased in magnitude for 2004. Table 7. NALT Scale Score Which Predicts Passage of the Minnesota Basic Skills Tests (MBST) in 8 th Grade Grade Reading Scale Score Math Scale Score 2 183 191 3 190 199 4 197 206 5 204 214 6 209 218 7 213 225 8 218 232 Page 19Page 19

Figure 14. Average Math Scale Score on NALT 2002-2004 250 NALT Scale Score 200 150 100 50 0 Grade 2 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Grade 9 2002 192.0 202.5 208.5 215.4 218.2 222.4 231.0 2003 190.5 201.2 208.5 216.0 218.4 222.5 231.6 2004 193.6 203.0 210.8 217.8 219.0 224.2 231.7 MBST Cut 191 199 206 214 218 225 Figures 15 and 16 plot average scale scores for all 2004 test takers by ethnic group against the basic standard cut score that predicts success on the eighth grade MBST. The reading and math performance patterns were similar across the ethnic groups. For White students the average reading and math scale scores were considerably above the basic standard cut score. Generally speaking the average Asian students tended to be at the math and reading basic standard cut score and the remaining ethnic groups tended to be somewhat below the basic standard cut score at several grades. Mean Scale Score 250 240 230 220 210 200 190 180 170 160 Figure 15. Average Reading Scale Score on NALT by Ethnic Group Grade 2 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Indian African Asian Hispanic White MBST Cut The gap between Students of Color and White students was about the same from grade to grade and the gap is less pronounced in math than reading. With respect to the basic standard cut score, Students of Color were somewhat stronger in grades 3-5 than grades 2, 6, and 7. Overall, Page 20Page 20

Indian and Asian students tended to score higher than Hispanic and African students. Within each ethnic group, students in homes where English is spoken did much better than students in homes where English is not spoken. This is true for math and reading. The English vs. non- English difference is greater in math than reading. Mean Scale Score 250 240 230 220 210 200 190 180 170 160 Figure 16. Average Math Scale Score on NALT by Ethnic Group Grade 2 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Indian African Asian Hispanic White MBST Cut Mean scale scores for all English Language Learner (ELL) or Special Education students taking the NALT during the 2002-2003 school year are shown in Figures 17 and 18. ELL students consistently outscored Special Education students at all grades in both reading and math. In math, ELL is near or at the basic standard but below for other grades. In reading, both groups fell below the basic standard cut score that predicts passing the MBST and as one moves up the grades, they fall further below. 240 Figure 17. Average Reading Scale Score on NALT for English Language Learners and Special Education Students 230 Mean Scale Score 220 210 200 190 180 170 160 Grade 2 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Grade 9 English Language Learners Special Ed. Students MBST Cut Page 21Page 21

240 Figure 18. Average Math Scale Score on NALT for English Language Learners and Special Education Students 230 Mean Scale Score 220 210 200 190 180 170 160 Grade 2 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Grade 9 English Language Learners Special Ed. Students MBST Cut Growth of Continuously Enrolled Students Table 8 lists the expected yearly growth for grades 3-7 for both reading and math based on the 2002 NALT norms. The table documents much higher growth expectation for students who start out in the lower blocks of the national norms with smaller expected change at the upper blocks. Also included are the percent of students who made one or more year s growth in the national group. Table 8. NALT Expected One-Year Scale Score (SS) Growth by Grade & Scale Score Block, 8 of National Norms Grade 140-149 150-159 160-169 170-179 180-189 190-199 200-209 210-219 220-229 230-239 240-249 250-259 260-269 Natl Growth Reading 3 13.8 12.8 12.9 11.1 10.0 8.8 7.4 5.8 64% 4 11.0 11.2 9.4 7.9 6.8 5.6 4.4 2.8 0.0 61% 5 9.8 10.6 9.3 7.9 6.7 5.5 4.8 3.6 2.6 62% 6 9.8 8.2 7.4 6.1 5.0 4.3 3.1 2.1 0.0 58% 7 8.2 7.7 6.4 5.4 4.5 3.8 3.0 2.2 1.4 6 Math 3 14.3 11.9 11.3 10.6 10.1 8.8 7.8 65% 4 13.8 11.2 10.0 8.6 7.7 7.1 6.8 6.7 6.1 61% 5 11.1 9.6 8.0 7.5 7.3 7.4 7.6 7.0 6.6 61% 6 7.0 6.1 7.3 4.9 5.2 5.5 6.0 5.4 3.4 58% 7 7.8 6.2 4.8 5.0 5.6 5.8 6.4 6.1 5.0 3.2 6 The following graphs are based on continuously enrolled students (students who took NALT administered in both spring of 2003 and in spring of 2004). Figures 19-26 represent the percentage of students who made at least one year s scale score growth by grade, ethnic group, and special programs. What constitutes a year s growth is the gain in scale score from the previous year to the present year and is dependent on grade and a student s initial level of achievement. For example, in reading much of the gains are achieved in the first few Page 22Page 22

years of instruction. Therefore, second grade students need to make faster progress than third or fourth grade students in order to keep up with national norms and district standards. It is also true that students with low initial levels of reading performance need to make greater progress than students who start out well above average. Students continuously enrolled in Minneapolis Public Schools from spring 2003 to spring 2004 were compared to the national growth norms (Table 8) in order to determine if students made less than one year s growth, about one year s growth, or greater than one year s growth. Students with less than 8 of one year s growth were classified as less than one year s growth. Students with 80-12 of one year s growth were classified as making about one-year s growth, and students who made more than 12 of one year s growth were classified as making greater than one year s growth. Figures 19 and 20 compare the percent of students making about a year s growth or more in Minneapolis with the national growth norm by grade. Reading growth was good for the younger grades especially grade 4. There was a decline in growth for grades 5 and 6 with grade 7 remaining steady; in math the growth rates increased for grades 3 through 5, decreased significantly for grade 6 with a smaller decrease for grade 7. With respect to the national growth norm, there were some differences among the grades. At grade 4 the Minneapolis 2003 to 2004 growth was above the national norm in reading and math and above the national norm in math for grade 5. The remaining grades were below the national norm in both reading and math. 8 Figure 19. Percent of Students Making About One Year's Growth or More in NALT Reading by Grade Percent of Students 6 4 2 51% 52% 58% 55% 55% 55% 51% 51% 53% 53% Grade 2-3 Grade 3-4 Grade 4-5 Grade 5-6 Grade 6-7 MPS 2002 to 2003 MPS 2003 to 2004 NATIONAL NORM Page 23Page 23

8 Figure 20. Percent of Students Making About One Year's Growth or More in NALT Math by Grade Percent of Students 6 4 2 59% 61% 59% 6 51% 5 43% 36% 51% 49% Grade 2-3 Grade 3-4 Grade 4-5 Grade 5-6 Grade 6-7 MPS 2002 to 2003 MPS 2003 to 2004 NATIONAL NORM Growth of continuously enrolled Minneapolis ELL students was higher than Special Education students, especially in math (Figure 21). Reading gains for special education students was below the all district gain. For ELL students, growth was similar to the district for both reading and math and for Special Education students, growth was below the district in both subjects. 10 Figure 21. Percent of Students Making About One Year's Growth or More on NALT Reading and Math by Program Percent of Students 8 6 4 2 53% 53% 52% 53% 44% 43% 55% 55% All (n=13,308; 13,305) ELL (n=3,214; 3,227) Reading Spec Educ (n=1,593; 1,596) Math General Educ (n=8,501; 8,482) Figure 22 shows that more White, Asian, and Hispanic students made growth in both reading and math than Indian and African students. In Math, White students had the highest growth followed by Asian and Hispanic students. In both subject areas, Indian students had the lowest growth. Page 24Page 24

Figure 22. Percent of Students Making About One Year's Growth or More on NALT by Ethnicity 10 Percent of Students 8 6 4 2 45% 44% 47% 48% 55% 5 53% 51% 64% 61% Indian (n=578; 576) African (n=5412; 5404) Asian (n=1940; 1941) Reading Hispanic (n=1748; 1761) Math White (n=3629; 3622) Figures 23 and 24 illustrate for reading and math respectively, the percentage of students making at least one year s growth across both ethnicity and ELL services. For Indian students, there were fewer than twenty-five students having a non-english home language, so they are not included in the two figures. All ethnic groups receiving ELL services except Asian s, made more gain in reading than those not receiving ELL services. In math the reverse was true. All ethnic groups not receiving ELL services except African students made more gain than those receiving ELL services. For African s the discrepancy between ELL and non-ell was quite large for both subjects. 10 Figure 23. Percent of Students Making About One Year's Growth or More in Reading by Ethnicity and ELL Status Percent of Students 8 6 4 2 45% 46% 6 58% 47% 52% 54% 73% 64% Indian African Asian NonELL ELL Hispanic White Page 25Page 25

Figure 24. Percent of Students Making About One Year's Growth or More in Math by Ethnicity and ELL 10 Percent of Students 8 6 4 2 44% 47% 6 58% 54% 61% 54% 5 51% Indian African Asian Hispanic White NonELL ELL Figure 25 depicts NALT reading and math growth for the eight most predominant home languages in Minneapolis. Math growth exceeded reading growth for Arabic, Hmong, Laotian, and Spanish. The gains of Oromo, Spanish, and Vietnamese home language students was highest in reading while the gains Arabic, Laotian, and Spanish students was highest in math. 10 Figure 25. Percent of Students Making About One Year's Growth or More on Reading and Math by Home Language Percent of Students 8 6 4 2 57% 6 62% 57% 53% 53% 56% 52% 47% 47% 54% 51% 59% 6 58% 55% Arabic English Hmong Laotian Oromo Somali Reading Math Spanish Vietnamese Seven out of ten students in MPS received free or reduced price lunches. Figure 26 compares growth of students on free or reduced lunches. The growth of students who received free or reduced price lunch was less than students who did not receive free or reduced price lunches for both subjects. Page 26Page 26

Figure 26. Percent of Students Making at Least One Year's Growth on Reading and Math by Economic Status Percent of Students 10 8 6 4 2 48% 49% Eligible for Free/Reduced Lunch (n=9,130; 9,126) 63% 61% Not Eligible (n=4,177; 4,178) Reading Math Growth for Students at the Margins For the past several years, the Minneapolis Public Schools Research, Evaluation and Assessment Department (REA) has provided individual schools with an analysis of achievement growth at several points in the distribution. This analysis, known as the 20/20 analysis, shows the shift in the achievement for the middle scoring student (i.e., the median), the 20 th percentile student (at the level that separates the bottom 1/5 th of the students) and the 80 th percentile student (at the level that separates the top 1/5 th ) compared to national norms. The district wide 20/20 analysis is presented in Figures 27 and 28. The figures are based on the scores of all students who were continuously enrolled in grades 3-7 from 2003 to 2004. The three Nation lines on the graph represent the median, the 20 th, and the 80 th percentiles for the national norm group. 90 Figure 27. 20/20 Analysis for NALT Reading Growth Grades 3-7 n=12,524 Normal Curve Equivalent (NCE) 80 70 60 50 40 30 Nation Top 1/5 District Top 1/5 Nation Median District Median Nation Bottom 1/5 20 10 2003 2004 District Bottom 1/5 Page 27Page 27

Reading 20/20 Analysis 2003 2004 NCE Percentile NCE Percentile Top 1/5 63.9 74 64.2 75 Median 43.6 38 43.0 37 Bottom 1/5 27.2 14 27.2 14 In Figure 27 the district reading performance is somewhat below the national average for the high, median, and low levels of the distribution. All three levels remained fairly stable indicating that about one year s growth was made. 90 Figure 28. 20/20 Analysis for NALT Math Growth Grades 3-7 n=12,576 Normal Curve Equivalent (NCE) 80 70 60 50 40 30 Nation Top 1/5 District Top 1/5 Nation Median District Median Nation Bottom 1/5 20 10 2003 2004 District Bottom 1/5 2003 2004 NCE Percentile NCE Percentile Top 1/5 70.1 83 70.9 84 Median 50.5 51 50.0 50 Bottom 1/5 32.3 20 32.3 20 Figure 28 shows a very slight increase in performance in math for the top 2 of the distribution. The median and bottom 2 are stable indicating one year s growth. Longitudinal Analysis of Achievement Gaps Figures 29-32 show longitudinal trends in NALT scale scores for all students who were continuously enrolled in the district for five years (grades 3-7). The basic standard cut score (predicting passing on the eighth grade MBST) and a high standard cut score is shown. The high standard cut score is equivalent to a Minnesota Page 28Page 28

Comprehensive Assessments score of 1420, the approximate dividing line between Level 2 and Level 3 in the new 5 level configuration. Figures 29 and 30 show longitudinal trends in NALT scale scores for the 2003-2004 seventh grade students. In reading the scores are above the basic cut for all grades but starts below the high standard cut in the lower grades and catches up by grade 7. The same is true in math except by grade 5 the group catches up to the high standard cut score. 230 Figure 29. Longitudinal NALT Reading Gains Class of 2009 Mean Scale Score 220 210 200 190 High Standard Basic Standard Class of 2009 (n = 1991) 180 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Grade 8 Grade9 250 Figure 30. Longitudinal NALT Math Gains Class of 2009 Mean Scale Score 240 230 220 210 200 190 High Standard Basic Standard Class of 2009 (n =1986) 180 Grade 2 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Grade 8 Grade 9 Page 29Page 29

Figures 31 and 32 below, show reading growth of continuously enrolled students in the class of 2009 by ethnic group as compared to the basic and high standards. The average White student scored above the high standard at all grades in reading. Students of Color scored below the high standard for the past five years with Indian students above the basic standard for grades 5-7. Asian, Hispanic, and African students performed similarly for grades 3-5 with Hispanic performing similarly to Indian for grades 6 and 7. Mean Scale Score 240 230 220 210 200 190 Figure 31. Longitudinal NALT Reading Gains by Ethnic Group High Standard Basic Standard White Indian African 180 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Grade 8 Grade 9 Mean Scale Score 240 230 220 210 200 190 Figure 32. Longitudinal NALT Reading Gains by Ethnic Group High Standard Basic Standard White Asian Hispanic 180 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Grade 8 Grade 9 Five-year longitudinal growth in math for the class of 2009 is shown in Figures 33 and 34. White students are above high standard cut for all grades. Scores were lowest for African students in math across Page 30Page 30

all grades. Asian students had the highest average math scores for students of color across all grades except grade 3 and were above the basic standard cut in grades 4-7. 250 240 Figure 33. Longitudinal NALT Math Gains by Ethnic Group High Standard Mean Scale Score 230 220 210 200 190 Basic Standard White Indian African 180 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Grade 8 Grade 9 A very large dip in growth for Indian and African students was observed between 5 th and 6 th grade for NALT math. Overall gaps for African and Indian students increased markedly during the middle school years for this cohort of students. 250 240 Figure 34. Longitudinal NALT Math Gains by Ethnic Group High Standard Mean Scale Score 230 220 210 200 190 Basic Standard White Asian Hispanic 180 Grade 3 Grade 4 Grade 5 Grade 6 Grade 7 Grade 8 Grade 9 In contrast, Asian and Hispanic averages continued to grow sufficient to stay even (i.e. parallel lines on figure 34) with the state standards but did not accelerate versus the standards as White student averages. In other words, Asian and Hispanic gaps versus the standards did not increase but gaps compared to White student performance did increase. Page 31Page 31