Interpreting the Minnesota Comprehensive Assessment …

1 1 Office of Educational AccountabilityUniversity Technology Center1313 SE Fifth Street, Suite B-1 Minneapolis, MN 55414-4504tel (612)627-1009fax (612)627-1058 Minnesota schools and students havereceived their results on the MinnesotaComprehensive Assessments (MCA) forreading and mathematics in third and fifthgrade. The interpretive materials thataccompany the test contain descriptions offour performance levels (see sidebar). Eachstudent will receive a report showing theirlevel of performance, along with a statepercentile rank showing how the studentcompared to other students who took thetest. Neither the description of performancelevels nor the state percentile rank providesinformation about how Minnesota studentscompare to students around the year (1997-98), many students tookboth the MCA and a nationally standardizednorm referenced test.

2 Staff in some districtsused the data to approximate the nationalpercentile ranks corresponding to each MCAlevel. While only approximate, the districtsfound this national information helped themunderstand the MCA levels. Using dataprovided by members of the MinnesotaAssociation of Educational Assessment andEvaluation (MAEAE), we have pulledtogether data from several districts usingvarious standardized tests. The data approxi-mate the national percentile ranks coveredby the MCA levels. We hope that thisinformation can supplement the interpretivetools provided with the test itself. EstimatesInterpreting the Minnesota Comprehensive AssessmentLevels: A Link to National Percentile Ranksare based on the 1998 MCA test information presented requiresseveral cautions. First, the estimatednational percentile ranks associated withthe MCA levels vary across school districtsand norm referenced tests.

3 For this reason,the final estimates can only be consideredapproximate. Second, not all standardizedtests and districts in the state are repre-sented by this effort; the test used by yourdistrict may not be included. Third, theestimates were obtained using the equipercentile method described in thebox on page 4. Other methods wouldproduce somewhat different , the Northwest Evaluation Associa-tion Achievement Levels Test (NWEA) hasbeen included. The norm group for theNWEA is not necessarily representative ofthe nation s student population. It is a user norm group composed of studentstaking the NWEA, and may not be repre-sentative of students nationally. Theremaining tests the California Achieve-ment Test (CAT/5), the Iowa Test of BasicSkills (ITBS), and the MetropolitanAchievement Test (MAT/7) all usenationally representative norm state percentile ranks provided onthe MCA reports indicate how a student sperformance compares to that of otherAchievement LevelsAchievement levelsdescribe Minnesotastudent progresstoward the state sHigh Standards inreading, mathematicsand for fifth-graders, IV: Studentsdemonstrate superiorperformance, wellbeyond what is expec-ted at the grade III: Students areworking above gradelevel.

4 Many are profi-cient with challengingsubject II: Moststudents in Minne-sota fall within thislevel. This includes awide range of stu-dents, from thosewith partial knowledgeand skills to studentswho are increasinglyproficient with gradelevel I: Studentshave gaps in theknowledge and skillsnecessary forsatisfactory 1999 Vol. 1, No. 12students in Minnesota . The MCAlevels provide a more qualitativedescription of the student s perfor-mance. The descriptions of thelevels provided with test materialsconstitute the major informationwith which to understand the national percentile rank esti-mates provided in this reportconstitute additional information tosupplement our understanding ofthose AND CONCLUSIONSThe table above shows our basicfindings. The first three columnsshow the lowest MCA score inLevels 2, 3, and 4 for reading andmathematics in grades 3 and 5.

5 Thenext eight columns show the eightestimates of the correspondingnational percentile rank. Eachestimate represents a different schooldistrict or collection of districts. Thelast column ( Median NPR Esti-mate ) is the median of the eightschool district median NPR estimates areapproximate estimates of thenational percentile rank where eachlevel begins. For instance, in thirdgrade reading, the median NPRestimates suggest that Level 2 beginsat about the 31st national percentile,Level 3 at the 75th national percen-tile, and Level 4 at the 97th the NWEA norm groupis not representative of the nation asa whole, one could argue that theNWEA data should not have beenused in obtaining the median NPRestimates. However, including theNWEA provided additional data,and seldom changed the final NPRestimate by as much as 3 points.

6 Inthose cases where it changed theresult by more than 3 points, theestimate obtained by including theNWEA is given in bold; the estimateexcluding the NWEA is given inplain the median NPR estimatesin the table, the figure approximatesthe range of national percentileranks associated with each MCAlevel. To see the relationshipsbetween the table and the figure,consider third grade reading. Themedian national percentile rankestimates for the lowest scores inLevels 2, 3, and 4 are 31, 75, and 97respectively. Therefore, in the figure,Level 2 begins at the 31st percentile,Table 1. National Percentile Rank Estimates Corresponding to the Lowest Score in MCA Levels 2, 3, and 4 Based on Eight Data Sets from One or Several DistrictsGrade/MCAL owestCAT5 ITBSMAT7 NWEAM edianContent AreaLevelScoreNPR Est(1)(2) (3) (4) (5) (6) (7) (8) Lowest ScoreSee page 4 for notes.

7 2 1280 32 20 24 33 26 33 31 37 31/263 1500 80 61 67 83 80 73 76 74 75/804 1690 98 87 95 99 98 95 97 98 97 Grade 3 ReadingNational PercentileRank EstimatesGrade 3 Math-ematics NationalPercentile RankEstimatesGrade 5 ReadingNational PercentileRank EstimatesGrade 5 Math-ematics NationalPercentile RankEstimates2 121027 18 23 33 29 29 29 41 293 1500 80 65 73 83 80 68 79 86 794 1740 97 93 98 97 97 96 96 98 972 1260 25 27 27 31 29 31 29 30 293 1500 68 66 68 77 74 74 74 73 73/684 171096 91 92 96 97 94 97 98 962 1240 32 21 32 45 30 29 29 44 313 1500 86 67 79 90 87 83 81 86 854 171099 93 97 99 98 97 98 98 983 About the dataData provided by a consortium ofdistricts through the StatewideTesting Program at the University ofMinnesota.

8 Districts include: Ada-Borup, Adrian, Atwater-Grove City-Cosmos, Badger, Bagley, Barnum,Bellingham, Blue Earth, Centennial(Circle Pines), Dover-Eyota,Edgerton, Ely, Fulda, GrandMeadow, Heron Lake-Okabena,Hinkley, Holdingford, Lanesboro,LaPorte, LeSuer, Marshall County(Newfolden), Melrose, Minneota,Minnewaska, Moose Lake, Moun-tain Iron, Mountain Lake, NorthBranch, North St. Paul-Maplewood,Osakas, Pequot Lakes, Pierz, RedRock Central (Storden-Jeffers),Redwood Falls, Remer, Roseau,Ruthton, St. Francis, Sartell, SleepyEye, South Koochiching, Southland(Rose Creek), Truman, andWaconia. The NPRs are taken fromState MCA/ITBS Tests: Birds of aFeather? by Charles B. numbers of students rangedfrom 3,176 to 3,200 depending onthe grade. The correlation betweenthe MCA scales and the ITBS scalesranged from.

9 76 to . Achievement Test (CAT5)data:Provided by the Wayzata SchoolDistrict (the Reading Comprehensionand Mathematics Concepts andapplications subscales). The numberof students ranged from 652 to 697depending on the grade and correlations between the MCAand CAT5 scales ranged and . data (based on the 1992 norms):2 Data provided by the Anoka-Hennepin district. The number ofstudents ranged from 2879 to 2945depending on the grade and subjectarea. Correlations between MCAand ITBS scales ranged from .66 provided by the St. PaulPublic Schools. The number ofstudents in the data set ranged from2887 to 3388 depending on thegrade and subject area. Correla-tions between the MCA and theMAT7 ranged from .82 to . results (based on 1992 norms):Results provided by the DuluthPublic Schools. The number ofstudents varied from 825 to 848depending on the grade and subjectarea.

10 Correlations between theMCA and MAT7 ranged and NWEA 1996 norms (based on auser sample that may not be representa-tive of students in the nation s schools).The three districts are Moundsview (6),Robbinsdale (7), and Minneapolis (8).In Moundsview the number ofstudents varied from 779 to 834,depending on the grade and subjectarea. The correlations between theMCA and NWEA scores rangedbetween .79 and .86 depending onthe grade and subject Robbinsdale, the number ofstudents varied from 1001 to correlations between the MCAand NWEA ranged between .80 Minneapolis, the number of stu-dents varied from 3289 to correlations between the MCAand NWEA ranged from .87 to . 1 Level 2 Level 4 Level 3 Level 4 Level 4 Level 2 Level 2 Level 2 Level 1 Level 1Gr. 3 ReadingGr. 3 MathematicsGr. 5 ReadingGr.