College Board Research

Transcription

1 College Board Research June 2, 2016 Concordance Tables for the New and Old SAT As part of determining that scores from the new SAT are valid for intended uses, College Board used equipercentile methods to link scores from the new SAT with scores from the old SAT. In keeping with AERA/APA/NCME Standard 5.18, which states that when linking procedures are used to relate scores on tests or test forms that are not closely parallel, the construction, intended interpretation, and limitations of those linkings should be described clearly (AERA/APA/NCME, 2014, p. 106), this document discusses the concordance methods used to link the new SAT to the old SAT. The majority of the material in this document was initially produced as a chapter for a technical manual that describes the redesign of the SAT in detail. However, it is reformatted and slightly edited here, so that it may be presented as a single self-contained document. The material in this document will be rereleased when the technical manual is published at a later date. Background and Purpose of the Concordance Study While redesigning the SAT, the College Board put forth significant efforts to examine score comparability, test equitability, and scale continuity issues. We also made mindful and deliberate efforts in communicating the changes between the old SAT and the new SAT to various stakeholders and the general public during the redesign process (College Board, 2015). Due to the nature and the scope of the differences between the new SAT and the old SAT, using the term equating is not appropriate, as measures from the old and the new SAT do not result in scores that can be made interchangeable. In order to supply a table that relates the scores from the two measures under these circumstances, a procedure known as concordance is performed (e.g., Dorans, 2000; Kolen & Brennan, 2004). Two concordance studies were conducted as part of the assessment redesign project, an early preliminary study and a second final study. This document describes the methods and results from the final concordance study. The concordance tables were prepared based on the internal College Board research for various intended uses including benchmark research, score reporting, and predictive validity. They are also used by external stakeholders including, but not limited to, high school counselors, higher education, and the National Collegiate Athletic Association (NCAA) for score reporting and other purposes. The SAT concordance tables are also intended to be used as concordance for the new and old scores of PSAT10. 1

2 Methods Concordance Study Design and Methodology Concordance data were collected using the new SAT and the October, November, and December 2015 administrations of the old SAT. Using a single group design, one sample of participants completed an intact form from both measures (e.g., old SAT and new SAT). Various methodologies for linking exist as possible options for selecting a concordance table solution (e.g., equipercentile, linear, with smoothing or without smoothing [Dorans, 2000; Kolen & Brennan, 2004]). The ideal approach, which was used here, is to examine multiple solutions and select the most appropriate one. The criteria for deciding which concordance table to choose include coverage of the full area score scale (e.g., 200 to 800) and minimizing gaps between scores (.e.g., scale scores skip from 680 to 700). Using the equipercentile scaling method with cubic spline post-smoothing, the percentile ranks of the scale score points from the new SAT and the old SAT are linked to one another to determine which scores correspond from the two forms. For this study, 11th and 12th graders completed both the new SAT and the old SAT and then analyses were conducted to produce concordance tables linking the scores across these two forms. Instruments The two instruments used in the concordance study were the old SAT and the new SAT. Specifically, examinees who completed the October 2015, November 2015, or December 2015 administration of the old SAT were recruited to take the new SAT on December 9th, This was a single group design where students who took the new SAT on December 9th also had a previous old SAT score from October, November, or December If students completed more than one of the three old SAT administrations, only their scores from the most recent administrations were kept for the purpose of this study. The new SAT is comprised of five types of scores: (1) the Total score, (2) two Section scores, (3) two Cross-Test scores, (4) three Test Scores, and (5) seven subscores (for the specific concordances produced, see Table 1 in the Appendix). Concordances were not created for the Cross- Test scores or the subscores. Survey questions that ask about motivation level and the necessary demographics were administered as part of the answer sheets. Participants The sampling plan targeted 50% 11th and 50% 12th graders. A high-level sampling plan with target specifications for region, location, public vs. private, grade level, and percent underrepresented minority students was also created and intended to represent a typical SAT cohort. All schools had to have sufficient training on administering secure assessments. Students with disabilities were allowed to participate in the study as well, but were not included in the analyses. The new SAT administration in December 2015 was conducted as an operational administration, and the scores students received from this administration were reportable. The target population was a typical SAT cohort, defined as the average of the old SAT cohorts. The cleaned unweighted dataset was comprised of N = 8,677 examinees (see Table 2 in the Appendix for the score 2

3 correlations and Tables 3 through 9, also in the Appendix, for more detailed descriptions of the unweighted data). A weighted version of this data was created to better represent the target population. Specifically, the 8,677 examinees case weights were obtained, which resulted in a weighted dataset that approximated recent SAT cohorts in terms of percentages of demographic variables such as grade, gender, ethnicity, region, best language, and first language, as well as the means and standard deviations of the old SAT scale scores (see Table 2 in the Appendix for the score correlations and Tables 10 through 16, also in the Appendix, for more detailed descriptions of the weighted data). Analysis, Methodology, and Results The LEGS (Linking with Equivalent Groups or Single Group Design) software written by the Center for Advanced Studies on Measurement and Assessment (CASMA) (Brennan, 2004) was used in the concordance analysis, with the specific analytic steps described below: Step 1: Data Cleaning and Screening Four data files were received from the technical operational group: three for the old SAT (one from each administration) and one for the new SAT. The data files had been matched in the system prior to being delivered to psychometricians, so the same students were in both the old and the new SAT assessment data files. After receiving the files, College Board psychometricians merged the files by the person_id identifier. The original matched sample with only the most recent old SAT scores contained 8,776 students. The sample of N=8,677 was obtained after applying the additional data screenings as listed below: 1. Removed SSD answer sheets (N = 64). 2. Removed students indicated by the student irregularity reports as having irregularities (N = 23) 3. Removed 10th graders (N = 9) 4. Removed examinees who skipped the entire Reading, Math, or Writing tests of the new SAT (i.e., operational screening rules, N=3). Step 2: Determine the Representativeness of the Sample and Weight the Sample Appropriately At this point, the cleaned sample was ready to be analyzed for the concordance. The next step was to determine if this sample closely approximated the target population, and if not, to determine a weighted sample that would be a better approximation of the target population. As mentioned previously, the uncleaned, unweighted dataset was weighted by the variables of interest that were deemed practically meaningful. The means and standard deviations of the old SAT scores were examined before and after the sample data were weighted. As expected, the weighted dataset was a closer approximation of the target population, and was used as input data for all correlation and concordance analyses. 3

4 Step 3: Correlation Analyses Table 2 in the Appendix shows the correlations between the score pairs of interest for this concordance study. The lower diagonal is based on the unweighted data, and the upper diagonal is based on the weighted data (which was used as inputs into the concordance analyses). Any cell highlighted in blue shows a correlation between two score pairs that were used to create a concordance table. Step 4: LEGS Analyses and the Resulted Concordance Solution for SAT All concordance analyses were done using the Linking with Equivalent Groups or Single Groups (LEGS) software program. For each score pair shown in Table 1 in the Appendix, two versions of the concordance tables are provided: 1. The old SAT to new SAT concordance table. 2. The new SAT to old SAT concordance table The unidirectional old SAT to new SAT and new SAT to old SAT tables were obtained by conducting two separate LEGS runs. The old SAT to new SAT tables include every possible old SAT score and provide a concorded new SAT score for each. The new SAT to old SAT tables include every possible new SAT score and provide a concorded old SAT score for each. Note that while each of these tables contains all possible values for the from score (e.g., the old SAT score in the old SAT to new SAT table), because of differences in number of score points on the scales compared and the resulting gaps and many-to-one concordances, each table does not contain all score points for the concorded to score (e.g., the new SAT score in the old SAT to new SAT table). Cubic spline post-smoothing was used for all concordance tables. All the concordance tables and their plots are provided in Tables 17 through 28 and Figures 1 through 12, all of which can be found in the Appendix. Several criteria were considered when determining the recommended final concordance table, specifically: 1. Top of the scale must be represented 2. Bottom of the scale must be represented 3. Many-to-one conversions should be minimized 4. Score gaps should be minimized As implemented in LEGS, the cubic spline post-smoothing method is replaced with linear interpolations for the lowest and highest scores with sparse data (usually 0.5%, meaning that linear interpolations are applied for from scores with percentile ranks less than 0.5 and above =99.5 [Kolen & Brennan, 2004]). The extent of interpolation can be manipulated, and was selected to achieve a minimal number of many-to-one conversions in each concordance table. Specifically, the extent of interpolation for the six concordance score pairs was determined as the following %s: 4

5 Math of the old SAT to Math of the new SAT concordance: 2% Critical Reading of the old SAT to Evidence-based Reading of the new SAT concordance: 2% Writing of the old SAT to Writing and Language of the new SAT concordance: 2% Critical Reading + Writing of the old SAT to Evidence-based Reading and Writing of the new SAT concordance: 0.5% Critical Reading + Math of the old SAT to Total of the new SAT concordance: 2% Critical Reading + Math + Writing of the old SAT to Total of the new SAT concordance: 1.5% interpolation ACT-SAT Concordance Tables An additional set of concordance tables that did not involve any new data collection was produced by merging with the existing ACT-SAT concordance tables (College Board, 2009). Specifically, the results of the final concordance tables from this December 2015 concordance study were leveraged to connect the existing ACT-SAT concordance tables to the new SAT scale scores. In other words, these concordance tables are derived concordances between the ACT and the new SAT. There are two sets of the existing concordance tables, one for the ACT composite and the old SAT total score (defined as Critical Reading plus Math), and a second for the ACT Writing (pre-september 2015, before the recent ACT Writing redesign) and the old SAT Writing. These new derived concordance tables are extensions of the existing tables with the intention of adding the new SAT. However, only single score points are included (whereas score ranges were in the existing concordance tables for the old SAT). Also, the newly derived concordance tables are shown in two directions ACT to new SAT and new SAT to ACT. There are four tables: (1) ACT Composite to new SAT Total (2) ACT Writing from pre-september 2015 to new SAT Writing and Language, (3) New SAT total to ACT Composite and (4) New SAT Writing and Language test to ACT Writing from pre-september These tables are shown in Tables 29 through 32 in the Appendix. For lower score points on these tables, there was not enough valid data to produce a valid concordance between the new SAT and ACT. Ideally, a new concordance study for the ACT and new SAT will be conducted in the future, to update the previous study with the new SAT scores, using actual student data. In the interim, to assist users during this transition phase to the new SAT scale, these derived concordance tables were created. SAT Concordance Table Implications and Cautions The concordances should be interpreted with the following cautions: 1. Concordance tables are sample-dependent. The sample that was used to create these tables took the old SAT in October, November, or December 2015, followed by the new SAT in a special administration on December 9 th, However, we weighted the sample according to demographic characteristics of recent SAT cohorts, so the sample used for the concordance study represents this target population, although the data was collected in the admin-specific samples. 5

6 2. Concordances are needed for students who take either the old SAT or the new SAT. An order effect was unavoidable in the data collection design, and thus the concordance datasets were assembled from students who took the old SAT followed by the new SAT. It was not possible to counterbalance the order of administration, or more ideally have an equivalent groups design with each student taking only one of the tests being concorded. Thus, interpretation of these tables should be done cautiously, keeping in mind the order effects resulting from the data collection limitation. 3. Additional invariance evaluations will be conducted to assess the accuracy of these concordances for examinees in several demographic subgroups. 6

7 References American Educational Research Association, American Psychological Association, National Council on Measurement in Education (AERA, APA, NCME) (2014). Standards for educational and psychological testing. Washington, DC: American Educational Research Association American Psychological Association, National Council on Measurement in Education. Brennan, R. L. (2004). Manual for LEGS Version 2.0 (CASMA Research Report No. 3). Iowa City, IA: Center for Advanced Studies in Measurement and Assessment, The University of Iowa (Available from the web address: College Board. (2009). ACT and SAT concordance tables (RN-40). New York: The College Board. College Board. (2015). Test Specifications for the redesigned SAT. New York: The College Board. Dorans, N. J. (2000). Distinctions among classes of linkages. College Board Research Note (RN-11). New York: The College Board. Kolen, M. J., & Brennan, R. L. (2004). Test equating, scaling, and linking, Second Edition. New York: Springer. 7

8 Appendix Table 1 Old SAT to New SAT Old SAT New SAT Math ( ) Math Section Score ( ) Critical Reading ( ) Reading Test Score (10-40) Writing ( ) Writing and Language Test Score (10-40) Critical Reading + Writing ( ) Evidence-Based Reading and Writing (ERW) Section Score ( ) Critical Reading + Math ( ) Total Score (ERW + Math Section; ) Critical Reading + Math + Writing ( ) Total Score (ERW + Math Section; ) 8

9 Table 2 Correlations between Old SAT and New SAT score pairs for concordances (Lower diagonal from unweighted sample, Upper diagonal from weighted sample) Old M Old CR Old W Old CR+W Old M + CR Old M + CR + W New M New R New W&L New ERW New Total Old M Old CR Old W Old CR + W Old M+ CR Old M + CR + W New M New R New W & L New ERW New Total M=Math; R=Reading; W=Writing; W&L = Writing and Language; CR = Critical Reading; ERW = Evidence-Based Reading and Writing 9

10 Table 3 Demographic Characteristics of the Unweighted SAT Concordance Data Overall Descriptive Statistics for the Unweighted Total Group (N =8,677) Subject New SAT Mean (SD) Corr. Old SAT Mean (SD) Old SAT Mean (SD) (Cohort) Reading (5.16) (101.21) (114.50) Math (97.96) (104.64) (118.00) Writing (5.12) (98.97) (114.00) 10

11 Table 4 Demographic Characteristics of the Unweighted SAT Concordance Data Weighted Group/ Subgroup N Percent Percent Grade Level (New SAT Response) Reading Scores New SAT Mean Corr. (SD) or Std.Diff Total (5.158) (101.21) (114.50) 11th Grade th Grade Weighted Group/ Subgroup N Percent Percent Math Scores New SAT Mean Corr. (SD) or Std.Diff Total (97.96) (104.64) (118.00) 11th Grade th Grade Weighted Group/ Subgroup N Percent Percent Writing Scores New SAT Mean Corr. (SD) or Std.Diff Total (5.116) (98.97) (114.00) 11th Grade th Grade *Cohort results are averages of the Cohort. 11

12 Table 5 Weighted Group/ Subgroup Demographic Characteristics of the Unweighted SAT Concordance Data Gender (New SAT Response) N Percent Percent Reading Scores New SAT Mean Corr. (SD) or Std.Diff Total (5.158) (101.21) (114.50) F M Weighted Group/ Subgroup N Percent Percent Math Scores New SAT Mean Corr. (SD) or Std.Diff Total (97.96) (104.64) (118.00) F M Weighted Group/ Subgroup N Percent Percent Writing Scores New SAT Mean Corr. (SD) or Std.Diff Total (5.116) (98.97) (114.00) F M *Cohort results are averages of the Cohort. 12

13 Table 6 Demographic Characteristics of the Unweighted SAT Concordance Data First Language (New SAT Response) Weighted Group/ Subgroup Reading Scores N Percent Percent New SAT Mean Corr. (SD) or Std.Diff Total (5.158) (101.21) (114.50) No Response English Only English and Another language Another language Weighted Group/ Subgroup N Percent Percent Math Scores New SAT Mean Corr. (SD) or Std.Diff Total (97.96) (104.64) (118.00) No Response English Only English and Another language Another language Weighted Group/ Subgroup N Percent Percent Writing Scores New SAT Mean Corr. (SD) or Std.Diff Total (5.116) (98.97) (114.00) No Response English Only English and Another language Another language

14 Table 7 Weighted Group/ Subgroup Demographic Characteristics of the Unweighted SAT Concordance Data N Percent Percent Best Language (New SAT Response) Reading Scores New SAT Mean Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (5.158) (101.21) (114.50) No Response English Only English and Another language Another language Weighted Group/ Subgroup N Percent Percent Math Scores New SAT Mean Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (97.96) (104.64) (118.00) No Response English Only English and Another language Another language Weighted Group/ Subgroup N Percent Percent Writing Scores New SAT Mean Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (5.116) (98.97) (114.00) No Response English Only English and Another language Another language

15 Table 8 Demographic Characteristics of the Unweighted SAT Concordance Data Ethnicity Weighted Group/ Subgroup Reading Scores N Percent Percent New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff Total (5.158) (101.21) (114.50) I do not wish to respond American Indian or Alaska Native Asian, Asian American, or Pacific Islander Black or African American Mexican or Mexican American Puerto Rican Other Hispanic, Latino, or Latin American White Other Weighted Group/ Subgroup Math Scores N Percent Percent New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff Total (97.96) (104.64) (118.00) I do not wish to respond American Indian or Alaska Native Asian, Asian American, or Pacific Islander Black or African American Mexican or Mexican American Puerto Rican Other Hispanic, Latino, or Latin American White Other

16 Weighted Group/ Subgroup Writing Scores N Percent Percent New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff Total (5.116) (98.97) (114.00) I do not wish to respond American Indian or Alaska Native Asian, Asian American, or Pacific Islander Black or African American Mexican or Mexican American Puerto Rican Other Hispanic, Latino, or Latin American White Other

17 Table 9 Weighted Group/ Subgroup Demographic Characteristics of the Unweighted SAT Concordance Data Region (New SAT Response) N Percent Percent Reading Scores New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff Total (5.158) (101.21) (114.50) Missing MRO MSRO NERO SRO SWRO WRO Weighted Group/ Subgroup N Percent Percent Math Scores New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff Total (97.96) (104.64) (118.00) Missing MRO MSRO NERO SRO SWRO WRO

18 Weighted Group/ Subgroup N Percent Percent Writing Scores New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff Total (5.116) (98.97) (114.00) Missing MRO MSRO NERO SRO SWRO WRO

19 Table 10 Demographic Characteristics of the Weighted SAT Concordance Data Overall Descriptive Statistics for the Weighted Total Group Subject New SAT Mean (SD) Corr. Old SAT Mean (SD) Old SAT Mean (SD) (Cohort) Reading (5.37) (108.83) (114.50) Math (104.36) (112.55) (118.00) Writing (5.37) (107.67) (114.00) 19

20 Table 11 Demographic Characteristics of the Weighted SAT Concordance Data Grade Level (New SAT Response) Weighted Group/ Subgroup Percent Percent Reading Scores New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff Total (5.368) (108.83) (114.50) 11th Grade th Grade Weighted Group/ Subgroup Percent Percent Math Scores New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (104.36) (112.55) (118.00) 11th Grade th Grade Writing Scores Weighted Group/ Subgroup Percent Percent New SAT Mean (SD) or Std.Diff. Corr. Old SAT Mean (SD) or Std.Diff. (SD) or Std.Diff Total (5.369) (107.67) (114.00) 11th Grade th Grade *Cohort results are averages of the Cohort. 20

21 Table 12 Weighted Group/ Subgroup Demographic Characteristics of the Weighted SAT Concordance Data Gender (New SAT Response) Percent Percent Reading Scores New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff Total (5.368) (108.83) (114.50) F M Weighted Group/ Subgroup Percent Percent Math Scores New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (104.36) (112.55) (118.00) F M Writing Scores Weighted Group/ Subgroup Percent Percent New SAT Mean (SD) or Std.Diff. Corr. Old SAT Mean (SD) or Std.Diff. (SD) or Std.Diff Total (5.369) (107.67) (114.00) F M *Cohort results are averages of the Cohort. 21

22 Table 13 Demographic Characteristics of the Weighted SAT Concordance Data First Language (New SAT Response) Weighted Group/ Subgroup Percent Percent Reading Scores New SAT Mean Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (5.368) (108.83) (114.50) No Response English Only English and Another language Another language Weighted Group/ Subgroup Percent Percent Math Scores New SAT Mean Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (104.36) (112.55) (118.00) No Response English Only English and Another language Another language Weighted Group/ Subgroup Percent Percent Writing Scores New SAT Mean Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (5.369) (107.67) (114.00) No Response English Only English and Another language Another language

23 Table 14 Weighted Group/ Subgroup Demographic Characteristics of the Weighted SAT Concordance Data Percent Best Language (New SAT Response) Percent Reading Scores New SAT Mean Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (5.368) (108.83) (114.50) No Response English Only English and Another language Another language Weighted Group/ Subgroup Percent Percent Math Scores New SAT Mean Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (104.36) (112.55) (118.00) No Response English Only English and Another language Another language Weighted Group/ Subgroup Percent Percent Writing Scores New SAT Mean Corr. (SD) or Std.Diff. (SD) or Std.Diff Total (5.369) (107.67) (114.00) No Response English Only English and Another language Another language

24 Table 15 Demographic Characteristics of the Weighted SAT Concordance Data Ethnicity Weighted Group/ Subgroup Percent Reading Scores Percent New SAT Mean (SD) or Corr. (SD) or Std.Diff Std.Diff. Total (5.368) (108.83) (114.50) I do not wish to respond American Indian or Alaska Native Asian, Asian American, or Pacific Islander Black or African American Mexican or Mexican American Puerto Rican Other Hispanic, Latino, or Latin American White Other Weighted Group/ Subgroup Percent Math Scores Percent New SAT Mean (SD) or Corr. (SD) or Std.Diff Std.Diff. Total (104.36) (112.55) (118.00) I do not wish to respond American Indian or Alaska Native Asian, Asian American, or Pacific Islander Black or African American Mexican or Mexican American Puerto Rican Other Hispanic, Latino, or Latin American White Other

25 Weighted Group/ Subgroup Percent Writing Scores Percent New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff Total (5.369) (107.67) (114.00) I do not wish to respond American Indian or Alaska Native Asian, Asian American, or Pacific Islander Black or African American Mexican or Mexican American Puerto Rican Other Hispanic, Latino, or Latin American White Other

26 Table 16 Demographic Characteristics of the Weighted SAT Concordance Data Region (New SAT Response) Weighted Group/ Subgroup Percent Reading Scores Percent New SAT Mean Corr. (SD) or Std.Diff Total (5.368) (108.83) (114.50) Missing MRO MSRO NERO SRO SWRO WRO Weighted Group/ Subgroup Percent Percent Math Scores New SAT Mean Corr. (SD) or Std.Diff Total (104.36) (112.55) (118.00) Missing MRO MSRO NERO SRO SWRO WRO

27 Weighted Group/ Subgroup Percent Percent Writing Scores New SAT Mean (SD) or Std.Diff. Corr. (SD) or Std.Diff Total (5.369) (107.67) (114.00) Missing MRO MSRO NERO SRO SWRO WRO

28 Table 17 Old SAT to New SAT Concordance Table for Math Scores Old SAT ( ) New SAT ( )

29 Old SAT ( ) New SAT ( )

30 Figure 1 Old SAT to New SAT Concordance Plot for Math Scores New SAT Math Old SAT Math 30

31 Table 18 New SAT to Old SAT Concordance Table for Math Scores New SAT ( ) Old SAT ( )

32 New SAT ( ) Old SAT ( )

33 Figure 2 New SAT to Old SAT Concordance Plot for Math Scores Old Math New SAT Math 33

34 Table 19 Old SAT to New SAT Concordance Table for Reading Scores Old SAT ( ) New SAT (10-40)

35 Old SAT ( ) New SAT (10-40)

36 Figure 3 Old SAT to New SAT Concordance Plot for Reading Scores New SAT Reading Old Critical Reading 36

37 Table 20 New SAT to Old SAT Concordance Table for Reading Scores New SAT (10-40) Old SAT ( )

38 Figure 4 New SAT to Old SAT Concordance Plot for Reading Scores Old Critical Reading New SAT Reading 38

39 Table 21 Old SAT to New SAT Concordance Table for Writing Scores Old SAT ( ) New SAT (10-40)

40 Old SAT ( ) New SAT (10-40)

41 Figure 5 Old SAT to New SAT Concordance Plot for Writing Scores 45 New SAT Writing and Language Old SAT Writing 41

42 Table 22 New SAT to Old SAT Concordance Table for Writing Scores New SAT (10-40) Old SAT ( )

43 Figure 6 New SAT to Old SAT Concordance Plot for Writing Scores

44 Table 23 Old SAT to New SAT Concordance Table for Combined Reading and Writing Scores Old SAT ( ) New SAT ( )

45 Old SAT ( ) New SAT ( )

46 Old SAT ( ) New SAT ( )

47 Old SAT ( ) New SAT ( )

48 Figure 7 Old SAT to New SAT Concordance Plot for Combined Reading and Writing Scores New SAT Evidence-Based Reading and Writing Old SAT Critical Reading + Writing 48

49 Table 24 New SAT to Old SAT Concordance Table for Combined Reading and Writing Scores New SAT ( ) Old SAT ( )

50 New SAT ( ) Old SAT ( )

51 Figure 8 New SAT to Old SAT Concordance Plot for Combined Reading and Writing Scores 1800 Old SAT Critical Reading + Writing New SAT Evidence-Based Reading and Writing 51

52 Table 25 Old SAT to New SAT Concordance Table for Old SAT Combined Reading and Math Scores to New SAT Total Scores Old SAT ( ) New SAT ( )

53 Old SAT ( ) New SAT ( )

54 Old SAT ( ) New SAT ( )

55 Old SAT ( ) New SAT ( )

56 Figure 9 Old SAT to New SAT Concordance Plot for Old SAT Combined Reading and Math Scores to New SAT Total Scores New SAT Total Old SAT Critical Reading + Math 56

57 Table 26 New SAT to Old SAT Concordance Table for New SAT Total Scores to Old SAT Combined Reading and Math Scores New SAT ( ) Old SAT ( )

58 New SAT ( ) Old SAT ( )

59 New SAT ( ) Old SAT ( )

60 New SAT ( ) Old SAT ( )

61 Figure 10 New SAT to Old SAT Concordance Plot for New SAT Total Scores to Old SAT Combined Reading and Math Scores 1800 Old SAT Critical Reading + Math New SAT Total 61

62 Table 27 Old SAT to New SAT Concordance Table for Total Scores Old SAT ( ) New SAT ( )

63 Old SAT ( ) New SAT ( )

64 Old SAT ( ) New SAT ( )

65 Old SAT ( ) New SAT ( )

66 Old SAT ( ) New SAT ( )

67 Figure 11 Old SAT to New SAT Concordance Plot for Total Scores New SAT Total Old SAT Critical Reading + Writing + Math 67

68 Table 28 New SAT to Old SAT Concordance Table for Total Scores New SAT ( ) Old SAT ( )

69 New SAT ( ) Old SAT ( )

70 New SAT ( ) Old SAT ( )

71 New SAT ( ) Old SAT ( )

72 Figure 12 Appendix H-4: New SAT to Old SAT Concordance Plot for Total Scores Old SAT Critical Reading + Writing + Math New SAT Total 72

73 Table 29 ACT Composite to SAT Total ACT Composite Score New SAT Total ( )

74 Table 30 ACT Writing (pre-2015) to SAT Writing and Language Test Score ACT English/Writing Score New SAT Writing and Language (10-40)

75 Table 31 SAT Total to ACT Composite New SAT Total ( ) ACT Composite Score

76 New SAT Total ( ) ACT Composite Score

77 New SAT Total ( ) ACT Composite Score

78 Table 32 SAT Writing and Language Test Score to ACT Writing (pre-2015) New SAT Writing and Language (10-40) ACT English/Writing Score