INTRODUCTION TO TEST THEORY Education 252 Spring 2004 MWF 9:00-10:50, School of Education Building, room 206 Instructor: Edward Haertel haertel@stanford.edu office: e331 650-725-1251 home: 650-494-6432 (Office hours by arrangement, individual or small group) This course introduces classical test theory, including definitions and formulas for test reliability, standard error of measurement, and related statistics. Additional topics include test validity, item statistics useful in test construction, score scales and norms commonly used in educational testing, item bias and test bias, and ideas of fairness and equity in educational and psychological testing. Factor analysis as well as the major extensions and alternatives to classical test theory, generalizability theory and item response theory (latent trait theory), are briefly introduced. This course is intended to equip students to read the literature in their own substantive areas more critically, to use tests more intelligently in research, and to pursue further studies in psychometrics. It is prerequisite to Education 353A (Item Response Theory) and Education 353C (Generalizability Theory). Although the focus of the course is on paper-and-pencil measures of cognitive abilities and academic achievement, most of the concepts and methods developed apply equally to performance testing, as well as the assessment of attitudes and personality constructs, ratings based on systematic observations, and other kinds of assessments of individuals or groups. Materials The text for the course is Introduction to Classical and Modern Test Theory, by Linda Crocker and James Algina (1986). There is also a collection of "programmed instruction" materials for this course, which provide a systematic introduction to and practice in using the notation, statistical concepts, and algebraic manipulations required. Various authors, primarily L. J. Cronbach, have contributed to these materials over the years. In 1982, the teaching assistant for this course, David King, retyped and extensively revised these materials to standardize their format, correct minor errors, and provide coverage of some topics that had been omitted. Some students have found these materials quite helpful. Two copies are on reserve in Cubberley Library. Anyone wishing to purchase a copy (at cost, about $12) should talk with me or send me an email. Use of the PI materials is optional. References to the PI materials are indicated on the course schedule. Students are also encouraged to pursue the readings listed in the attached bibliography. These are intended both to provide additional review and practice for students especially insecure about their statistical preparation, and to offer more thorough coverage of selected topics, for those who wish to go beyond what is provided in the text or who are concerned with specific kinds of applications of measurement theory. In previous years, optional readings have been placed on reserve. The library has found, however, that the demand for optional readings is generally low, and that the restricted loan period for materials on reserve inconveniences those students who do make use of them. For the past several years, I have followed the library's recommendation in not placing these materials on reserve, and students have not reported any difficulties. If access to or competition for these materials should pose any problem, please let me know, and I'll try to locate additional copies for you to use. Course Meetings, Requirements, and Grading Attendance is expected at lectures, 9:00-10:50. The regular meeting times are Mondays and Wednesdays when possible, but the AERA and NCME Annual Meetings and other commitments will necessitate one regular class meeting on Friday April 9. I have found that scheduled office hours do not provide sufficient flexibility to accommodate students' varied schedules, which is why I have listed "office hours by arrangement, individual or small group." Please do not hesitate to send me an email to set up a time to meet if necessary. Early in the quarter, on some Fridays when there is no lecture scheduled, I plan to conduct optional discussion sections for students wishing extra help or practice with course material. The course is listed as "3-4 units." For most students, this should be a four-unit course, but enrollment for three units is offered as a courtesy for those with limited tuition grants or stipends. Enrollment for 3 or 4 units makes no difference in the work expected. The satisfactory/no credit (+/NC) option is available for this course. Brief in-class quizzes and homework exercises will be assigned to provide practice and to check comprehension. Completion of these quizzes and exercises is required, but grades will be based solely on the take-home midterm and final examinations, with more weight on the final. This course is not designed for students with exceptionally strong mathematical and statistical preparation, but such students may nonetheless desire a systematic introduction to measurement theory. If you fall into this category, please see me individually to arrange some special project in connection with the course, such as a brief paper on some measurement topic, a critique of measurement methods in one or more pieces of published research, a critique of the measurement of some psychological construct, or a review of one or more published tests. I will be happy to arrange one or more additional units of credit for such special projects in conjunction with the course, if you wish. Small groups of students are also welcome to undertake such projects. Students with Documented Disabilities Students who have a disability that may necessitate an academic accommodation or the use of auxiliary aids and services in a class must initiate the request with the Disability Resource Center (DRC). The DRC will evaluate the request with required documentation, recommend appropriate accommodations, and prepare a verification letter dated in the current academic term in which the request is being made. Please contact the DRC as soon as possible; timely notice is needed to arrange for appropriate accommodations. The DRC is located at 563 Salvatierra Walk (phone 7231066; TDD 725-1067). Tentative Course Schedule Date Topic Readingsa W 3/31 F 4/ 2 Introduction and Overview No class C&A 3-15; PI 1-39 M 4/ 5 Properties of composite scores C&A 16-50, 60-64, 87-101; PI 40-111; WE&C 56-73, Ch. 12 W 4/ 7 The classical (weak) true-score model C&A 36-42(review),105-30;PI 112-133, §§B,Q; WE&C 178-192; Gulliksen Chs. 2, 3 F 4/ 9 Estimating reliability (1) C&A 131-43; PI C1-C11; Feldt & Brennan 105-113; Gulliksen Chs. 6-7 M 4/12 W 4/14 F 4/16 AERA/NCME Annual meetings: Class cancelled AERA/NCME Annual meetings: Class cancelled AERA/NCME Annual meetings: Class cancelled M 4/19 Estimating reliability (2); stratified alpha Feldt & Brennan 113-118; Feldt (1990) W 4/21 F 4/23 Useful applications (1) Optional discussion section C&A 143-56; Gulliksen Chs. 8,10 M 4/26 Useful applications (2); rel. of group means Feldt & Brennan §§ 2.7.1, 2.8.8 W 4/28 F 4/30 Conditional standard errors of measurement Optional discussion section Feldt & Brennan §§ 2.8.3 M 5/ 3 Generalizability theory C&A 157-91; Shavelson & Webb W 5/ 5 Validity (1) C&A 217-42; PI C11-C17; Campbell & Fiske (1959); APA Standards F 5/ 7 Optional discussion section M 5/10 Validity (2); MIDTERM DUE Messick (1989, 1995) W 5/12 F 5/14 Factor Analysis No class C&A 287-308 M 5/17 Test construction and item analysis C&A 66-86, 311-38; Haertel (1985); Millman & Greene; WE&C 196-203 W 5/19 Scores, scales, and norms C&A 399-409,431-32,438-55; Petersen, Kolen, & Hoover F 5/21 No class M W F M W Item Bias, Test Bias, & Equity Item response theory No class Memorial Day Observed (University Holiday) Testing and Educational Policy 5/24 5/26 5/38 5/31 6/ 2 C&A 267-78, 283-85; Cronbach (1976) C&A 339-50,352-54,361-71 W 6/ 9 TAKE-HOME FINAL EXAMINATION DUE, 12:00 noon _________________________________________________________________________ a"C&A" refers to Crocker and Algina (1986) and “WE&C” to Welkowitz, Ewen, and Cohen (2000). "PI" refers to the "Programmed Instruction" materials available as a course reader. Written examinations will cover both lectures and reading assignments from C&A. References to PI and WE&C are to provide additional explanatory material for students who lack a strong statistical background. The remaining readings treat specific topics in greater depth. References American Educational Research Association (AERA), American Psychological Association (APA), and the National Council on Measurement in Education (NCME). (1999). Standards for educational and psychological testing. Washington, D.C.: American Psychological Association. Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitraitmultimethod matrix. Psychological Bulletin, 56, 81-105. Crocker, L., and Algina, J. (1986). Introduction to Classical and Modern Test Theory. New York: CBS College Publishing. Cronbach, L. J. (1976). Equity in selection: When psychometrics and political philosophy meet. Journal of Educational Measurement, 13, 31-41. Feldt, L.S. (1990). The sampling theory for the intraclass reliability coefficient. Applied Measurement in Education, 3, 361-367. Feldt, L. S., & Brennan, R. L. (1989). Reliability. In R. L. Linn (Ed.), Educational measurement (3rd ed., pp. 105-146). New York: Macmillan. Gulliksen, H. (1950). Theory of mental tests. New York: Wiley. Haertel, E. H. (1985). Construct validity and criterion-referenced testing. Review of Educational Research, 55, 23-46. Messick, S. (1989). Validity. In R. L. Linn (Ed.), Educational measurement (3rd ed., pp. 13-103). New York: Macmillan. Messick, S. (1995). Validity of Psychological Assessment: Validation of Inferences from Persons' Responses and Performances as Scientific Inquiry into Score Meaning. American Psychologist, 50, 741-749. Millman, J., & Greene, J. (1989). The specification and development of tests of achievement and ability. In R. L. Linn (Ed.), Educational measurement (3rd ed., pp. 335-366). New York: Macmillan. Petersen, N. S., Kolen, M. J., & Hoover, H. D. (1989). Scaling, norming, and equating. In R. L. Linn (Ed.), Educational measurement (3rd ed., pp. 221-262). New York: Macmillan. Shavelson, R. J., & Webb, N. M. (1991). Generalizability theory: A primer. Newbury Park, CA: Sage. Welkowitz, J., Ewen, R.B., Cohen, J. (2000). Introductory statistics for the behavioral sciences (5th Ed.).Orlando: Harcourt Brace College.