Guide to the Use and Interpretation of Scores for the EXADEPTM Test
Transcription
Guide to the Use and Interpretation of Scores for the EXADEPTM Test
Guide to the Use and Interpretation of Scores for the EXADEP Test TM www.ets.org/exadep The Examen de Admisión a Estudios de PosgradoTM (EXADEPTM) is administered by Educational Testing Service. Please send all your questions to the ETS office in Puerto Rico. ETS Puerto Rico Office Global Division American International Plaza 250 Muñoz Rivera Ave. 3rd Floor, Suite 315 Hato Rey, PR 00918 Phone: TDD: Fax: E-mail: Web site: 787-753-6363 787-758-4598 787-250-7426 PRO@ets.org www.ets.org/exadep 8th Edition, 2011. Copyright © 2011 by Educational Testing Service. All rights reserved. ETS, the ETS logo and LISTENING. LEARNING. LEADING. are registered trademarks of Educational Testing Service (ETS) in the United States of America and other countries. EXADEP and EXAMEN DE ADMISIÓN A ESTUDIOS DE POSGRADO are trademarks of ETS. 2 TABLE OF CONTENTS Page PREFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 DESCRIPTION OF THE TEST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 EXADEP Measures Specific Reasoning Skills That Are Developed Over a Long Period of Time . . . . 7 GUIDELINES FOR THE USE OF EXADEP SCORES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Normally Appropriate and Inappropriate Uses of EXADEP Scores . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Appropriate Uses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Inappropriate Uses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 CONSIDERATIONS IN SCORE INTERPRETATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Repeat Test Takers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Examinees with EXADEP Scores More Than Five Years Old . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Examinees with Disabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 POLICY AND USE OF EXADEP SCORES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Score Reporting Policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Reporting Revised Scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Confidentiality and Authenticity of EXADEP Scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 EXADEP Scores and Graduate Admissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Validity Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 SCORE INTERPRETATION AND STATISTICAL INFORMATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Verbal, Mathematics and Analytical Reasoning, Written Expression, and English Sections of the Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Standard Error of Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Reliability Coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 STATISTICAL TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 How to Compare the Scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Mean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Percentile Ranks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Table 1: Business . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Table 2: Science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Table 3: Education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Table 4: Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Table 5: All Other Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Table 6: All Candidates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 STEPS THAT ETS TAKES TO ENSURE FAIRNESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Involving External Faculty Members in the Development of the Test. . . . . . . . . . . . . . . . . . . . . . . . . 23 Differential Item Functioning Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3 PREFACE Our mission at Educational Testing Service® (ETS®) is to advance quality and equity in education by providing fair and valid assessments, research, and related services. The Guide to the Use and Interpretation of Scores for the EXADEP™ supports this mission by helping to ensure that the test is created and scores are used under fair and appropriate ETS guidelines. The purpose of this guide is to help institutions or university members responsible for graduate admissions and fellowship awards to use EXADEP scores as an additional aid in their decision making. Understanding the contents of this guide will help prevent misuse and misinterpretation of the test scores and potential unfairness to applicants. An absolute commitment to excellence, integrity, and fairness is at the core of everything we do. Our products and services are meant to help educators, educational institutions, businesses, and governments further education for all people worldwide. Additional copies or information about the test or services may be obtained from the ETS Puerto Rico Office by e-mail or visiting our Web site at: www.ets.org/exadep. 4 INTRODUCTION EXADEP scores can be used by admissions or fellowship panels to supplement undergraduate records and other qualifications for graduate study. The scores provide common measures for comparing the qualifications of applicants and aid in the evaluation of grades and recommendations. Any accredited graduate or professional school or any department or division within a school may require or recommend that its applicants take the test. An institution not accredited by a recognized agency can become a score user if approved by the EXADEP Program. The weight to be given to EXADEP scores can generally be established by relating what the tests measure to the orientation, curriculum, and aims of a department. Specifically, the content validity of the tests for a graduate department should be determined by reviewing each test carefully and then making subjective decisions as to the weight, if any, the scores on EXADEP tests should receive in relation to other admission factors. Before taking the EXADEP, students should have completed at least two years of undergraduate study in an institution where the language of instruction is Spanish. Students should indicate on their registration form the institutions to which they want their test scores sent. The institutions or the faculties within an institution that require or recommend that the candidates take the EXADEP should announce it in their publications and on their Web site, and they should notify the candidates. It is also important to let the candidates know the deadlines for receiving scores. For institutions that are considering using the EXADEP test or for those planning to require EXADEP as a requisite of admission, additional information and convenient forms can be found on the EXADEP Web site (www.ets.org/exadep). 5 DESCRIPTION OF THE TEST The purpose of the EXADEP test is to measure an examinee’s quantitative, analytical reasoning and verbal abilities in Spanish and in English as a foreign language in order to contribute to the prediction of the examinee’s performance in graduate or professional schools and for fellowship selection. The test comprises five sections. The first four are in Spanish and the fifth is in English. All questions in the test are multiple-choice questions. Section 1: Verbal Aptitude (90 minutes) Part A: Antonyms Analogies Sentence Completion Part B: Reading Comprehension Section 2: Mathematics (40 minutes) Arithmetic Algebra Geometry Section 3: Analytical Reasoning (40 minutes) Section 4: Written Expression (30 minutes) Part A: Language Usage Part B: Sentence Correction Section 5: English (45 minutes) Part A: Sentence Completion Antonyms Part B: Reading Comprehension The total testing time is four hours and five minutes. There is a ten-minute break between Sections 2 and 3. • The Verbal section measures the ability to analyze and evaluate written material and synthesize information obtained from it, to analyze relationships among component parts of sentences, and to recognize relationships between words and concepts. In each test edition, the passages are a balance among three different subject-matter areas: humanities, social sciences, and natural sciences. • The Mathematics section measures basic mathematical skills and understanding of elementary mathematical concepts, as well as the ability to reason quantitatively and to solve problems in a quantitative setting. There is a balance of questions requiring arithmetic, algebra, geometry, and data analysis. • The Analytical Reasoning section measures the ability to think analytically. It tests the ability to draw inferences and think deductively based on a partially defined scenario and a set of conditions that must hold in any fully developed version of the scenario. It does not test knowledge of any particular subject nor does it require training in formal logic. • The Written Expression section measures the ability to recognize the use of language essential to a finished piece of writing that would be considered acceptable by most educated readers and writers of Spanish. • The English section measures certain elements in the spectrum of abilities required to reason effectively in a verbal medium. 6 EXADEP Measures Specific Reasoning Skills That Are Developed Over a Long Period of Time. EXADEP is designed to measure verbal, quantitative, and analytical reasoning abilities considered important for successful performance in graduate school. The test is designed to be independent of particular courses of study. Instead, it represents knowledge, skills, and abilities that are developed across many courses and that reflect exposure to demanding courses of study over a long period of time. Research on other ETS tests used for similar purposes has shown that short-term study will not alter scores greatly for most people. It is important that test takers be thoroughly familiar with the test—its contents and procedures—before the actual testing day in order to avoid receiving lower scores than they might otherwise obtain. The long-term reasoning component measured by EXADEP makes it a test of developed skills. Both words are significant here. The skills are reasoning skills of the type most important to success in graduate study. The developed nature of these skills comes from a lifetime as a reader, as one who thinks through quantitative problems, and as one who is accustomed to communicating and critiquing arguments in writing. Unlike high school or undergraduate grade reports, EXADEP scores reflect performance on tasks that are common to ALL applicants. The complementarity between EXADEP scores and other elements of the application means that EXADEP scores should always be used in conjunction with those other elements. Standardized test scores provide information that is highly comparable across examinees and that references a well-documented set of knowledge and skills. Such scores do not, however, provide information about other important characteristics, such as motivation and persistence. Furthermore, they do not typically provide a good indicator for understanding such complexities as changes in level of attainment over time or obstacles overcome to reach a particular level of attainment. Grades and letters of recommendation, in contrast, are less comparable across examinees because they are based on different requirements for different individuals. Grades represent different combinations of courses and are based on grading standards unique to each institution, department, and professor. Letters of recommendation may provide valuable information about an individual, but typically offer little systematic basis for comparing one applicant to another. Despite these weaknesses, grades and letters of recommendation probably provide better information than standardized tests about such traits as persistence, motivation, and ability to overcome obstacles, all of which may be of considerable importance to success in graduate study. This is why it is important to combine grades, letters of recommendation, and test scores to develop a holistic understanding of an applicant’s abilities and of achievement in relation to abilities. The complementarity of test scores and grades is made clear by the results of validity studies conducted by the EXADEP Program in collaboration with various universities. These studies show that the combination of EXADEP scores and undergraduate GPA predicts first-year grades more effectively than any single piece of information. These various sources of information represent a system of checks and balances in decision making. Fairness is enhanced by using multiple measures because systematic over- or under-prediction will be decreased as additional measures are used. For example, students coming from schools or fields that employ strict grading practices will be less disadvantaged when tests are used along with grades. Conversely, test scores may represent peak performance capacity, while grades and letters of recommendation can offer testimony to an individual’s ability to sustain performance over longer periods of time. 7 GUIDELINES FOR THE USE OF EXADEP SCORES Introduction These guidelines have been adopted by the EXADEP Program to provide information about the appropriate use of the EXADEP scores for those who use the scores in graduate admissions and fellowship selection processes and for other approved purposes. They are also intended to protect students from unfair decisions that may result from inappropriate uses of scores. Adherence to the guidelines is important. EXADEP is designed to assess academic knowledge and skills relevant to graduate study. As measures with known statistical properties and high-quality technical characteristics, the scores from these tests, when used properly, can improve graduate admissions and fellowship selection processes. Any EXADEP test, however, has two primary limitations: (1) it does not and cannot measure all the qualities that are important in predicting success in graduate study or in confirming undergraduate achievement, and (2) it is an inexact measure—that is, only score differences that exceed the standard error of measurement of a given score can serve as reliable indications of real differences in academic knowledge and developed abilities. Although limitations and cautions apply to all admissions measures, the EXADEP Program is obligated to inform users of the appropriate uses of EXADEP scores and to identify and try to rectify instances of misuse. To this end, the following policies and guidelines are available to all EXADEP test takers, institutions, and organizations that receive EXADEP scores. Policies In recognition of its obligation to ensure the appropriate use of EXADEP scores, the EXADEP Program has developed policies designed to make score reports available only to approved recipients, to encourage these score users to become knowledgeable about the validity of the tests, to protect the confidentiality of examinees’ scores, and to follow up on cases of possible misuse of scores. The policies are discussed below. Score Recipients Accredited undergraduate and graduate institutions and non-degree-granting organizations that award graduate fellowships are eligible for consideration as score recipients. Institutions and organizations that do not meet either one of these requirements are, in general, not eligible to be score recipients. The EXADEP Program retains the right to make exceptions to this policy in special circumstances. Validity The general appropriateness of using EXADEP scores for graduate admissions, fellowship selection, and other approved purposes has been established by research studies carried out by ETS. EXADEP scores may be appropriate for some other purposes, but it is important for the user to validate their use for those purposes. To assist institutions in evaluating proposed uses, these guidelines include information about appropriate and inappropriate uses. Confidentiality EXADEP scores, whether for an individual or aggregated for an institution, are confidential and can be released only by authorization of the individual or institution or by compulsion of legal process. Use of Scores in Aggregated Form Information based on EXADEP scores may be useful to prospective students, but use of a precise mean or median should be avoided. Graduate programs and institutions are urged to report EXADEP scores in ranges such as the highest and lowest scores of the middle 50 percent of the admitted students. Presenting information by score ranges emphasizes the diversity of individual scores for any one graduate program or 8 institution and also makes clear the overlap of scores among graduate programs and institutions. Use of EXADEP scores in aggregated form as a measure for ranking or rating graduate programs, institutions, university systems, or states is strongly discouraged except when the scores are used as one indicator among several appropriate indicators of educational quality. Encouragement of appropriate use and investigation of inappropriate use All users of EXADEP scores have an obligation to use the scores in accordance with published EXADEP Program policies and guidelines. Institutions have a responsibility to ensure that all individuals using EXADEP scores are aware of the EXADEP Program score-use policies and guidelines and to monitor the use of the scores, correcting instances of inappropriate use when they are identified. The EXADEP Program staff is available to assist institutions in resolving issues of inappropriate score use. Guidelines Use Multiple Criteria Regardless of the decision to be made, multiple sources of information should be used to ensure fairness and balance the limitations of any single measure of knowledge, skills, or abilities. These sources may include undergraduate grade point average, letters of recommendation, personal statement, samples of academic work, and professional experience related to proposed graduate study. EXADEP scores should not be used exclusively. The use of multiple criteria is particularly important when using EXADEP scores to assess the abilities of educationally disadvantaged students, as well as those who are returning to school after an extended absence. Score users are urged to become familiar with factors affecting score interpretation for these groups, as discussed in this publication. Accept Only Official EXADEP Score Reports The only official reports of EXADEP scores are those issued by ETS and sent directly to approved institutions and organizations designated by the test takers. Scores obtained from other sources should not be accepted. If there is a question about the authenticity of a score report, it should be referred to ETS. ETS will verify the accuracy of the scores and whether an official report was issued. Conduct Validity Studies Departments using EXADEP scores for graduate admissions, fellowship awards, and other approved purposes are encouraged to collect validity information by conducting their own studies. The EXADEP Program staff will provide advice on the design of appropriate validation studies without charge. Maintain Confidentiality of EXADEP Scores All individuals who have access to EXADEP scores should be aware of the confidential nature of the scores and agree to maintain their confidentiality. Institutional policies should be developed to ensure that confidentiality is maintained. For example, EXADEP scores should not be placed on documents sent outside the institution. Consider Verbal Aptitude, Quantitative and Analytical Reasoning, Written Expression, and English Subscores Since the level of skills required for success in graduate school varies by field or department, the subscores provided for each section of the test should be taken into consideration. Avoid Decisions Based on Small Score Differences Small differences in EXADEP scores (as defined by the standard error of measurement) should not be used to make distinctions among examinees. Standard errors of measurement (SEMs) vary by test and are available in this publication. 9 Do Not Compare Scores from Different Tests EXADEP scores are not directly comparable with scores on other graduate or undergraduate admission tests. Percentile ranks should be compared only if they are based on the same reference population. Recognize Limitations of Scores Earned on Tests Taken Under Special Conditions EXADEP tests are offered with special arrangements and test materials to accommodate the needs of students with visual, physical, hearing, and learning disabilities. Depending upon the nature and extent of the disability, the scores may not accurately reflect a student’s educational achievement. For some students, the nature of their disabilities may make it advisable to waive EXADEP score requirements. Normally Appropriate and Inappropriate Uses of EXADEP Scores The suitability of EXADEP for a particular use should be explicitly examined before using test scores for that purpose. The following lists of appropriate and inappropriate uses of EXADEP scores are based on the policies and guidelines previously outlined. The lists are meant to be illustrative, not exhaustive, in nature. There may be other appropriate uses of EXADEP scores, particularly for non-admissions purposes, but any use other than those listed below must be discussed in advance with EXADEP Program staff to determine their appropriateness. If a use other than those appropriate uses listed below is contemplated, it will be important for the user to validate the use of scores for that purpose. The EXADEP Program staff will provide advice on the design of such validity studies without charge. Appropriate Uses Provided all applicable guidelines are adhered to, EXADEP scores are suitable for the following uses: • selection of applicants for admission to graduate school • selection of graduate fellowship applicants for awards • selection of graduate teaching or research assistants • guidance and counseling for graduate study Inappropriate Uses The EXADEP scores are not suitable for the following uses: • Requirement of a minimum score on the EXADEP for conferral of a degree, credit-by-examination, advancement to candidacy, or any noneducational purpose • Requirement of scores on the EXADEP for employment decisions, including hiring, salary, promotion, tenure, or retention (except for the awarding of assistantships to graduate students) Comments, complaints, inquiries, and suggestions about the use of EXADEP scores are welcome. To contact the EXADEP Program office, see the inside front cover. 10 CONSIDERATIONS IN SCORE INTERPRETATION EXADEP scores should be used to supplement the information provided in a person’s application, such as undergraduate record and letters of recommendation. Officials responsible for admission at each institution must determine the significance of EXADEP scores for each applicant. Particular attention must be paid to the use of EXADEP scores for individuals described below. The experience of institutions or departments should continue to be the best guide to interpretation of EXADEP scores in these instances. Repeat Test Takers Individuals are permitted to take EXADEP tests more than once. Score recipients are cautioned not to view an increase in scores necessarily as a reflection of academic gain, especially over a short time period. There are several ways in which graduate departments can judge multiple scores for an individual (e.g., use average of all scores, use most recent score, use highest score). Using the mean score may be the best technique because it is the most objective. Whatever approach is adopted, it should be used consistently with all applicants. Examinees with EXADEP Scores More Than Five Years Old The EXADEP Program established a policy of retaining and reporting EXADEP scores earned during the fiveyear period after the test date. Scores more than five years old are not reported. Candidates who took the test more than five years ago must take it again if they want scores to be sent to institutions. When institutions decide to consider older scores, they should be aware that the applicant’s competence may have changed in the intervening time. The problem lies in determining how much the applicant’s competence has changed in either direction in the intervening years and how the change affects present potential for graduate work. Institutions may prefer to request that the applicant retake the test. Examinees with Disabilities ETS makes special arrangements for individuals who have recently documented visual, physical, hearing, or learning disabilities and are unable to take the tests under standard conditions. The tests are administered in a nonstandard manner chosen to minimize any adverse effect of the examinee’s disability on test performance and to help ensure that, insofar as possible, the final scores should closely represent the examinee’s educational achievement. Depending on the nature and extent of the disability, an examinee’s scores may not fully reflect his or her educational achievement and, because there are so few persons with disabilities taking EXADEP tests and their circumstances vary so widely, it has not been possible to provide special interpretive data for these examinees. Therefore, graduate schools should seriously consider waiving EXADEP requirements for applicants with certain disabilities. 11 POLICY AND USE OF EXADEP SCORES Score Reporting Policies EXADEP score reporting policies have been adopted by the EXADEP Program to encourage the appropriate use of EXADEP scores and to protect the right of individuals to control the distribution of their own score reports. Current EXADEP Program policy states that scores are reportable for five years. Score reports for the test are released approximately three to four weeks after the test date to the examinees and to accredited institutions of higher education granting the baccalaureate or higher and approved graduate fellowshipgranting sponsors designated by the examinees. Absentees are reported to institutions. Their names will appear on the institution roster with no scores. Percentile ranks shown on the tables included in this guide and on the score reports are based on the performance of the current reference group. The percentile rank for any score may vary over the years depending on the scores of the group with which the score is compared. Thus, when two or more applicants are being compared, the comparison should be made on the basis of their respective scores; if percentile ranks are considered, they should all be based on the percentile rank tables in the most recent edition of this guide. Reporting Revised Scores ETS routinely follows extensive review and quality control procedures to detect and avoid flawed questions and consequent errors in scoring. Nonetheless, if an error is discovered after scores have been reported, the specific circumstances are reviewed carefully, and a decision is made about how best to take corrective action that is fairest to all concerned. Revised scores are sent to the affected students, who may request that ETS send the revised scores to any graduate schools or fellowship sponsors still considering their applications. Confidentiality and Authenticity of EXADEP Scores EXADEP scores are confidential and must not be released by an institutional recipient without the explicit permission of the examinee. EXADEP scores are not to be included in academic transcripts. Dissemination of score records should be kept to a minimum, and all staff who have access to them should be explicitly advised of their confidential nature. To ensure the authenticity of scores, the EXADEP Program urges that institutions accept only official reports of EXADEP scores received directly from ETS. The EXADEP Program recognizes the privacy rights of both institutions and individuals with regard to information supplied by and about them. ETS therefore safeguards from unauthorized disclosure all information stored in its data or research files. Information about an institution (identified by name) will be released only in a manner consistent with a prior EXADEP agreement or with the consent of the institution. EXADEP Scores and Graduate Admissions Many factors play a role in an applicant’s admissibility and expectation of success as a graduate student. EXADEP scores are only one element in this total picture and should be considered along with other data. The EXADEP Program believes that EXADEP scores should never be the sole basis for an admissions decision and that it is inadvisable to reject an applicant solely on the basis of EXADEP scores. A cutoff score should not be used without consideration of other admission factors. 12 EXADEP scores permit comparison of one applicant to a graduate school with other applicants for the same program at that institution as well as with everyone else who took the test. Subscores provide further information for consideration. These subscores, which reflect a student’s general strengths and weaknesses in the major areas on which the total score is based, aid in the interpretation of the total score. Often the subscores can suggest areas in which the student may require extra work. A low subscore, however, may be the result of lack of exposure to a particular field. As a result, subscores should always be reviewed in relation to the applicant’s undergraduate history. For admissions decisions, the weight to be given to EXADEP scores can generally be established by relating what the tests measure to the orientation, curriculum, and aims of the department. Specifically, the content validity of the tests for a graduate department should be determined by reviewing the test carefully and then making subjective decisions as to the weight, if any, the scores on EXADEP should receive in relation to other admission factors. Validity Studies One way to determine the weight to give to test scores is to conduct validity studies. Validity is an ongoing process of assembling knowledge supporting interpretations that are made using test scores. A primary way of determining the validity of a test is to examine the correlation between test scores (and perhaps other predictors, such as undergraduate grade point average) and one or more criteria of success in graduate study. It should be noted, however, that where there are small numbers of students, major problems can occur in attempting to carry out adequate validity studies. In addition to correlation studies, institutions might consider other approaches. One would be to prepare a table of EXADEP scores for those students who do poorly and/or drop out of graduate school. Another approach would be to independently assess the general skills or abilities needed for success in the particular graduate school and compare them to those assessed by the EXADEP. The EXADEP Program strongly recommends that institutions using EXADEP scores conduct such validity studies, and it will assist institutions that wish to do so. Institutions interested in such assistance should contact the ETS office in Puerto Rico. 13 SCORE INTERPRETATION AND STATISTICAL INFORMATION Verbal, Mathematics and Analytical Reasoning, Written Expression, and English Sections of the Test The range of EXADEP scores for each section is from 20 to 80, in one-point increments. If no answers are given for a section, a score of 20 is reported for that section. The total scale score ranges from 200 to 800, and it is the sum of the results of the different sections multiplied by their individual weights. The weight of each section is proportional to the number of questions in the section. The distribution of weights is given in the following table: Section 1 Verbal Aptitude 2 Mathematics and 3 Analytical Reasoning 4 Written Expression 5 English TOTAL Section Score A Weight 3.5 B 2.5 C D 2.0 2.0 3.5A + 2.5B + 2.0C + 2.0D Note: One score is reported for sections 2 and 3 combined. There may be candidates who do not answer all questions in a given section of the test. Because the number of answers is incorporated into the calculation of the scores, it is important that test takers answer every question. The following table indicates the mean, standard deviation, SEMs, and reliability coefficient based on the results of all the candidates who took the test between February 2008 and December 2010. EXADEP SUMMARY TABLE (Based on the scale scores of 24,969 candidates who took the test between February 2008 and December 2010) Mean Standard Deviation SEM Reliability Coefficient 45.1 11 4 0.868 49.5 10 4 0.832 4 Writing Expression 49.2 12 5 0.798 5 English 47.5 12 5 0.836 475.7 91 22 0.938 Section 1 Verbal Aptitude 2 Mathematics and 3 Analytical Reasoning TOTAL SEM = Standard Error of Measurement 14 The EXADEP Summary Table (on page 14) provides data on the standard error of measurement and reliability. Standard Error of Measurement As with all educational measurements, the scores obtained by an individual could easily vary from one administration to another, although there may have not been any changes in the candidate’s true skills. The individual score is considered an estimate of the knowledge or skills of the person in the area examined. If a person could take different editions of the test without changes in his or her knowledge and skills, the average of all the scores obtained would be a precise measurement of the knowledge and the skills of the person in the area examined. This hypothetical average is known as the “true score”. The difference between the “true score” and the score obtained at a test administration is called “error of measurement.” The term tries to describe the imprecise measurement of a unique score obtained by the candidate as the measurement of knowledge or skills of the person. It does not mean that an error was committed when developing or evaluating the test. It is statistically possible to estimate the average of the error of measurement for a large group of candidates. The statistic outcome is expressed in a score that is known as “standard error of measurement” (SEM). If the “true scores” for a group of candidates are distributed according to a normal distribution, the probability is that nearly 68 percent of the candidates will obtain scores within a standard error of measurement above or below their true scores. The standard error of measurement for the total score of the EXADEP, expressed in scale scores, is 22. Therefore, for the 68 percent of the EXADEP candidates, it can be assumed that their true scores would fall within a range of 22 points above and 22 points below the obtained scores. The importance of the standard error of measurement is in understanding that, when comparing the performance in the examination of two candidates, the variations in the scores perhaps do not represent a real difference as far as the skills of the respective candidates. The value of the standard error of measurement must be taken into consideration when interpreting the individual results. Reliability coefficient The reliability of a test indicates the degree to which individual candidates scores would change the relative position of the scores if the examination had been administered under conditions to a certain extent different (for example, if the candidates were tested by means of different test edition). The reliability is represented by means of a statistical coefficient that is affected by measurement errors. Generally, while smaller the measurement errors in a test, the greater the reliability. The reliability coefficient ranges from 0 to 1. The reliability index of 1 indicates a perfectly reliable test (that is, that there are no measurement errors in the test). The 0 indicates that the test obtains completely inconsistent results. 15 STATISTICAL TABLES To aid in the interpretation of scale scores, the EXADEP Program describes scores in terms of their standing in appropriate reference groups. Tables 1–6 provide percentile ranks (i.e., the percentage of candidates in a group who obtained scores lower than each possible scale score) for the test. All candidates who took the test between February 2008 and December 2010 were divided into five reference groups according to the graduate program to which they requested the scores to be sent. Tables 1–5 show percentile ranks according to those five groups. Table 6 shows percentile ranks for all candidates. The tables also indicate the means and standard deviations for each group. • • • • • • Table 1: Business Table 2: Science Table 3: Education Table 4: Law Table 5: All Other Programs* Table 6: All Candidates How to Compare the Scores Mean A way to interpret the scores is to compare the results of a candidate with the mean of scores obtained by all candidates that took the test. It can be done for an EXADEP candidate by comparing the results of the candidate with the mean of results in the Table of Summary of the EXADEP that appears on page 14. However, it is even better to compare the scores of the candidate with the reference group in the tables on pages 17–22. Percentile Ranks Another method to interpreting the EXADEP scores is to compare the score of one candidate with the score of another candidate, as it indicates in percentile rank tables, Table 1 – Table 6. The candidate’s score can be compared in relation to other candidates by means of the use of the tables. For example, in Table 1, if the scaled score of a candidate is 56 in Section 1 (Verbal Aptitude), by using the Table, then 86 percent of the other candidates obtained lower scores. Also, if a candidate obtained a scaled score of 60 in Sections 2 and 3 (Mathematical and Analytical Reasoning), according to the table 83 percent of the candidates of the group obtained a lower score. If a candidate obtained a scaled score of 68 in Section 4 (Written Expression), then according to the table, 97 percent of the candidates obtained a lower score. If a candidate obtained a scaled score of 61 on Section 5 (English), then the table shows that between 86 and 89 percent of the candidates obtained a lower score. Also note that not all the total scores appear in the tables, but as multiples of 10. For example, in Table 1, if the total scaled score of a candidate is 563, then (according to the table, 560 is the closest number) we can assume that 84 percent of the candidates obtained a lower score. *The fifth reference group includes all graduate programs that, from a statistical point of view, did not have enough candidates. 16 Table 1 PERCENTILE RANKS FOR BUSINESS PROGRAM CANDIDATES Percentage of candidates who obtained scale scores lower than each possible scale score ENGLISH TOTAL SCALE SCORE PERCENTILE RANK OF THE TOTAL SCALE SCORE 99 98 98 97 97 95 94 93 91 89 86 83 80 76 73 68 62 57 48 40 31 23 14 8 5 2 1 1 1 1 1 800 780 760 740 720 700 680 660 640 620 600 580 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 99 99 99 99 98 97 96 94 91 88 84 79 73 65 57 49 41 32 24 17 11 7 4 2 1 1 1 4,316 4,316 4,316 4,316 n 4,316 Mean 43.7 50 47.9 46.8 Mean 467.6 Standard Deviation 11 10 12 11 Standard Deviation 91 80 78 76 74 72 70 68 66 64 62 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 99 99 99 99 99 99 98 96 95 92 89 86 82 77 71 66 57 51 45 38 31 25 20 15 12 8 6 5 3 1 n MATHEMATICS/ ANALYTICAL REASONING 99 99 99 99 99 98 97 95 92 88 84 79 73 66 61 54 49 42 36 30 26 21 17 14 11 8 6 4 3 2 1 VERBAL APTITUDE 99 99 99 98 97 95 94 92 89 86 83 78 74 68 62 54 47 40 32 23 17 12 7 4 2 1 1 1 1 SECTION SCALE SCORE WRITING EXPRESSION (Based on the results of 24,969 candidates who took the test between February 2008 and December 2010) 17 Table 2 PERCENTILE RANKS FOR SCIENCE PROGRAM CANDIDATES Percentage of candidates who obtained scale scores lower than each possible scale score TOTAL SCALE SCORE PERCENTILE RANK OF THE TOTAL SCALE SCORE 99 99 99 99 98 97 95 91 87 82 76 71 65 58 52 47 41 35 29 24 19 15 12 9 7 5 4 2 2 1 99 99 98 97 96 96 95 93 91 90 88 85 82 78 74 69 63 57 48 41 32 24 16 9 5 2 1 1 1 1 1 800 780 760 740 720 700 680 660 640 620 600 580 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 99 99 99 99 99 98 97 95 92 88 84 79 72 65 56 48 39 31 24 17 12 7 4 2 1 1 1 n 5,981 5,981 5,981 5,981 n 5,981 Mean 44.7 48.5 48.5 46.4 Mean 468.2 Standard Deviation 11 10 12 11 Standard Deviation 90 WRITING EXPRESSION 99 99 99 99 98 98 97 95 93 90 87 83 78 72 67 60 52 44 36 27 21 14 8 4 2 1 1 1 1 1 1 MATHEMATICS/ ANALYTICAL REASONING 99 99 99 99 99 99 97 96 93 90 87 83 79 74 68 62 54 47 41 35 28 22 18 14 11 7 5 3 2 1 VERBAL APTITUDE 80 78 76 74 72 70 68 66 64 62 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 SECTION SCALE SCORE ENGLISH (Based on the results of 24,969 candidates who took the test between February 2008 and December 2010) 18 Table 3 PERCENTILE RANKS FOR EDUCATION PROGRAM CANDIDATES Percentage of candidates who obtained scale scores lower than each possible scale score PERCENTILE RANK OF THE TOTAL SCALE SCORE TOTAL SCALE SCORE 99 99 99 99 97 95 93 90 87 82 78 72 66 61 54 47 41 35 29 24 19 15 12 9 7 5 3 2 1 99 99 99 98 98 97 97 96 95 94 92 90 88 85 83 79 74 69 61 53 44 34 24 15 9 4 2 1 1 1 1 800 780 760 740 720 700 680 660 640 620 600 580 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 3,482 3,482 3,482 n 3,482 42.3 45.2 46.5 43.4 Mean 441.6 11 9 12 10 Standard Deviation 86 80 78 76 74 72 70 68 66 64 62 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 99 99 99 99 99 98 97 95 93 90 87 84 79 74 69 63 57 50 44 37 31 25 20 15 11 8 5 3 1 n 3,482 Mean Standard Deviation 99 99 99 99 99 98 98 97 96 95 93 91 88 84 80 75 68 60 50 39 31 22 13 7 3 2 1 1 1 WRITING EXPRESSION ENGLISH MATHEMATICS/ ANALYTICAL REASONING VERBAL APTITUDE SECTION SCALE SCORE (Based on the results of 24,969 candidates who took the test between February 2008 and December 2010) 19 99 99 99 99 99 98 97 95 93 90 87 82 76 69 61 53 43 34 25 18 11 6 3 1 1 1 Table 4 PERCENTILE RANKS FOR LAW PROGRAM CANDIDATES Percentage of candidates who obtained scale scores lower than each possible scale score ENGLISH TOTAL SCALE SCORE PERCENTILE RANK OF THE TOTAL SCALE SCORE WRITING EXPRESSION MATHEMATICS/ ANALYTICAL REASONING VERBAL APTITUDE SECTION SCALE SCORE (Based on the results of 24,969 candidates who took the test between February 2008 and December 2010) 99 99 98 97 96 94 92 88 85 80 75 69 63 56 49 41 33 26 19 13 9 6 3 1 1 1 1 1 1 99 99 99 99 99 97 95 92 88 82 76 68 61 53 45 38 32 26 21 17 13 10 8 6 4 3 3 2 1 1 1 99 96 95 94 92 91 88 86 83 81 77 73 69 64 59 53 46 40 32 26 18 13 8 4 2 1 1 1 1 1 1 800 780 760 740 720 700 680 660 640 620 600 580 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 99 99 99 99 99 99 97 96 93 90 86 80 74 65 56 46 36 28 20 14 9 6 4 2 1 1 1 1 7,844 7,844 7,844 7,844 n 7,844 Mean 48.2 53.2 52.2 51.2 Mean 509 Standard Deviation 10 10 11 12 Standard Deviation 84 80 78 76 74 72 70 68 66 64 62 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 99 99 99 99 99 98 97 95 92 88 83 78 72 65 56 48 40 33 26 20 15 11 8 6 4 3 2 1 1 1 n 20 Table 5 PERCENTILE RANKS FOR ALL OTHER PROGRAMS Percentage of candidates who obtained scale scores lower than each possible scale score WRITING EXPRESSION ENGLISH TOTAL SCALE SCORE PERCENTILE RANK OF THE TOTAL SCALE SCORE MATHEMATICS/ ANALYTICAL REASONING VERBAL APTITUDE SECTION SCALE SCORE (Based on the results of 24,969 candidates who took the test between February 2008 and December 2010) 80 78 76 74 72 70 68 66 64 62 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 99 99 99 99 99 98 98 96 94 91 88 84 80 75 70 65 58 52 46 40 33 26 21 17 13 9 6 4 3 1 99 99 99 99 99 98 97 96 95 93 90 87 84 79 74 67 60 52 43 33 26 18 11 6 3 1 1 1 1 1 1 99 99 99 99 99 98 96 94 91 87 83 78 73 66 60 55 49 44 38 32 27 22 18 14 12 10 7 5 4 3 1 99 98 98 97 96 95 94 92 91 89 87 85 82 79 75 71 64 59 52 44 36 27 18 11 6 3 1 1 1 1 1 800 780 760 740 720 700 680 660 640 620 600 580 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 n 3,346 3,346 3,346 3,346 n 3,346 Mean 43.6 46.7 47.7 46 Mean 457.4 Standard Deviation 11 10 12 12 Standard Deviation 92 21 99 99 99 99 99 99 98 97 95 93 90 86 81 75 69 61 53 45 37 28 21 15 10 6 3 1 1 1 Table 6 PERCENTILE RANKS FOR ALL CANDIDATES Percentage of candidates who obtained scale scores lower than each possible scale score WRITING EXPRESSION ENGLISH TOTAL SCALE SCORE PERCENTILE RANK OF THE TOTAL SCALE SCORE MATHEMATICS/ ANALYTICAL REASONING VERBAL APTITUDE SECTION SCALE SCORE (Based on the results of 24,969 candidates who took the test between February 2008 and December 2010) 80 78 76 74 72 70 68 66 64 62 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 22 20 99 99 99 99 99 98 97 96 94 90 87 83 78 73 66 60 52 45 39 33 26 21 17 13 10 7 5 3 2 1 99 99 99 98 97 96 95 93 90 87 84 79 75 69 63 56 49 41 33 24 19 13 7 4 2 1 1 1 1 1 1 99 99 99 99 99 98 96 94 91 86 81 75 70 62 56 50 44 38 32 27 22 17 14 11 9 7 5 3 2 2 1 99 98 97 96 95 94 93 91 89 87 85 81 78 74 70 66 59 54 45 38 30 22 15 9 5 2 1 1 1 1 1 800 780 760 740 720 700 680 660 640 620 600 580 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 99 99 99 99 99 99 98 97 96 93 90 87 82 76 69 61 53 44 36 28 21 15 10 6 4 2 1 1 1 n 24,969 24,969 24,969 24,969 n 24,969 Mean 45.1 49.5 49.2 47.5 Mean 475.7 Standard Deviation 11 10 12 12 Standard Deviation 91 22 STEPS THAT ETS TAKES TO ENSURE FAIRNESS ETS has designed several procedures intended to build fairness into its tests: involving external faculty members in the design and oversight of the tests, the fairness review process, and the differential item functioning (DIF) analysis. The purpose of involving faculty members in the design and oversight of the tests is to make sure that the perspectives of a diverse group of people are considered in planning and ongoing operational activities. The purpose of the fairness review process is to ensure that tests reflect the multicultural nature of society and to screen out any material that might be offensive or less accessible to certain groups of test takers, such as those based on age, disability, ethnic group, race, or gender. The purpose of the DIF analysis is to identify any test questions on which members of a particular group of test takers perform differently than would be expected on the basis of their overall ability in the areas covered by the test. Involving External Faculty Members in the Development of the Test The EXADEP Program involves undergraduate and graduate faculty members in the design and oversight of the test. The EXADEP Advisory Committee is made up of men and women from different academic disciplines and different Spanish-speaking countries representing a variety of institutions. Members are drawn from different ethnic groups. Drawing on a diverse group of educators who are not ETS employees is one way ETS seeks to ensure the fairness of the test. Every question in an ETS test (and all materials published by ETS) must pass a fairness review. This review is based on a set of written guidelines; each review is conducted by an ETS staff member specifically trained in the application of these guidelines. Any test question that does not pass the fairness review must be revised to comply with the guidelines or be removed from the test. The fairness review does not guarantee that women, minority group members, or individuals with disabilities will perform well on the test, but it does guard against the possibility of distraction caused by language or content that might be found offensive or inaccessible. Differential Item Functioning Analysis Differential item functioning occurs when people from different groups and of approximately equal knowledge and skill perform in substantially different ways on a particular test question. Differential item functioning analysis is a statistical technique used as part of the testing process that is designed to identify test questions that are more difficult for members of one group than for members of some other group, controlling for overall ability. It is important to realize that DIF is not synonymous with bias. DIF may occur if a perfectly fair question happens to be measuring a skill that is not well represented in the test as a whole. Each DIF analysis involves a set of comparisons between a group of examinees that is the focus of the study (focal group) and the group with which it is compared (reference group). For example, if the focal group is women, the reference group is men. The DIF analysis is based on a comparison between groups of test takers of the same overall ability as determined by their performance on the test as a whole. A DIF statistic is computed for each test question, indicating the extent to which members of the focal group perform differently from members of the reference group who have similar ability levels. DIF analyses are run before scoring is performed. A test question that appears, on the basis of the DIF analysis, to be functioning in a substantially different way for the focal and reference groups, is reviewed and subject to removal from scoring. The EXADEP Program encourages test takers to report concerns about specific test questions directly to the test center administrator or to the EXADEP Program immediately following the test administration. Subject-matter specialists will review these questions and eliminate them from scoring if potential bias is determined. The test specialists will also respond in writing to the examinees. If a response does not resolve an examinee’s concern, the examinee may pursue the matter further with ETS. 23 Guide to the Use and Interpretation of Scores for the EXADEP Test TM Copyright © 2011 by Educational Testing Service. All rights reserved. ETS, the ETS logo and LISTENING. LEARNING. LEADING. are registered trademarks of Educational Testing Service (ETS) in the United States and other countries. EXADEP, EXAMEN DE ADMISIÓN A ESTUDIOS DE POSGRADO are trademarks of ETS. 16463