This House would abolish standardized tests for University Admission

Standardized tests are inherently arbitrary. They reduce an applicant’s entire academic career to a single one-day session. The result is an inherently unrepresentative test which fails to paint an accurate picture. What if a student has a bad day? What if they do poorly on the specific test questions?  In the SAT’s there is an error of measurement of about 30 points either way out of 800, this is the potential difference between where the student really is and what his or her score on the day was.[1] By contrast, looking at their entire academic record ensures that admissions officers will get a far more comprehensive picture of their actual ability. The law of averages means that bad days and tests will be balanced out with good ones, with the result that their academic record, the result of years of work, will reflect their true performance.

[1] Cloud, John, ‘What’s Good about the New SAT Test’, Time, 1 September 2006, http://www.time.com/time/nation/article/0,8599,1521184,00.html

A student’s academic record tends to measure very different things from what a standardized test does. GPA tends to be based on repetitive assignments such as homework, and in many cases students receive at least some academic credit for simply attending class. By contrast, standardized tests reward ability, by seeing whether or not at the end of the process students actually learned the material in question.  Performance under pressure is an important skill to measure, especially for top institutions, while sifting through the differing standards for what goes into the grades in different school districts is simply not possible.

Standardized tests are inherently skewed in favour of high-income students. Tutoring is and industry unto itself and the resulting rise in scores among high income students skews the scale against students whose families cannot afford it. By contrast, even if the money is available, the impact that tutoring can have on a student’s academic performance is limited. A tutor cannot write papers or take tests for them, and nor can they answer questions in class, or participate in sports or other extracurricular activities. As a consequence, eliminating easily tutored standardized tests helps even the playing field between richer and poorer students.

Every other indicator is also skewed in favour of high-income students. They tend to have parents who are better educated and are interested and much more involved in their children’s education, as well as greater access relevant materials such as books and computers. Tutoring academically, while it may not involve having tutors test for a student, is probably much more impactful in the long-run, they may not be able to do extra-curricular activities but can help with homework (as can engaged parents), so would be just as likely to have an impact on coursework or another method of assessment as it does on the SATs.

Minority students perform less well on tests like the SAT, even when correction is made for income. The test with its reading comprehension test and being done in English is obviously biased against students from households where English is not the dominant language as they are much less likely to understand the questions.

Even when English language skills are not a problem the question choice can contain a cultural bias. One analogy question on the 2003 SAT used the word “regatta” which minority students were unlikely to be familiar with. Having this kind of bias is obviously unfair and can never be fully accounted for in a diverse country such as the United States of America where those who set the wording of the questions will never know all possible perspectives. 

Language will be as much of an issue in understanding in college as it will be for the tests and is likely to be as much of a difficulty in any other method of determining college admissions.

Otherwise these are objections to individual questions on the test, and not to the system itself. The very fact they have made have ensured that the same problems have not re-occurred. Indeed the analogies were dropped in 2005 and scores for the poorest students increased as a result showing that SATs can simply be changed when problems are found with the testing.[1] The problems with a purely grade based system are however intrinsic, and much harder for individuals to improve.

[1] Cloud, ‘What’s Good about the New SAT Test’, 2006

Most tests were designed in an earlier era when far fewer students took them. The large number of students who now take tests like the SATs results in a situation in which the scoring scale magnifies small differences in performance into large differences in scores. Two questions wrong out of 80 on the math section of the SAT may well produce a score of 760 while three questions wrong will result in a 720. 40 points can mean difference between admission and rejection for many candidates, while telling us nothing about the different abilities of the students involved. Indeed on average for 88% of applicants their SAT score will predict their final college grade rank no more accurately than a pair of dice.[1]

[1] Elert, Glenn, ‘The SAT Aptitude of Demographics?’, 5 May 1992, http://hypertextbook.com/eworld/sat.shtml

All measurements are to one degree or another arbitrary. If necessary the scale can be increased, or there can be movement towards a more essay-based exam like the British A-Levels rather than the current multiple-choice format. These changes in fact have already been made with the transformation of the SAT in 2006 which increased the maximum score from 1600 to 2400 and added an essay.[1]

[1] Cloud, ‘What’s Good about the New SAT Test’, 2006

The importance attached to such tests leads to teachers actively “teaching the test.” The result is that many teachers, rather than instilling useful skills or providing a balanced curriculum, end up trying to focus on things that occur on given tests. While this is not a huge problem with the SAT itself, it is a serious problem with subject tests like the SAT 2s, AP Exams, and the British A-Levels. This undermines the provision of education in the country.

There is a degree of hypocrisy in arguing that the tests are classist and racist and then complaining that schools take too long in preparing students for them. Ideally the tests should be on relevant subjects that will be useful to the student and is needed as part of a well-rounded education that prepares the student for life, and if they are not they are flawed.

Many of the skills required for a successful performance on A-Levels or on the AP Essays are remarkably similar to those needed for University level written work. As a consequence it is wrong to argue they are of no relevance.

Standardized tests are inherently arbitrary. They reduce an applicant’s entire academic career to a single one-day session. The result is an inherently unrepresentative test which fails to paint an accurate picture. What if a student has a bad day? What if they do poorly on the specific test questions?  In the SAT’s there is an error of measurement of about 30 points either way out of 800, this is the potential difference between where the student really is and what his or her score on the day was.[1] By contrast, looking at their entire academic record ensures that admissions officers will get a far more comprehensive picture of their actual ability. The law of averages means that bad days and tests will be balanced out with good ones, with the result that their academic record, the result of years of work, will reflect their true performance.

[1] Cloud, John, ‘What’s Good about the New SAT Test’, Time, 1 September 2006, http://www.time.com/time/nation/article/0,8599,1521184,00.html

A student’s academic record tends to measure very different things from what a standardized test does. GPA tends to be based on repetitive assignments such as homework, and in many cases students receive at least some academic credit for simply attending class. By contrast, standardized tests reward ability, by seeing whether or not at the end of the process students actually learned the material in question.  Performance under pressure is an important skill to measure, especially for top institutions, while sifting through the differing standards for what goes into the grades in different school districts is simply not possible.

Standardized tests are inherently skewed in favour of high-income students. Tutoring is and industry unto itself and the resulting rise in scores among high income students skews the scale against students whose families cannot afford it. By contrast, even if the money is available, the impact that tutoring can have on a student’s academic performance is limited. A tutor cannot write papers or take tests for them, and nor can they answer questions in class, or participate in sports or other extracurricular activities. As a consequence, eliminating easily tutored standardized tests helps even the playing field between richer and poorer students.

Every other indicator is also skewed in favour of high-income students. They tend to have parents who are better educated and are interested and much more involved in their children’s education, as well as greater access relevant materials such as books and computers. Tutoring academically, while it may not involve having tutors test for a student, is probably much more impactful in the long-run, they may not be able to do extra-curricular activities but can help with homework (as can engaged parents), so would be just as likely to have an impact on coursework or another method of assessment as it does on the SATs.

Minority students perform less well on tests like the SAT, even when correction is made for income. The test with its reading comprehension test and being done in English is obviously biased against students from households where English is not the dominant language as they are much less likely to understand the questions.

Even when English language skills are not a problem the question choice can contain a cultural bias. One analogy question on the 2003 SAT used the word “regatta” which minority students were unlikely to be familiar with. Having this kind of bias is obviously unfair and can never be fully accounted for in a diverse country such as the United States of America where those who set the wording of the questions will never know all possible perspectives. 

Language will be as much of an issue in understanding in college as it will be for the tests and is likely to be as much of a difficulty in any other method of determining college admissions.

Otherwise these are objections to individual questions on the test, and not to the system itself. The very fact they have made have ensured that the same problems have not re-occurred. Indeed the analogies were dropped in 2005 and scores for the poorest students increased as a result showing that SATs can simply be changed when problems are found with the testing.[1] The problems with a purely grade based system are however intrinsic, and much harder for individuals to improve.

[1] Cloud, ‘What’s Good about the New SAT Test’, 2006

Most tests were designed in an earlier era when far fewer students took them. The large number of students who now take tests like the SATs results in a situation in which the scoring scale magnifies small differences in performance into large differences in scores. Two questions wrong out of 80 on the math section of the SAT may well produce a score of 760 while three questions wrong will result in a 720. 40 points can mean difference between admission and rejection for many candidates, while telling us nothing about the different abilities of the students involved. Indeed on average for 88% of applicants their SAT score will predict their final college grade rank no more accurately than a pair of dice.[1]

[1] Elert, Glenn, ‘The SAT Aptitude of Demographics?’, 5 May 1992, http://hypertextbook.com/eworld/sat.shtml

All measurements are to one degree or another arbitrary. If necessary the scale can be increased, or there can be movement towards a more essay-based exam like the British A-Levels rather than the current multiple-choice format. These changes in fact have already been made with the transformation of the SAT in 2006 which increased the maximum score from 1600 to 2400 and added an essay.[1]

[1] Cloud, ‘What’s Good about the New SAT Test’, 2006

The importance attached to such tests leads to teachers actively “teaching the test.” The result is that many teachers, rather than instilling useful skills or providing a balanced curriculum, end up trying to focus on things that occur on given tests. While this is not a huge problem with the SAT itself, it is a serious problem with subject tests like the SAT 2s, AP Exams, and the British A-Levels. This undermines the provision of education in the country.

There is a degree of hypocrisy in arguing that the tests are classist and racist and then complaining that schools take too long in preparing students for them. Ideally the tests should be on relevant subjects that will be useful to the student and is needed as part of a well-rounded education that prepares the student for life, and if they are not they are flawed.

Many of the skills required for a successful performance on A-Levels or on the AP Essays are remarkably similar to those needed for University level written work. As a consequence it is wrong to argue they are of no relevance.

All measures applied during University admissions processes are arbitrary to one extent or another. Grades depend on the whims on individual instructors, curriculum choices, and school quality. Every single one of the flaws the proposition applies to the SATs or other standardized tests may well be true, but the fact is that they are true for everyone who takes them. The SATs are the only part of the application to a university that every single applicant will likely have been taken, and therefore the only one where all applicants can be objectively compared.

Applicants are not the same. They come from different backgrounds, attend different schools, and have different opportunities. It is “false meritocracy” to compare them with students from an entirely different background. Students should be examined individually, and their performance compared with other students with a similar background. The best way to do this is to allow individual schools to grade their student body, and to then have Universities compare how those students did vis-à-vis their peers within that school.

SATs are mathematical and it is therefore possible to objectively evaluate them. This is why they are so popular, they provide a benchmark of comparison across the whole education system in a way that any non-standardized assessment never could. This does not only benefit universities in providing an objective measure to compare admissions candidates but it also gives the government statistics with which to measure the progress of schools.

Any other form of assessment would mean switching to much more subjective factors. Traditionally such factors, such as extracurricular activities, volunteer work, and even access to references are all more easily available to high income students. Opportunities may not even be offered in poorer school districts. Complaining that poorer and minority students do less well on the SAT ignores the fact that the test provides one of their best opportunities to impress admissions officials.

The objectivity of SATs does not matter to individual students, many of whom are discriminated against by the tests. Subjective factors can be comparative if students’ backgrounds are taken into account. By contrast, the SAT, by its very use and existence, legitimizes its results as an “objective” measure of ability and capacity, which in turn legitimizes schools in discriminating against the vast majority of poor students who do worse on the exam, even if it helps a few out on an individual level. It would be much better to offer preferences in admission to poorer students than to legitimize discrimination.

Subjective criteria can also be successful in determining how successful someone will be in university. In a 1972 study by psychologist Jonathan R. Warre it was found that “Motivation was the quality most frequently cited by over 3400 college teachers” when asked what it takes to succeed in college.[1] Motivation cannot be objectively measured.

[1] Elert, ‘The SAT Aptitude of Demographics?’, 1992