… what IQ tests actually assess is not some universal scale of cognitive strength but the presence of skills and knowledge structures more likely to be acquired in some groups than in others. (Richardson, 2017: 98)
For the past 100 years, the black-white IQ gap has puzzled psychometricians. There are two camps—hereditarians (those who believe that individual and group differences in IQ are due largely to genetics) and environmentalists/interactionists (those who believe that individual and group differences in IQ are largely due to differences in learning, exposure to knowledge, culture and immediate environment).
However, one of the most forceful arguments for the environmentalist (i.e., that the cause for differences in IQ are due to the cultural and social environment; note that an interactionist framework can be used here, too) side is one from Fagan and Holland (2007). They show that half of the questions on IQ tests had no racial bias, whereas other problems on the test were solvable with only a specific type of knowledge – knowledge that is found specifically in the middle class. So if blacks are more likely to be lower class than whites, then what explains lower test scores for blacks is differential exposure to knowledge – specifically, the knowledge to complete the items on the test.
But some hereditarians say otherwise – they claim that since knowledge is easily accessible for everyone, then therefore, everyone who wants to learn something will learn it and thus, the access to information has nothing to do with cultural/social effects.
A hereditarian can, for instance, state that anyone who wants to can learn the types of knowledge that are on IQ tests and that they are widely available everywhere. But racial gaps in IQ stay the same, even though all racial groups have the same access to the specific types of cultural knowledge on IQ tests. Therefore, differences in IQ are not due to differences in one’s immediate environment and what they are exposed to—differences in IQ are due to some innate, genetic differences between blacks and whites. Put into premise and conclusion form, the argument goes something like this:
P1 If racial gaps in IQ were due specifically to differences in knowledge, then anyone who wants to and is able to learn the stuff on the tests can do so for free on the Internet.
P2 Anyone who wants to and is able to learn stuff can do so for free on the Internet.
P3 Blacks score lower than whites on IQ tests, even though they have the same access to information if they would like to seek it out.
C Therefore, differences in IQ between races are due to innate, genetic factors, not any environmental ones.
This argument is strange. One would have to assume that blacks and whites have the same access to knowledge—we know that lower-income people have less access to knowledge in virtue of the environments they live in. For instance, they may have libraries with low funding or bad schools with teachers who do not care enough to teach the students what they need to succeed on these standardized tests (IQ tests, the SAT, etc are all different versions of the same test). (2) One would have to assume that everyone has the same type of motivation to learn what amounts to answers for questions on a test that have no real-world implications. And (3) the type of knowledge that one is exposed to dictates what one can tap into while they are attempting to solve a problem. All three of these reasons can cascade in causing the racial performance in IQ.
Familiarity with the items on the tests influences a faster processing of information, allowing one to correctly identify an answer in a shorter period of time. If we look at IQ tests as tests of middle-class knowledge of skills, and we rightly observe that blacks are lower class than whites who are more likely to be middle class, then it logically follows that the cause of differences in IQ between blacks and whites are cultural – and not genetic – in origin. This paper – and others – solves the century-old debate on racial IQ differences – what accounts for differences in IQ scores is differential exposure to knowledge. Claiming that people have the same type of access to knowledge and, thusly, won’t learn it if they won’t seek it out does not make sense.
Differing experiences lead to differing amounts of knowledge. If differing experiences lead to differing amounts of knowledge, and IQ tests are tests of knowledge—culturally-specific knowledge—then those who are not exposed to the knowledge on the test will score lower than those who are exposed to the knowledge. Therefore, Jensen’s Default Hypothesis is false (Fagan and Holland, 2002). Fagan and Holland (2002) compared blacks and whites on for their knowledge of the meaning of words, which are highly “g”-loaded and shows black-white differences. They review research showing that blacks have lower exposure to words and are therefore unfamiliar with certain words (keep this in mind for the end). They mixed in novel words with previously-known words to see if there was a difference.
Fagan and Holland (2002) picked out random words from the dictionary, then putting them into a sentence to attempt to give the testee some context. They carried out five experiments in all, and each one showed that, when equal opportunity was given to the groups, they were “equal in knowledge” (IQ). So, whites were more likely to know the items more likely to be found on IQ tests. Thus, there were no racial differences between blacks and whites when looked at from an information-processing point of view. Therefore, to expain racial differences in IQ, we must look to differences in the cultural/social environment. Fagan (2000) for instance, states that “Cultures may differ in the types of knowledge their members have but not in how well they process. Cultures may account for racial differences in IQ.”
The results of Fagan and Holland (2002) are completely at-ends with Jensen’s Default Hypothesis—that the 15-point gap in IQ is due to the same environmental and cultural factors that underlie individual differences in the group. However, as Fagan and Holland (2002: 382) show that:
Contrary to what the default hypothesis would predict, however, the within racial group analyses in our study stand in sharp contrast to our between racial group findings. Specifically, individuals within a racial group who differed in general knowledge of word meanings also differed in performance when equal exposure to the information to be tested was provided. Thus, our results suggest that the average difference of 15 IQ points between Blacks and Whites is not due to the same genetic and environmental factors, in the same ratio, that account for differences among individuals within a racial group in IQ.
Exposure to information is critical, in fact. For instance, Ceci (1996) shows that familiarity with words dictates speed of processing to use in identifying the correct answer to the problem. In regard to differences in IQ, Ceci (1996) does not deny the role of biology—indeed, it’s a part of his bio-ecological model of IQ, which is a theory that postulates the development of intelligence as an interaction between biological dispositions and the environment in which those dispositions manifest themselves. Ceci (1996) does note that there are biological constraints on intelligence, but that “… individual differences in biological constraints on specific cognitive abilities are not necessarily (or even probably) directly responsible for producing the individual differences that have been reported in the psychometric literature.” That such potentials, though may be “genetic” in origin, of course, does not license the claim that genetic factors contribute to variance in IQ. “Everyone may possess them to the same degree, and the variance may be due to environment and/or motivations that led to their differential crystallization.” (Ceci, 1996: 171)
Ceci (1996) also further shows that people can differ in intellectual performance due to 3 things: (1) the efficiency of underlying cognitive potentials that are relevant to the cognitive ability in question; (2) the structure of knowledge relevant to the performance; and (3) contextual/motivational factors relevant to crystallize the underlying potentials gained through one’s knowledge. Thus, if one is lacking in the knowledge of the items on the test due to what they learned in school, then the test will be biased against them since they did not learn the relevant information on the tests.
Cahan and Cohen (1989) note that nine-year-olds in fourth grade had higher IQs than nine-year-olds in third grade. This is to be expected, if we take IQ scores as indices of—cultural-specific—knowledge and skills and this is because fourth-graders have been exposed to more information than third-graders. In virtue of being exposed to more information than their same-age cohort in different grades, they then score higher on IQ tests because they are exposed to more information.
Cockroft et al (2015) studied South African and British undergrads on the WAIS-III. They conclude that “the majority of the subtests in the WAIS-III hold cross-cultural biases“, while this is “most evident in tasks which tap crystallized, long-term learning, irrespective of whether the format is verbal or non-verbal” so “This challenges the view that visuo-spatial and non-verbal tests tend to be culturally fairer than verbal ones (Rosselli and Ardila, 2003)”.
IQ tests “simply reflect the different kinds of learning by children from different (sub)cultures: in other words, a measure of learning, not learning ability, and are merely a redescription of the class structure of society, not its causes … it will always be quite impossible to measure such ability with an instrument that depends on learning in one particular culture” (Richardson, 2017: 99-100). This is the logical position to hold: for if IQ tests test class-specific type of knowledge and certain classes are not exposed to said items, then they will score lower. Therefore, since IQ tests are tests of a certain kind of knowledge, IQ tests cannot be “a measure of learning ability” and so, contra Gottfredson, ‘g’ or ‘intelligence’ (IQ test scores) cannot be called “basic learning ability” since we cannot create culture—knowledge—free tests because all human cognizing takes place in a cultural context which it cannot be divorced from.
Since all human cognition takes place through the medium of cultural/psychological tools, the very idea of a culture-free test is, as Cole (1999) notes, ‘a contradiction in terms . . . by its very nature, IQ testing is culture bound’ (p. 646). Individuals are simply more or less prepared for dealing with the cognitive and linguistic structures built in to the particular items. (Richardson, 2002: 293)
Heine (2017: 187) gives some examples of the World War I Alpha Test:
1. The Percheron is a kind of
(a) goat, (b) horse, (c) cow, (d) sheep.
2. The most prominent industry of Gloucester is
(a) fishing, (b) packing, (c) brewing, (d) automobiles.
3. “There’s a reason” is an advertisement for
(a) drink, (b) revolver, (c) flour, (d) cleanser.
4. The Knight engine is used in the
(a) drink, (b) Stearns, (c) Lozier, (d) Pierce Arrow.
5. The Stanchion is used in
(a) fishing, (b) hunting, (c) farming, (d) motoring.
Such test items are similar to what are on modern-day IQ tests. See, for example, Castles (2013: 150) who writes:
One section of the WAIS-III, for example, consists of arithmetic problems that the respondent must solve in his or her head. Others require test-takers to define a series of vocabulary words (many of which would be familiar only to skilled-readers), to answer school-related factual questions (e.g., “Who was the first president of the United States?” or “Who wrote the Canterbury Tales?”), and to recognize and endorse common cultural norms and values (e.g., “What should you do it a sale clerk accidentally gives you too much change?” or “Why does our Constitution call for division of powers?”). True, respondents are also given a few opportunities to solve novel problems (e.g., copying a series of abstract designs with colored blocks). But even these supposedly culture-fair items require an understanding of social conventions, familiarity with objects specific to American culture, and/or experience working with geometric shapes or symbols. [Since this is questions found on the WAIS-III, then go back and read Cockroft et al, 2015 since they used the British version which, of course, is similar.]
If one is not exposed to the structure of the test along with the items and information on them, how, then, can we say that the test is ‘fair’ to other cultural groups (social classes included)? For, if all tests are culture-bound and different groups of people have different cultures, histories, etc, then they will score differently by virtue of what they know. This is why it is ridiculous to state so confidently that IQ tests—however imperfectly—test “intelligence.” They test certain skills and knowledge more likely to be found in certain groups/classes over others—specifically in the dominant group. So what dictates IQ scores is differential access to knowledge (i.e., cultural tools) and how to use such cultural tools (which then become psychological tools.)
Lastly, take an Amazonian people called The Pirah. They have a different counting system than we do in the West called the “one-two-many system, where quantities beyond two are not counted but are simply referred to as “many”” (Gordon, 2005: 496). A Pirah adult was shown an empty can. Then the investigator put six nuts into the can and took five out, one at a time. The investigator then asked the adult if there were any nuts remaining in the can—the man answered that he had no idea. Everett (2005: 622) notes that “Piraha is the only language known without number, numerals, or a concept of counting. It also lacks terms for quantification such as “all,” “each,” “every,” “most,” and “some.””
(hbdchick, quite stupidly, on Twitter wrote “remember when supermisdreavus suggested that the tsimane (who only count to s’thing like two and beyond that it’s “many”) maybe went down an evolutionary pathway in which they *lost* such numbers genes?” Riiiight. Surely the Tsimane “went down an evolutionary pathway in which they *lost* such numbers genes.” This is the idiocy of “HBDers” in action. Of course, I wouldn’t expect them to read the actual literature beyond knowing something basic (Tsimane numbers beyond “two” are known as “many”) and the positing a just-so story for why they don’t count above “two.”
Take a non-verbal test, such as the Bender-Gestalt test. There are nine index cards which have different geometrical designs on them, and the testee needs to copy what he saw before the next card is shown. The testee is then scored on how accurate his recreation of the index card is. Seems culture-fair, no? It’s just shapes and other similar things, how would that be influenced by class and culture? One would, on a cursory basis, claim that such tests have no basis in knowledge structure and exposure and so would rightly be called “culture-free.” While the shapes that come on Ravens tests are novel, the rules governing them are not.
Hoffmann (1966) studied 80 children (20 Kickapoo Indians (KIs), 20 low SES blacks (LSBs), 20 low SES whites (LSWs), and 20 middle-class whites (MCWs)) on the Bender-Gestalt test. The Kickapoo were selected from 5 urban schools; 20 blacks from majority-black elementary schools in Oklahoma City; 20 whites in low SES areas of Oklahoma; and 20 whites from middle-schools in Oklahoma from majority-white schools. All of the children were aged 8-10 years of age and in the third grade, while all had IQs in the range of 90-110. They were matched on a whole slew of different variables. Hoffman (1966: 52) states “that variations in cultural and socio-economic background affect Bender Gestalt reproduction.”
Hoffman (1966: 86) writes that:
since the four groups were shown to exhibit no significant differences in motor, or perceptual discrimination ability it follows that differences among the four groups of boys in Bender Gestalt performance are assignable to interpretative factors. Furthermore, significant differences among the four groups in Bender performance illustrates that the Bender Gestalt test is indeed not a so called “culture-free” test.
Hoffman concluded that MCWs, KIs, LSBs, and LSWs did not differ in copying ability, nor did they differ significantly in discriminating in different phases in the Bender-Gestalt; there also was no bias in figures that had two of the different sexes on them. They did differ in their reproductions of Bender-Gestalt designs, and their differing performance can be, of course, interpreted differently by different people. If we start from the assumption that all IQ tests are culture-bound (Cole, 2004), then living in a different culture from the majority culture will have one score differently by virtue of having differing—culture-specific knowledge and experience. The four groups looked at the test in different ways, too. Thus, the main conclusion is that:
The Bender Gestalt test is not a “culture-free” test. Cultural and socio-economic background appear to significantly affect Bender Gestalt reproduction. (Hoffman, 1966: 88)
Drame and Ferguson (2017) and Dutton et al (2017) also show that there is bias in the Raven’s test in Mali and Sudan. This, of course, is due to the exposure to the types of problems on the items (Richardson, 2002: 291-293). Thus, their cultures do not allow exposure to the items on the test and they will, therefore, score lower in virtue of not being exposed to the items on the test. Richardson (1991) took 10 of the hardest Raven’s items and couched them in familiar terms with familiar, non-geometric, objects. Twenty eleven-year-olds performed way better with the new items than the original ones, even though they used the same exact logic in the problems that Richardson (1991) devised. This, obviously, shows that the Raven is not a “culture-free” measure of inductive and deductive logic.
The Raven is administered in a testing environment, which is a cultural device. They are then handed a paper with black and white figures ordered from left to right. Note that Abel-Kalek and Raven (2006: 171) write that Raven’s items “were transposed to read from right to left following the custom of Arabic writing.” So this is another way that the tests are biased and therefore not “culture-free.”) Richardson (2000: 164) writes that:
For example, one rule simply consists of the addition or subtraction of a figure as we move along a row or down a column; another might consist of substituting elements. My point is that these are emphatically culture-loaded, in the sense that they reflect further information-handling tools for storing and extracting information from the text, from tables of figures, from accounts or timetables, and so on, all of which are more prominent in some cultures and subcultures than others.
Richardson (1991: 83) quotes Keating and Maclean (1987: 243) who argue that tests like the Raven “tap highly formal and specific school skills related to text processing and decontextualized rule application, and are thus the most systematically acculturated tests” (their emphasis). Keating and Maclean (1987: 244) also state that the variation in scores between individuals is due to “the degree of acculturation to the mainstream school skills of Western society” (their emphasis). That’s the thing: all types of testing is biased towards a certain culture in virtue of the kinds of things they are exposed to—not being exposed to the items and structure of the test means that it is in effect biased against certain cultural/social groups.
Davis (2014) studied the Tsimane, a people from Bolivia, on the Raven. Average eleven-year-olds 78 percent or more of the questions correct whereas lower-performing individuals answered 47 percent correct. The eleven-year-old Tsimane, though, only answered 31 percent correct. There was another group of Tsimane who went to school and lived in villages—not living in the rainforest like the other group of Tsimane. They ended up scoring 72 percent correct, compared to the unschooled Tsimane who scored only 31 percent correct. “… the cognitive skills of the Tsimane have developed to master the challenges that their environment places on them, and the Raven’s test simply does not tap into those skills. It’s not a reflection of some kind of true universal intelligence; it just reflects how well they can answer those items” (Heine, 2017: 189). Thus, measures of “intelligence” are not an innate skill, but are learned through experience—what we learn from our environments.
Heine (2017: 190) discusses as-of-yet-to-be-published results on the Hadza who are known as “the most cognitively complex foragers on Earth.” So, “the most cognitively complex foragers on Earth” should be pretty “smart”, right? Well, the Hadza were given six-piece jigsaw puzzles to complete—the kinds of puzzles that American four-year-olds do for fun. They had never seen such puzzles before and so were stumped as to how to complete them. Even those who were able to complete them took several minutes to complete them. Is the conclusion then licensed that “Hadza are less smart than four-year-old American children?” No! As that is a specific cultural tool that the Hadza have never seen before and so, their performance mirrored their ignorance to the test.
The term “logical” comes from the Greek term logos, meaning “reason, idea, or word.” So, “logical reasoning” is based on reason and sound ideas, irrespective of bias and emotion. A simple syllogistic structure could be:
If X, then Y
We can substitute terms, too, for instance:
If it rains today, then I must bring an umbrella.
It’s raining today.
∴ I must bring an umbrella.
Richardson (2000: 161) notes how cross-cultural studies show that what is or is not logical thinking is not objective nor simple, but “comes in socially determined forms.” He notes how cross-cultural psychologist Sylvia Scribner showed some syllogisms to Kpelle farmers, which were couched in terms that were familiar to them. One syllogism given to them was:
All Kpelle men are rice farmers
Mr. Smith is not a rice farmer
Is he a Kpelle man? (Richardson, 2002: 162)
The individual then continuously replied that he did not know Mr. Smith, so how could he know whether or not he was a Kpelle man? Another example was:
All people who own a house pay a house tax
Boima does not pay a house tax
Does Boima own a house? (Richardson, 2000: 162)
The answer here was that Boima did not have any money to pay a house tax.
In regard to the first syllogism, Mr. Smith is not a rice farmer so he is not a Kpelle man. Regarding the second, Boima does not pay a house tax, so Boima does not own a house. The individual could give a syllogism that is something like:
All the deductions I can make are about individuals I know.
I do not know Mr. Smith.
Therefore I cannot make a deduction about Mr. Smith. (Richardson, 2000: 162)
They are using what are familiar terms to them, and so, they get the answer right for their culture based on the knowledge that they have. These examples, therefore, show that what can pass for “logical reasoning” is based on the time and place where it is said. The deductions the Kpelle made were perfectly valid, though they were not what the syllogism-designers had in mind. In fact, I would say that there are many—equally valid—ways of answering such syllogisms, and such answers will vary by culture and custom.
The bio-ecological framework, culture, and social class
The bio-ecological model of Ceci and Bronfenbrenner is a model of human development that relies on gene and environment interactions. The model is a Vygotskian one—in that learning is a social process where the support from parents, teachers, and all of society play an important role in the ontogeny of higher psychological functioning. (For a good primer on Vygotskian theory, see Vygotsky and the Social Formation of Mind, Wertsch, 1985.) Thus, it is a model of human development that, most hereditarians would say, that “they use too.” Though this is of course, contested by Ceci who compares his bio-ecological framework with other theories (Ceci, 1996: 220, table 10.1):
Cognition (thinking) is extremely context-sensitive. Along with many ecological influences, individual differences in cognition are understood best with the bio-ecological framework which consists of three components: (1) ‘g’ doesn’t exist, but multiple cognitive potentials do; (2) motivational forces and social/physical aspects of a task or setting, how elaborate a knowledge domain is not only important in the development of the human, but also, of course, during testing; and (3) knowledge and aptitude are inseparable “such that cognitive potentials continuously access one’s knowledge base in the cascading process of producing cognitions, which in turn alter the contents and structure of the knowledge base” (Ceci, 1996: 123).
Block (1995) notes that “Blacks and Whites are to some extent separate cultural groups.” Sternberg (2004) defines culture as “the set of attitudes, values, beliefs and behaviors shared by a group of people, communicated from one generation to the next via language or some other means of communication.” In regard to social class—blacks and whites differ in social class (a form of culture), Richardson (2002: 298) notes that “Social class is a compound of the cultural tools (knowledge and cognitive and psycholingustic structures) individuals are exposed to; and beliefs, values, academic orientations, self-efficacy beliefs, and so on.” The APA notes that “Social status isn’t just about the cars we drive, the money we make or the schools we attend — it’s also about how we feel, think and act …” And the APS notes that social class can be seen as a form of culture. Since culture is a set of attitudes, beliefs and behaviors shared by a group of people, social classes, therefore, are forms of culture as different classes have different attitudes, beliefs and behaviors.
Ceci (1996 119) notes that:
large-scale cultural differences are likely to affect cognition in important ways. One’s way of thinking about things is determined in the course of interactions with others of the same culture; that is, the meaning of a cultural context is always negotiated between people of that culture. This, in turn, modifies both culture and thought.
While Manstead (2018) argues that:
There is solid evidence that the material circumstances in which people develop and live their lives have a profound influence on the ways in which they construe themselves and their social environments. The resulting differences in the ways that working‐class and middle‐ and upper‐class people think and act serve to reinforce these influences of social class background, making it harder for working‐class individuals to benefit from the kinds of educational and employment opportunities that would increase social mobility and thereby improve their material circumstances.
In fact, the bio-ecological model of human development (and IQ) is a developmental systems-type model. The types of things that go into the model are just like Richardson’s (2002) “sociocognitive affective nexus.” Richardson (2002) posits that the sources of IQ variation are mostly non-cognitive, writing that such factors include (pg 288):
(a) the extent to which people of different social classes and cultures have acquired a specific form of intelligence (or forms of knowledge and reasoning); (b) related variation in ‘academic orientation’ and ‘self-efficacy beliefs’; and (c) related variation in test anxiety, self-confidence, and so on, which affect performance in testing situations irrespective of actual ability
Cole (2004) concludes that:
Our imagined study of cross-cultural test construction makes it clear that tests of ability are inevitably cultural devices. This conclusion must seem dreary and disappointing to people who have been working to construct valid, culture-free tests. But from the perspective of history and logic, it simply confirms the fact, stated so clearly by Franz Boas half a century ago, that “mind, independent of experience, is inconceivable.”
It is a noteworthy feature of all preceding (cognitive approaches) that they make no reference whatsoever to the environment in which the person actually lives and grows. The implicit assumption is that the attributes in question are constant across place; the person carries them with her wherever she goes. Stating the issue more theoretically, the assumption is that the nature of the attribute does not change, irrespective of the context in which one finds one’s self.
Such contextual differences can be found in the intrinsic and extrinsic motivations of the individual in question. Self-efficacy, what one learns and how they learn it, motivation instilled from parents, all form part of the context of the specific individual and how they develop which then influences IQ scores (middle-class knowledge and skills scores).
If blacks and whites are, to some extent, different cultural groups, then they will—by definition—have differing cultures. So “cultural differences are known to exist, and cultural differences can have an impact on psychological traits [also in the knowledge one acquires which then is one part of dictating test scores] (see Prinz, 2014: 67, Beyond Human Nature). If blacks and whites are “separate cultural groups” (Block, 1995) and if they have different experiences by virtue of being cultural groups, then they will score differently on any test of ability (including IQ; see Fagan and Holland, 2002, 2007) as all tests of ability are culture-bound (see Cole, 2004).
1 Blacks and whites are different cultural groups.
2 If (1), then they will have different experiences by virtue of being different cultural groups.
3 So blacks and whites, being different cultural groups, will score differently on tests of ability, since they are exposed to different knowledge structures due to their different cultures and so, all tests of ability are culture-bound.
So, what accounts for the intercorrelations between tests of “cognitive ability”? They validate the new test with older, ‘more established’ tests so “based on this it is unlikely that a measure unrelated to g will emerge as a winner in current practice … [so] it is no wonder that the intelligence hierarchy for different racial/ethnic groups remains consistent across different measures. The tests are highly correlated among each other and are similar in item structure and format” (Suzuki and Aronson, 2005: 321).
Therefore, what accounts for differences in IQ is not intellectual ability, but cultural/social exposure to information—specifically the type of information used in the construction of IQ tests—along with the test constructors attempting to construct new tests that correlate with the old tests, and so, they get the foregone conclusion of their being racial differences, for example, in IQ which they trumpet as evidence for a “biological cause”—but it is anything but: such differences are built into the test (Simon, 1997). (Note that Fagan and Holland, 2002 also found evidence for test bias as well.)
Thus, we should take the logical conclusion: what explains racial IQ differences are not biological factors, but environmental ones—specifically in the exposure of knowledge—along with how new tests are created (see Suzuki and . All human cognizing takes place in specific cultural contexts—therefore “culture-free tests” (i.e., tests devoid of cultural knowledge and context) are an impossibility. IQ tests are experience-dependent so if one is not exposed to the relevant experiences to do well in a testing situation, then they will score lower than they would have if they were to have the requisite culturally-specific knowledge to perform well on the test.