Journal of Autism and Developmental Disorders, Vol. 31, No.

1, 2001

The Autism-Spectrum Quotient (AQ): Evidence from Asperger Syndrome/High-Functioning Autism, Males and Females, Scientists and Mathematicians
Simon Baron-Cohen,1 Sally Wheelwright,1 Richard Skinner,1 Joanne Martin,1 and Emma Clubley1

Currently there are no brief, self-administered instruments for measuring the degree to which an adult with normal intelligence has the traits associated with the autistic spectrum. In this paper, we report on a new instrument to assess this: the Autism-Spectrum Quotient (AQ). Individuals score in the range 0–50. Four groups of subjects were assessed: Group 1: 58 adults with Asperger syndrome (AS) or high-functioning autism (HFA); Group 2: 174 randomly selected controls. Group 3: 840 students in Cambridge University; and Group 4: 16 winners of the UK Mathematics Olympiad. The adults with AS/HFA had a mean AQ score of 35.8 (SD = 6.5), significantly higher than Group 2 controls (M = 16.4, SD = 6.3). 80% of the adults with AS/HFA scored 32+, versus 2% of controls. Among the controls, men scored slightly but significantly higher than women. No women scored extremely highly (AQ score 34+) whereas 4% of men did so. Twice as many men (40%) as women (21%) scored at intermediate levels (AQ score 20+). Among the AS/HFA group, male and female scores did not differ significantly. The students in Cambridge University did not differ from the randomly selected control group, but scientists (including mathematicians) scored significantly higher than both humanities and social sciences students, confirming an earlier study that autistic conditions are associated with scientific skills. Within the sciences, mathematicians scored highest. This was replicated in Group 4, the Mathematics Olympiad winners scoring significantly higher than the male Cambridge humanities students. 6% of the student sample scored 32+ on the AQ. On interview, 11 out of 11 of these met three or more DSM-IV criteria for AS/HFA, and all were studying sciences/mathematics, and 7 of the 11 met threshold on these criteria. Test–retest and interrater reliability of the AQ was good. The AQ is thus a valuable instrument for rapidly quantifying where any given individual is situated on the continuum from autism to normality. Its potential for screening for autism spectrum conditions in adults of normal intelligence remains to be fully explored.
KEY WORDS: Autism-Spectrum Quotient; Asperger syndrome; high-functioning autism; normal intelligence.

INTRODUCTION Autism is defined in terms of abnormalities in social and communication development, in the presence of marked repetitive behavior and limited imagination
1

Departments of Experimental Psychology and Psychiatry, University of Cambridge, Downing Street, Cambridge, CB2 3EB, United Kingdom.

(American Psychiatric Association [APA], 1994). Asperger syndrome (AS) is defined in terms of the individual meeting the same criteria for autism but with no history of cognitive or language delay, and not meeting the criteria for Pervasive Development Disorder (PDD), (ICD-10; World Health Organization, 1994). Language delay itself is defined as not using single words by 2 years of age, and/or phrase speech by 3 years of age. There is growing evidence that autism and AS are of genetic 5
0162-3257/01/0200-0005$19.50/0 © 2001 Plenum Publishing Corporation

rating their own behavior as more appropriate than it might really be. 1995. still under debate. poor imagination. made up of 10 questions assessing 5 different areas: social skill (items 1. 47. where we had the opportunity to ask them about their responses. or the degree of caseness of an individual. in scientific comparisons).11.6. self-administered scale for identifying the degree to which any individual adult of normal IQ may have “autistic traits” or what has been called “the broader phenotype” (Bailey et al.46). Abnormality = poor social skill.18.39). Bolton & Rutter. parents independently completed an AQ for their child with AS/HFA. For both of these reasons. but this instrument is designed for high-functioning individuals who are perfectly able to read or discuss issues. rather than only asking them to judge their 2 The term “quotient” is not used in the arithmetic sense (the result of dividing one quantity by another) but as derived from the Latin root quotiens (how much or how many). screening for possibly affected individuals to assist in making referrals for a full diagnostic assessment). Lord. attention to detail (items 5.22. mentioned above.4. over several years. if anything. 35. and Clubley the respondent records the abnormal or autistic-like behavior either mildly or strongly (see below for scoring each item. with AS as the bridge between autism and normality (Baron-Cohen.2 DESIGN OF THE AQ The AQ was designed to be short. that autism and AS lie on a continuum of social-communication disability. lies on this continuum. & Robertson.49). Due to the concern over whether a condition like HFA or AS might impair one’s ability to understand the items in the questionnaire. There is also an assumption. and half an “agree” response. and age-matched controls. It is shown in the Appendix. attention switching (items 2.45. that there is an autism spectrum (Wing. Wheelwright. the ADOS-G (Autism Diagnostic Observation Schedule) are fairly time-consuming to administer. Comprehension of wording might be a greater problem in a less able population.8. This would be useful for both scientific reasons (e. 1979). Reichler..28.17. 1995.25. we included questions in both the social and communication domains that ask about the person’s preferences.14. items were randomized with respect to both the expected response from a high-scorer. Existing instruments. 1978. What is needed is a short. Following piloting. 1989.g. 1994. with normal intelligence. Mortimore.9. 1990.. due to subtle mind-reading problems (Baron-Cohen. Folstein & Rutter.15. 1981. 29.23. Rutter. Currently there are no brief. 1995.26. We did this by calling some patients into our lab. such as the ADI-R (Autism Diagnostic Interview) (Le Couteur et al. this would lead a person to score lower on the AQ. however. Baron-Cohen. & Renner.41. Jolliffe. poor communication skill.g. Frith. imagination (items 3. Approximately half the items were worded to produce a “disagree” response.31. 1986).40. 1991.38. 1994)..27.19. 1997).21. easy to use. . Rutter.6 Baron-Cohen. A related issue is whether a condition like AS or HFA might impair the subject’s ability to judge their own social or communicative behavior. and required the coding of responses. 1988). exceptional attention to detail. If this occurred. is not self-administered (Schopler.43. Family pedigrees of AS also implicate heritability (Gillberg. Each of the items listed above scores 1 point if INSTRUMENT DEVELOPMENT Items were selected from the domains in the “triad” of autistic symptoms (APA.20. we checked comprehension with the patients in our pilot study.44. and with respect to their domain. Martin. 37.. items which controls scored on as often. selected at random.36. The AQ as shown in the Appendix is the outcome of piloting multiple versions. to guard against false negatives. Any inaccuracies of this kind would therefore. This was to avoid a response bias either way. It comprises 50 questions.13.30. origin. & Le Couteur. The evidence is strongest for autism. communication (items 7.16.10. and easy to score. 42. and comes from twin and behavioral genetic family studies (Bailey et al. The instrument’s name was chosen because of the assumption.34.24. Wing. The instrument was piloted on adults with AS or high-functioning autism (HFA). 1991). lead to a conservative estimate of the person’s true AQ score.32. An early version was also interview-based. and from demonstrated areas of cognitive abnormality in autism.33. establishing who is “affected” and who is not. 1995). The continuum view shifts us away from categorical diagnosis and towards a quantitative approach.. Following this. poor attention-switching/strong focus of attention).12. 1977. 1988). Wing & Gould. we developed the Autism-Spectrum Quotient (AQ).48).50). Skinner. 1988). and potentially for applied reasons (e. self-administered instruments available for measuring where any given individual adult. or more often. and the CARS (Childhood Autism Rating Scale) which can be brief. than did people with autism/AS were omitted. in a high scoring person with AS/HFA. However. For caution.

0). Cicchetti. They were included as a retest of this same association. and their mix of occupations was similar to that Group 1. 14. and is straightforward to score since it does not depend on any interpretation in the scoring. 386 female). 34. and mathematicians. Finally.The Autism-Spectrum Quotient own behavior. They were drawn from 500 adults sent the AQ by post. 30. The return rates from the different disciplines did not differ significantly. On this basis. 17. Volkmar. “Definitely disagree” or “slightly disagree” responses scored 1 point. 11. small sounds). 46. 43. items in the other domains ask about their attentional preferences or focus of attention (e.8. 47. 49. that there was a main effect . on the following items: 1. See below for a check of this. giving a return rate of 20.1). 41.5. 7 They were drawn from 4. 20. engineers. (M = 105. Group 3 was included to test if they showed a similar profile to the randomly selected controls (Group 2. In Groups 1 and 2. range = 16. Method Participants were sent the AQ by post. 1 female). 2. giving a return rate of 34. 27.5:1 (M:F) is similar to that found in other samples (Klin. 19. The final sample of 58 were those who responded from a larger sample of 63. 42. 22. The final version of the AQ has a forced-choice format. rather than attempting to separate them into AS versus HFA.6–51. and Sex Differences Mean total and subcategory AQ scores from each group are displayed in Table I. To confirm the diagnosis of adults in Group 1 being high functioning. using four subtests of the WAIS-R (see below). 8.6 years (SD = 11. 32. on the following items: 3. Their mean age was 31. 29. Their mean number of years in education was 14. 36. Participants in Group 2 had the option to complete this anonymously or not. 31.. Block Design.34).7). including the National Autistic Society (UK). given earlier reports (Baron-Cohen et al. can be self-administered.3).41). 15 were randomly selected and invited into the lab for intellectual assessment using four subtests of the WAIS-R (Wechsler.8%.1–60. 48. Their mean number of years in education was 13. to dates. Group 3 comprised 840 students in Cambridge University (454 male. 50. and they were instructed to complete it as quickly as possible (to avoid thinking about responses too long). 21.3). p > . 7. 23. that is. 15 individuals were randomly selected from the individuals who had returned an AQ and invited into the lab to check prorated IQ. Comparing Groups 1 and 2 using an ANOVA of total AQ score by Group and Sex. 37. range = 17. Sparrow. They were recruited via several sources. SD = 6. RESULTS AS/HFA Versus Controls. and Picture Completion. They had all attended mainstream schooling and were reported to have an IQ in the normal range. these individuals are grouped together.8. This sex ratio of 3. despite the difference in both IQ and educational level of the two groups. all of these had a prorated IQ of at least 85. and advertisements in newsletters/web pages for adults with AS/HFA.5–58. They were all living in the East Anglia area. range = 18. 89 had university degrees. Scoring the AQ “Definitely agree” or “slightly agree” responses scored 1 point. 26. Group 4 comprised 16 winners of the UK Mathematics Olympiad (15 male. 18. 28. 12. specialist clinics carrying out diagnostic assessments. Piloting revealed that such able subjects were certainly able to report on their own preferences and what they find easy or difficult. 45.9. Because we could not confirm age of onset of language with any reliability (due to the considerable passage of time).5).0 years (SD = 7. & Rourke. 13. 1994). There is no reason to expect that a highfunctioning person with autism or AS would be at all impaired in being able to report faithfully on such items. Similarities.9 ( SD = 2. as predicted.1%. Equally. 24. and did not differ significantly from the subsample ( n = 15) selected from Group 1 (t test. Their occupations reflected their mixed socioeconomic status (SES). 1998) suggesting that autism is more common in families of physicists. 25.g. Their mean age was 21. Their mean age was 37.4 years (SD = 1. 39. 1995). Thirty-two had higher educational qualifications (university degrees). 35. 9. 5. 1958). 38. Subjects Four groups of subjects were tested: Group 1 comprised 58 adults with AS/HFA (45 male.0.175 students sent an AQ. above). numbers. All subjects in this group had been diagnosed by psychiatrists using established criteria for autism or AS (APA. in the normal range (M = 106. range = 15. SD = 8. Their mean age was 17. 33. 4. The four subtests of the WAIS-R were Vocabulary. 15. Group 2 comprised 174 adults selected at random (76 male and 98 female).. 10. 16.2 ( SD = 2. Group 3 also allowed us to test if scientists differed from students in the humanities.0 years (SD = 2.3–18. 44. 40. we found.7.0). 6. and to complete it on their own. 13 female).

7 76 98 840 2.0001.2 2.8 2. males scoring higher than females.8 16.9 1.3 2. 1010) = 3.0 1.1 5.9 7.7 2.4 2.9 1.6 1. neurophysiology. F(1. F(1.4 2.3 2. Martin. p = .2 4.3 biological sciences. biology.0 2.015.9 2.6 2.0 7.0001).8 6.0 1. Group means on each subscore are also shown in Table I.5 6. p = .3 3.84. p < .9 2.0 7.3 1.6 2.7 1.1 1.01. Scientists Versus Nonscientists Table II shows the AQ scores for subjects in Group 3. physical natural sciences.4 6. material science.1 6.0001).0 4.01). the control males scoring significantly higher than the control females (t = 2.5 1. 1010) = 0.7 5.6 2. 1010) = 19. There was no difference between mean AQ scores of men and women with AS/HFA.2 2. chemistry. communications.0001. and no Group × Sex interaction. and molecular ecology.5 1.0 1.042.9 1.0 4.0 2.56.9 24.1 3.8 2.9 38.9 2. Combining Groups 2 and 3. neuroscience. 228) = 328.1 6.2 2. We compared students studying Science (i. all p < .5 5. p = .4 18. and a two-way interaction of Group × Sex.07. and geophysics.3 5.9 4.9 3. This means that on the AQ the students do not differ from the general population sample. ..3 4.2 2.8 6.5 2.9 1.2 2. despite the differ- ences in IQ and educational level between the two groups.0 7. F(1. Skinner.3 5.3 1.7 2.4 1. and Clubley Table I.3 2. chemical engineering.0 2.1 7. df = x.1 8.7 2.3 6.9 2.9 8. computer 3 Physical sciences included physics.9.2 2.3 5.4 5.0 2.8 2.6 6.7 of Group.3 4.5 2. Comparing the students (Group 3) to the randomly selected controls (Group 2).9 1. Mean AQ and Subscale Scores by Group n Communication Social Imagination Local details Attention switching Total AQ Group 1 AS/HFA M SD AS/HFA males M SD AS/HFA females M SD Group 2 Controls M SD Control males M SD Control females M SD Group 3 Students M SD Student males M SD Student Females M SD Group 4 Olympiad M SD 58 7.4 mathematics. broken down according to their Degree/area of study.4 2.4 2.8 2. 228) = 6. Wheelwright. biological natural sciences.8 7.8 Baron-Cohen. the AS/HFA group scoring higher than the controls.8 15.5 35. F(1.3 2.0 2. 4 Biological sciences included experimental psychology.3 2.0 35. p = .3 2.9 2. there was no main effect of Group.1 4.6 6.1 454 386 16 3. but there was a significant effect of Sex.7 1.7 1.2 5.0 2.3 17. F(1.2 2. The AS/HFA group differed from Group 2 on all subscores (t tests. See also Figs.3 2.6 2.9 1.9 2. p < .5 5. men and women differed on all subscales except local details (t tests.3 2.6 16.2 7.5 6.4 45 13 174 2. geology. physical sciences. bioanthroplogy.2.0 17.0 3.3 6.e.4 6. mineral science.4. p = . 1 and 2 for graphic displays of the Group and Sex differences.4 6.3 2.

2. AQ scores in male and female controls (Group 2).The Autism-Spectrum Quotient 9 Fig. . AQ scores in AS/HFA group and controls (Groups 1 and 2). 1. Fig.

languages.8 4.0 16.62).8 1. excluding the nonspecific category).2 2. theology. were compared in one ANOVA.6 2.7 5.2 2. physical and computer sciences.9 2. 5 6 There was no significant interaction of Sex × Degree. versus the Social Sciences (i. geography. in Group 3.1 6.9 3.6 2. The six types of Sciences (i.4. archaeology and anthropology.7 1. medicine.6 2.8 1.0 4. Skinner.5 2.7 5.8. 331) = 27. There was a main effect of type of Science.1 2. p < .0 2.8 1. F(1.2 2.9 4.2 5.0 2.8 17. Mean AQ and Subscale Scores for Students (Group 3) Studying Different Degrees n Communication Social Imagination Local details Attention switching Total AQ Sciences All M SD Male M SD Female M SD Humanities All M SD Male M SD Female M SD Social Sciences All M SD Male M SD Female M SD 454 3.2 2. or between male and female scientists.8 5.4 2. This confirms an earlier report of an association between science/maths skills. by Sex.8 2. p = .4 2.0001). ..0 3.1 2.1 2. which could have been any of the sciences.9 1..1 2.3 6.1 2. p < .5 1. p = .1 15.73). F(1.8 2.6 1.4 5.10 Baron-Cohen.8 2.7 2.9 2. F(5.1 6.1 5.9 1. Student Neuman Keuls tests showed that mathematicians scored higher than engineers.9 5.9 1. The mean AQ’s and subscale scores (and SDs for each type of science are shown in Table III. or music). There was also a main effect of Sex. F(1. but no main effect of Sex.4 1.8 119 157 110 2. replicating the sex difference found in Group 2 above. or management).9 2.2 5.0.5 1. who did not differ from each other.3.0 5.5 and nonspecific science6). There was a main effect of Degree. law.3 5. economics. and Clubley Table II.9.. and no Sex × Science-type interaction.8 1. land economy.0 3.32. Subscale differences were not compared Medicine included both medicine and veterinary science.5 2.7 5. 1998).5 51 59 science. This last category included those subjects who simply listed their Degree as natural sciences. who scored higher than medicine and biology.5 6.8. 834) = 5.8 19..0 2.0001. classics.3 1. p = . engineering.0 2.001.003. 834) = 11.8 5. F(2. t tests. p = .e. Wheelwright. scientists scoring higher than both humanities and social scientists. social and political sciences. versus Humanities (i.5 284 170 276 2. and autistic conditions (Baron-Cohen et al. p = .1 4. architecture. Analyzed differently.8 1.2 2.7 2.8 15.2 4.9 17.9 2.1 2.0 1. F(5.0 4. F(2.8 2.8 1. 450) = 16.5 4.7 5.4 1.e.0001. English.7 1.9 2. Martin.3 2. 834) = 0. p = .e.7 2.0 4.0 18.2 2. Differences Between Types of Science We then tested if there were differences between the different types of scientists. mathematicians scored significantly higher than the nonmathematician scientists.3 1.08. 331) = 3. philosophy.8 1.1 2.2 2.5 2. Scientists differed from nonscientists on two subscales (social and imagination.1 1.8 17.0 2. history.2 2.1 5.0 2. males scoring higher than females over all.3 2.7 2.0 16.3 4. 331) = 7.1 1.

9 4. 23. more conservative than would be estimated by another judge. Determining a Useful Cutoff Percentage of each group scoring at or above each AQ score is shown in Table V.8 points (SD = −0.9 5.6 3. Twenty-two of these families agreed to do this. Imagination = .77. p = . p = .6 2.5 1.6 3. Item Analysis and Internal Consistency An item analysis (percentage of each group scoring on each item) is shown in Table IV.4 2.002).6 7. t(16) = 0.6 1.5 1.0 2. To retest the finding from Group 3 that mathematicians score significantly higher than controls.2 5.6 3.9 2. strongly confirming the value of these items for discriminating HFA/AS versus controls established at the pilot phase.2 23 77 85 69 47 122 within the science types in order to avoid multiple statistical testing.63.8 18. 6.3.7 2.5 3.6 2. The internal consistency of items in each of the five domains was also calculated.4 2.0 2.7 1.The Autism-Spectrum Quotient Table III. p = . To test if self-report by adults with AS/HFA was leading to inflated scores.4 5. Group 4 scored significantly higher than the male humanities students.4 4.7 3. Social.1 7. There were no differences between Group 4 and mathematicians from Group 3.1 1.9 1.7 2. the final analysis compared Group 4 (Maths Olympiad) versus male humanities students from Group 3. Scores from the first and second AQs did not differ statistically.1 4.7. The parent version of the AQ omitted 10 items out of 50 (items 3. if anything.1 3.7 17.0 2.09.4 1.4 15.5 6. parents scoring their child more highly than their child’s self-report.65.8 2.8 3. and Self Versus Parent Report To establish test–retest reliability.0 1. 20.3 2.9 4. all subjects in Group 1 were asked if a parent could also compete an AQ on them.0 2. and Cronbach’s alpha coefficients were all moderate to high (Communication = . 12.4 2. On only 2 items out of 50 (items 29 and 30) did controls score more than adults with AS/HFA. and the same analysis for science versus nonscience students in Group 3.8 2.9 5. t(133) = −4.7 5.3 1.0 3.7. 17 students from Group 3 were asked to complete a second AQ 2 weeks after the first administration. and 42). t(99) = −1. a reanalysis differentiating these in terms of 1 versus 2 points led to the same pattern of results overall.0 1. 5.9 4. if anything.2 1.9 5.75) and were strongly correlated ( r = . This shows that scores in Group 1 are. Mean AQ and subscale scores from Group 4 are shown in Table I. Mean AQ and Subscale Scores for Student Scientists (Group 3) Studying Different Subjects n Biological sciences M SD Computer science M SD Engineering M SD Mathematics M SD Medicine M SD Physical sciences M SD Nonspecific science M SD Communication Social Imagination Local details Attention switching 11 Total AQ 31 2.2 5.8 1.3 3.0 19.4 3.5 3.6 2.2 2. The mean difference in AQ score between self-report and parent report for the 40-item AQ was 2.67). 36. 8. in .7 1.6).8 2.1 1.4 2.4 5.7 21.65.7 2.8 5. they served to reduce the size of group differences.5 1.4 1.3 1.5 7. since these could only be answered subjectively.1 2.42. 27.3 4. Attention Switching = . Regarding the decision to score “slightly agree” and “definitely agree” responses using 1 point only.6 21.2 2. Local Details = .8 1. p = .4 2. These two items were conservatively retained in the analysis since.1 14. Test–Retest Reliability.4 5.9 3.0 4.0 2.6 2.0001. = .

0 19.9 27.3 81.3 58. but that significantly more males than females in the general population show moderate levels of “autistic traits” (i.2 70. an experienced clinician (S. Note also that at AQ score 20+.5 81. Indeed.0 71.2 23.9 35.4 15.3 32.9 Baron-Cohen.6 40.3 77.6 32.9 38. .9% of control males do.0 16.5 25.9 52.3 36.9 25.4 31.0 10.0 54.5 20. of whom 11 agreed to be interviewed.0 77.2 12.2 61. and Clubley the AS/HFA group scored at this level. Validation of the AQ Among Controls We cannot determine the rate of false negatives in Group 2 (general population controls).3 58.6 29.2 31.0 69.2 55.7 40.1 43.0 27.4 23. Item 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Group 3 Students (n = 840) 29.9 72.6 58.2 96.6 60.5 19.0 79.8 10. 1994).3% of 7 Only DSM-IV criteria were applied.0 75.5 37.6 69. being bullied.8 65.0 36. This suggests that there is not only a sex difference on the AQ overall (as reflected in the male mean AQ being higher than the female mean).9 38.1 58.1 33. 7 met criteria for HFA or AS.7 36.7 35. an AQ score of 32+ was chosen.9 23.1 69. while only 2% of controls did so.e. as for individuals of this age it was not appropriate to also use instruments such as the ADI-R or the ADOS (Lord et al.3 65.1 34.3 74.) sought to establish the number of criteria each subject met.6 67. To validate the AQ in Group 3. Skinner.9 22.4 50.8 74.2 55.0 32. Using DSM-IV7 criteria for autistic disorder.0 6. In all 11 cases however.3% scoring at this point or above) versus control females (1% of whom score at this point or above).1 60.4 19. Of the 11 subjects scoring 32+.2 19.7 25.6 81.B-C.4 18.6 73.0 23.5 39.6 82.9 37. there are twice as many males (40%) as females (21%) in the control group scoring at this intermediate point on the scale.4 9.2 59.7 59. Of the other 4.9 Group 2 Controls (n = 174) 36.0 65. and was even valued. Considering Table V. all met at least three criteria.9 65..6 14.1 56.8 91.1 21.6 20. whereas 3.0 81. as the majority of these completed the AQ anonymously.9 81.9 43.0 48.0 87. and because none of those meeting criteria complained of any current unhappiness.0 19. and using the rule that a useful cutoff would discriminate the groups with as many true positives and as few false positives as possible.5 14.0 87. many of them reported that within a University setting their desire not to be sociable.7 17.5 60.6 54.5 28. Normal Sex Differences on the AQ Table V also shows that control females never score as high as 34+.0 60.0 11.5 81.8 35.9 27. Wheelwright.0 17.4 79.4 50.7 63.2 81.5 65.2 36.8 64.6 45. together with their desire to pursue their narrow or repetitive interests (typically mathematics and computing) was not considered odd.3 17. since 79.8 38.2 65.9 84.0 24.2 32.5 75. there was evidence from self-report of significant impairment in functioning during the school years (social isolation.0 86.1 13.0 47. Martin.8 40. No diagnoses were actually made for two reasons: No parent was present to provide independent developmental data.5 93.8 49.9 31.1 14.6 84.3 32..8 18. 32+ also seems to be a useful cutoff for distinguishing females with AS/HFA (92.6 81.1 12. we called in for clinical interview all subjects scoring 32+.0 81. those traits that people with AS or HFA tend to endorse on the AQ).0 65.0 25.3 38.4 29.0 41.1 20.6 13. Item Analysis for Groups 1–3 Group 1 AS/HFA (n = 58) 67.12 Table IV.3 38. The clinician remained blind to the AQ score of the subject being interviewed.6 74. and a sex difference at high levels on the AQ (reflected in the sex ratio in Group 1 being 3.9 58.5 26.0 Table VI.5 29.5:1). and difficulty in making friendships).8 25.0 84.

9 76.6 36.2 1.4 63.0 47.8 95.4 97.8 8.6 22.5 46.3 3.2 2.7 37.8 33.9 94.5 85.6 75.8 94.7 86.0 1.7 90.0 4.7 89.6 66.0 4.3 2.1 46.9 2.0 93.4 7.5 13.1 23.7 68.7 86.9 3.0 22.7 19.3 84.6 5.9 25.8 79. Percentage of Subjects in Groups 1 and 2 Scoring at or Above each AQ Score AS/HFA (n = 58) 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 98.6 77.4 50.0 71.1 2.3 96.1 53.6 93.2 27.4 14.2 64.0 95.2 80.7 0 0 AS/HFA males (n = 45) 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 97.9 76.2 10.4 18.1 4.0 2.9 3.1 40.9 73.7 13.3 0 0 0 0 0 0 0 0 0 0 0 0 0 Control females (n = 98) 100 100 100 100 100 99.9 61.2 38.6 6.5 60.4 14.4 96.4 18.0 2.8 75.3 4.8 81.2 0 0 AS/HFA females (n = 13) 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 92.1 15.7 89.3 93.8 46.8 29.5 70.6 94.7 0 0 0 0 0 Controls (n = 174) 100 100 100 100 100 98.0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 AQ Score 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 .6 76.7 1.8 87.3 86.6 96.5 82.3 1.0 1.3 93.2 2.8 62.4 97.4 62.6 0.3 5.2 6.0 13.3 1.4 2.0 77.9 51.3 24.8 69.4 19.8 5.6 0.8 94.2 6.7 82.6 96.2 85.6 96.2 57.1 81.3 70.5 61.2 55.9 91.5 23.2 22.The Autism-Spectrum Quotient 13 Table V.0 2.3 48.0 16.7 60.6 96.6 95.6 21.9 38.6 0 0 0 0 0 0 0 0 0 0 0 0 0 Control males (n = 76) 100 100 100 100 100 97.7 63.7 1.5 32.8 65.1 23.9 2.8 13.0 1.3 24.3 1.5 78.2 17.4 22.7 0.3 12.0 1.9 54.7 27.6 95.7 39.0 57.4 22.8 32.8 9.6 78.1 93.6 6.9 1.0 43.4 89.1 36.6 95.0 4.3 96.7 46.

7 0.3 0.4 54. physical scientists.4 0.5 0.6 73. and within the sciences.1 55.7 46. We wish to underline that the AQ is not diagnos- AQ Score 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Nonscience (n = 386) 100 100 100 99.2 15. communication. Skinner.4 16.0 90.8 19.6 8. Martin.6 1.7 4.5 51. 1999) and may have implications for the marked sex ratio in AS (Wing. 1998).3 67.6 32. Scahill. and engineering (Baron-Cohen et al. 1981).3 4. 1944.0 97. attention to detail.2 0 0 0 0 0 0 Baron-Cohen. The group differences between the AS/HFA group versus controls are if anything conservative..6 39.2 99. Wheelwright. 80% scored above a critical minimum of 32+.4 32.1 70. 1997.9 62. Baron-Cohen & Hammer.7 0.1 7. scientists score higher than nonscientists.3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 . & Lawson.3 12.1 3.4 0.9 0.2 0. since the questionnaire purports to measure autistic spectrum traits. This is consistent with the extreme male brain theory of autism (Asperger. & Rutherford.14 Table VI.7 98. both overall.6 1. implying that IQ and SES do not appear to influence AQ.0 39. Our recent single-case studies of very high-achieving mathematicians. Baron-Cohen.2 99. The AQ can also be said to have reasonable construct validity.3 6.4 89.2 23.3 62. computer scientists. Percentage of Subjects in Group 3 Scoring or at Above Each AQ Score Science (n = 454) 100 100 100 100 100 99. and at intermediate and high levels of autistic traits. Wheelwright.5 75. AQ scores from a general population sample and a Cambridge student sample were not significantly different. As predicted.4 96.7 5. Future work needs to test the false negative rate by carrying out diagnostic assessments on a larger number of subjects in the control group than was possible here.1 2.1 0.6 29.3 10.3 2.4 1. Converging evidence for a link between AS and talent in physics has been reported in an unselected sample of children with AS (Baron-Cohen. This demonstrates that the AQ has reasonable face validity. physicists. in press). and computer scientists with AS show that this condition need not be any obstacle to achieving the highest levels in these fields. Finally. in that items purporting to measure each of the five domains of interest (social. and Clubley DISCUSSION We have described a new self-assessment screening instrument.2 97. and people with a diagnosis involving these traits score highly on it.9 96.4 85. The AQ has excellent test–retest reliability.6 30. men score slightly but significantly higher than women.4 85. This latter finding replicates our earlier studies finding a link between autism spectrum conditions and occupations/skills in maths.4 0.7 99.5 81.5 1. the Autism-Spectrum Quotient (AQ).4 94.3 0. Within the control group. Spong. Stone. adults with AS/HFA scored significantly higher on the AQ than matched controls.2 81. whereas only 2% of controls did so. Wheelwright.3 96. and engineers score higher than the more human or life-centered sciences of medicine (including veterinary science) and biology.7 10. mathematics.9 93.3 46.6 19.2 26. and attention switching) show moderate to high alpha coefficients.0 23.4 8.4 12. imagination. given that parents score their child with AS/HFA higher than they score by selfreport. physics. for measuring the degree to which an individual of normal intelligence shows autistic traits.6 3.

I tend to have very strong interests. 21. 80% of those scoring 32+ met DSM-IV criteria for HFA. I usually notice car number plates or similar strings of information. Future work could include administering the AQ to other psychiatric control groups. I find making up stories easy. In a social group. I can easily keep track of several different people’s conversations. I often notice small sounds when others do not. and to replicate the current results from Group 1 with patients diagnosed using standardized instruments. 10. 2. 15. 14. I prefer to do things with others rather than on my own. I find myself drawn more strongly to people than to things. I would rather go to a library than a party. 4. 13. 6. but did not merit a diagnosis as they were not suffering any significant distress. Such a high score on the AQ however does not mean an individual has AS or HFA. It does not upset me if my daily routine is distubed. I would rather go to the theatre than a museum. We suggest the AQ fills a gap for a brief assessment instrument for HFA/AS in adults of normal intelligence. 18.The Autism-Spectrum Quotient tic. 12. I frequently find that I don’t know how to keep a conversation going. If an adult 15 scores above 32 on the AQ. 25. even though I think it is polite. I can easily imagine what the characters might look like. 22. 26. I am fascinated by numbers. A score of 32+ appears to be a useful cutoff for distinguishing individuals who have clinically significant levels of autistic traits. since a diagnosis is only merited if the individual is suffering a clinical level of distress as a result of their autistic traits. but may serve as a useful instrument in identifying the extent of autistic traits shown by an adult of normal intelligence. it isn’t always easy for others to get a word in edgeways. 11. I am fascinated by dates. 17. definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely definitely definitely definitely definitely agree agree agree agree agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly slightly slightly slightly slightly agree agree agree agree agree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly slightly slightly slightly slightly disagree disagree disagree disagree disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely definitely definitely definitely definitely disagree disagree disagree disagree disagree definitely agree definitely agree definitely agree definitely agree definitely agree definitely definitely definitely definitely definitely agree agree agree agree agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly slightly slightly slightly slightly agree agree agree agree agree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly slightly slightly slightly slightly disagree disagree disagree disagree disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely definitely definitely definitely definitely disagree disagree disagree disagree disagree definitely agree slightly agree slightly disagree definitely disagree . 5. and is suffering some distress. Other people frequently tell me that what I’ve said is impolite. 20. As shown in the subsample of students in Group 3 above. When I’m reading a story. I find it difficult to work out the characters’ intentions. APPENDIX The Autistic-Spectrum Quotient 1. I prefer to do things the same way over and over again. When I’m reading a story. 16. 23. which I get upset about if I can’t pursue. I frequently get so strongly absorbed in one thing that I lose sight of other things. I enjoy social chit-chat. 24. I don’t particularly enjoy reading fiction. A limitation of this instrument is that it may not be appropriate for patients with low IQ. I find it hard to make new friends. 9. since the AQ assumes reading comprehension skills. in order to further determine its specificity. I notice patterns in things all the time. 8. I find it very easy to create a picture in my mind. we suggest this merits a referral to an expert clinician for a full diagnostic assessment. I tend to notice details that others do not. If I try to imagine something. 3. 19. When I talk. I find social situations easy. 7.

Diagnostic and statistical manual of mental disorders (4th ed. (1944).. 40. Mindblindness: An essay on autism and theory of mind. P. 11. Baron-Cohen.g. and to Hugh Osborne and Imre Leader for access to Group 4. 36. 28. When I was young. I find it easy to work out what someone is thinking or feeling just by looking at their face. S. I enjoy doing things spontaneously. S. Is autism an extreme form of the male brain? Advances in Infancy Research. A.).. Wheelwright. New situations make me anxious. Archiv fur Psychiatrie und Nervenkrankheiten. in Experimental Psychology. © MRC-SBC/SJW Feb 1998 Baron-Cohen. Boston: MIT Press/Bradford Books. I like to plan any activities I participate in carefully. S.. R. 38. 813–822. S. Stone. I’m not sure when it’s my turn to speak. 193–217. Short. 43.. to Nick Stone for data entry. P. J. types of car. & Scahill. Another advanced test of theory of mind: evidence from very high functioning adults with autism or Asperger Syndrome.. carried out Study 2 in part fulfilment of the B. M..).. 50. and Clubley definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree definitely agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly agree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree slightly disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely disagree definitely agree definitely definitely definitely definitely definitely definitely definitely agree agree agree agree agree agree agree slightly agree slightly slightly slightly slightly slightly slightly slightly agree agree agree agree agree agree agree slightly disagree slightly slightly slightly slightly slightly slightly slightly disagree disagree disagree disagree disagree disagree disagree definitely disagree definitely definitely definitely definitely definitely definitely definitely disagree disagree disagree disagree disagree disagree disagree definitely agree slightly agree slightly disagree definitely disagree ACKNOWLEDGMENTS The authors were supported by a grant from the Medical Research Council. Yuzda. 49. 31. Autism. In H. Mortimore. I enjoy meeting new people. 37. I don’t usually notice small changes in a situation. Bolton. I find it difficult to imagine what it would be like to be someone else. S... (1995). (1997). A. J. Neurodevelopmental disorders . & Robertson. 41.. Autism occurs more often in families of physicists. REFERENCES American Psychiatric Association. 76–136.). T. S. Wheelwright. 46. Skinner. E. V. I am not very good at remembering people’s date of birth. 48. Washington DC: Author. Gottesman. We thank the reviewers who improved this work through their feedback. & Hammer. & Rutter. Wheelwright. I like to collect information about categories of things (e.. Boston: MIT Press/Bradford Books. I am a good diplomat. I am not very good at remembering phone numbers.. Martin. 44. types of train. The extreme male brain theory of autism.16 27. University of Cambridge. Baron-Cohen. M. types of plant. or a person’s appearance. and E. I usually concentrate more on the whole picture. (1997). Asperger. types of bird. I find it difficult to work out people’s intentions. S. 25. (1995).. M. etc. 2. I can switch back to what I was doing very quickly. When I talk on the phone.. rather than the small details. 30. C. 117. Tager-Flusberg (Ed.. I am often the last to understand the point of a joke. 33. I. V. Autism as a strongly genetic disorder: evidence from a British twin study. I enjoy social occasions. Psychological Medicine. L. I find it very easy to play games with children that involve pretending. H. & Rutherford M. . (1994). engineers. Thanks also to Jessica Hammer for initial discussions. Baron-Cohen.. 38. C. S. 39. I am good at social chit-chat. I know how to tell if someone listening to me is getting bored. 47. If there is an interruption. G.. (1998). S. Bolton.. I used to enjoy playing games involving pretending with other children. (1999). 296–301. Simonoff. Die “Autistischen Psychopathen” im Kindesalter. I find it easy to do more than one thing at once. and the Three Guineas Trust during the period of this work. Baron-Cohen. Mead. People often tell me that I keep going on and on about the same thing.. 45. Baron-Cohen. T. Bailey. 42. 63–77. I find it easy to “read between the lines” when someone is talking to me. Smith.. E. 29. Journal of Child Psychology and Psychiatry. 34.. Jolliffe.Sc. the McDonnell-Pew Foundation. Baron-Cohen. 35. and mathematicians. 32. Le Couteur.

.. Wing. A. & Rutter. & Rutter. S. (1986). (1995). Asperger syndrome: A clinical account. Aspects of autism: Biological research. The childhood autism rating scale. Validity and neuropsychological characterization of Asperger syndrome: Convergence with nonverbal learning disabilities syndrome. and a computer scientist with Asperger syndrome: Performance on folk psychology and folk physics tests.. Cicchetti. (1979). 11–29. (1981). International Review of Psychiatry.. S.. M. Wing.. (1958).). Spong. P. J. & McLennan.The Autism-Spectrum Quotient (1999). S. Are intuitive physics and intuitive psychology independent? A test with children with Asperger syndrome. Le Couteur. Volkmar. & Rutter. Gillberg. J. New York: Plenum Press. Rutter. Autism: familial aggregation and genetic implications. Severe impairments of social interaction and associated abnormalities in children: Epidemiology and classification. Folstein. (1978). London: Gaskell/Royal College of Psychiatrists. A. Wing (Ed. Clinical and neurobiological aspects of Asperger syndrome in six family studies. (in press). M. In L. Psychological Medicine. B.). Autism Diagnostic Interview: A standard investigator-based instrument. F. (1989). Wing. Journal of Child Psychology and Psychiatry. MD: Williams & Wilkins. (1977). M. Journal of Autism and Developmental Disorders. Journal of Autism and Developmental Disorders. In U. (1994). The measurement and appraisal of adult intelligence. Sparrow. V. 36. 19. (1991). Wheelwright.. (1988). 2. .. B. Cambridge.. A. 18. a physicist. L. & Le Couteur. Infantile Autism: A Genetic Study of 21 Twin Pairs. M. Los Angeles: Western Psychological Services.). & Rourke. 67–80. Folstein. World Health Organization. A mathematician.. Autism and Asperger’s syndrome. Wechsler.). P. Frith. 9. S... Rutter. (1990). Journal of Developmental and Learning Disorders. Hold- 17 grafer. International classification of diseases (10th ed. Diagnosis and definition. Baron-Cohen. R. Journal of Child Psychology and Psychiatry. P. “Sex differences in intelligence”... & Renner.. Bolton. M. Cambridge. D. 5. UK: Cambridge University Press. C.. Robertson. UK: Cambridge University Press. Lord. Autism: a reappraisal of concepts and treatment (pp. 18. 1–26). Neurocase. Rios. 1127–1140. D. 363–387. 297–321. In M. M. Lord. S. Journal of Autism and Developmental Disorders. Schopler. Journal of Autism and Developmental Disorders. C. Autism Diagnostic Interview-Revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. Reichler. 659–685. Baltimore. 3–30. L. & Lawson.. Geneva. Switzerland: Author. The autistic continuum. & Gould.. 11. M. Schopler (Eds. Rutter. 475–483. Frith (Ed. Scahill. 115–130.. Autism and Asperger syndrome. C.. J. Rutter & E.. (1988). (1991). 24. A. L. (1994). Klin. E. U. Genetic influences in autism..

Sign up to vote on this title
UsefulNot useful