The IQ test for engineers: what spatial ability adds beyond math scores

0 The short answer

An engineer who wants a real IQ test should take a normed battery that reports quantitative, visual spatial and fluid reasoning scores separately, because the research ties engineering to those abilities and no occupational average describes any one engineer. A psychologist giving the WAIS-5 is the supervised route (Pearson lists 45 to 60 minutes of testing and a qualification level C instrument), and ACIS is one online option: the Full Scale form covers all six domains for 50 dollars in about 175 minutes, and the Custom Test returns the Quantitative Reasoning Index for 10 dollars or the Visual Spatial Index for 15 dollars (prices read October 29, 2026). The decisive limit is that an online, unsupervised test is not a clinical instrument, and a score from it describes performance on those tasks, not a profession.

1 Which IQ Test Can an Engineer Take, and What Does Each Option Return?

Five routes exist for an engineer who wants a cognitive score, they differ most in who supervises the session and who the score is compared with, and only some return a normed IQ at all. The table lists what each route measures, its reference group, whether it is supervised, its time, a cost where we could verify one, and what you receive. Prices and times were read on the pages named on October 29, 2026 and can change. None of the routes is built for engineers specifically: every normed battery measures a broad set of abilities for adults in general, and the engineering link comes from reading the profile, not from the test name.

Read the table as a map of what each route is for. The psychologist route is the one that can support a clinical or educational question. The Mensa route is a membership gate: American Mensa says on its testing page, read on October 29, 2026, that a qualifying score is one that is at or above 98 percent of the general population, and it lists a promotion that ends October 31, which we do not print here. The page on where to take an IQ test compares every route in more detail, and the page on the best online IQ tests compares the online batteries.

Pearson prices the instrument itself at 1,784 dollars for the WAIS-5 complete print kit on its product page, read October 29, 2026, with qualification level C, which marks it as a professional instrument and not something sold for use at home. That is why a psychologist route is priced as an evaluation. The University of Wisconsin school psychology training clinic fee page lists a psychoeducational evaluation at 600 dollars and says a sliding fee scale is available, but it does not say which ages it serves or which instruments it uses. We state that figure as one example of how clinic services are priced, not as a quote for any adult cognitive assessment, and we did not find a published private practice fee we could verify.

The employer row deserves a plain sentence. Timed numerical and abstract reasoning tests exist for selection processes, they are not IQ tests in the sense used here, and a result is reported to the employer. This page stays with the question of what an engineer who wants a personal score can take.

2 What Is the Average IQ of an Engineer?

No source we could open gives one verified average IQ for engineers as a whole, and the figures that do exist come from different samples, different tests and different scales, so they cannot be merged into one number. The honest answer is a set of dated findings, each with its sample. We report each with its source, its sample and its year, and none of them describes an individual.

The largest representative estimate we found is a ranking of 360 UK occupations. Wolfram (2023), in Intelligence, used Understanding Society, a UK household survey, with valid occupation records for 56,096 respondents across 367 unit groups. Respondents aged 16 or older took five short cognitive tests in 2011 to 2013: word recall, a number series subset from the Woodcock-Johnson III, everyday arithmetic problems, a serial subtraction task for working memory, and one minute of animal naming. A general factor was extracted, rank normalized and scaled to a mean of 100 and a standard deviation of 15. The paper reports the highest and lowest mean estimates as physical scientists at 114 and packers, bottlers, canners and fillers at 87, a gap of almost two standard deviations. Its figure of the 15 highest occupations includes physical scientists, chemical scientists, IT business analysts, architects and systems designers, and aircraft pilots and flight engineers. The paper's full ranking sits in an appendix that we did not open, so we report no engineering unit group figure from it. The same paper says roughly a quarter of the variation in cognitive ability occurs between occupations. Three quarters occurs within them.

A second source uses a different test and a different metric. Edwards and DeYoung (2026), in Intelligence and Cognitive Abilities, analyzed the Armed Services Vocational Aptitude Battery in two US cohorts, the National Longitudinal Survey of Youth 1979 with 11,914 participants and the 1997 cohort with 7,008. They extracted a general factor and three specific abilities that they named tech, speed and math-verbal, each scaled to a standard deviation of one, and scored occupations on each. In the 1979 cohort, engineers had a coefficient of 1.23 (standard error 0.08) on the general factor and 0.28 (standard error 0.12) on tech ability, and the paper says the tech result was not significant after correction for multiple comparisons. Engineers were among the occupations with significantly high math-verbal ability, and in the 1997 cohort engineers were one of only four groups with a significant deviation, again on math-verbal ability. If the 1.23 is read as standard deviations above the sample average, it is about 18 points on a scale with a standard deviation of 15. That conversion is our arithmetic, not the authors, and it describes the engineering occupation group in one cohort, tested in 1980.

The third source is an experiment built to compare two occupations. Usher and colleagues (2021), in the BMJ, invited aerospace engineers and neurosurgeons to take the Great British Intelligence Test, an online battery. Of 748 people who took part, 401 complete datasets were analyzed, 329 from aerospace engineers and 72 from neurosurgeons, and the authors say fewer than 20 percent of those invited completed the survey. Compared with the general population sample, the only two significant differences among six cognitive domains were found for neurosurgeons, not for aerospace engineers. The paper says the comparison group was mainly white, had completed secondary school and had a university degree, and that the normative data come from self selection rather than random sampling. A reasonable reading is that an online sample of aerospace engineers did not stand apart from an educated online sample on six domain scores. It is not a finding that engineers have average ability, and the authors do not claim it.

The fourth source is not an IQ at all but is often confused with one. The Educational Testing Service publishes the mean GRE scores of seniors and nonenrolled college graduates by intended graduate field. In its Interpreting Your GRE Scores: 2026 to 27, Table 2, based on people who tested between July 1, 2022 and June 30, 2025, engineering has a Quantitative Reasoning mean of 160, a Verbal Reasoning mean of 151 and an Analytical Writing mean of 3.4. Physical sciences is 162 on Quantitative, business and humanities and arts are each 160, and education and life sciences are each 150. A table of intended fields measures who chooses to apply to graduate study, and it mixes ability with schooling, so the page on average IQ by college major is the place for that ranking and its limits.

Putting the four together gives a modest conclusion. Engineering is one of the occupation groups with a high average on measures of general and mathematical ability, the sources differ in how high and on what scale, and none is a census of engineers. The page on average IQ by profession reports the estimates ACIS publishes, and every one of them is a group figure.

3 What Do Group Averages by Occupation Say About One Engineer?

A group average describes the group, and nothing in it tells you where one engineer sits, because most of the variation in ability lies within occupations and not between them. Wolfram (2023) says roughly a quarter of the variation in cognitive ability occurs between occupations, and for men and older participants it is close to one third. The remainder is the spread of people inside the same job title. An engineer at the group average and an engineer 15 points above it can hold the same title, work in the same office and have studied the same degree.

Two features of the published averages limit how far they can be pushed. The first is selection. People enter engineering through degree programs, admission tests, labor markets and their own choices, and those filters shape the group average. The page on IQ and job performance covers the selection research, and this page does not draw hiring conclusions from any group figure. The second is precision. Wolfram's figure shows the mean for each occupation with a 95 percent confidence interval, and for the top occupations those intervals overlap heavily, so the order of adjacent occupations is not stable. A ranking that places one occupation a point above another is a ranking of estimates, not of people.

The Usher study shows a third limit, the comparison group. When the reference sample is itself educated and self selected, a profession can look unremarkable against it. When the reference sample is a general population, a profession can look elevated against it. The same engineers produce both pictures. A score only has meaning relative to a reference group, which is why the page on how IQ scores are normed matters more than any occupational table.

The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) put the rule in Standard 1.1: no test permits interpretations that are valid for all purposes or in all situations, and each recommended interpretation for a given use requires validation. A group mean by occupation is a description of a sample. Using it to predict one engineer's score, or to judge whether a score is "enough" for the field, is an interpretation no occupational table has been validated for.

4 Which Abilities Does Engineering Draw On, in CHC Terms?

Engineering work draws most visibly on quantitative reasoning and knowledge, visual spatial processing and fluid reasoning, with working memory and processing speed in support, and the Cattell-Horn-Carroll model gives each its own name so that they are not blurred into one. The page on the CHC model explains the framework, and the six ACIS domains map onto it. For an engineer the useful point is that three of the abilities are easy to confuse with one another.

Quantitative knowledge (Gq) is acquired mathematical knowledge, number concepts and the use of quantitative procedures. Quantitative reasoning is a narrow ability inside fluid reasoning (Gf), the ability to reason with quantities and relations in a new problem. A strong calculus student can be strong on the first and ordinary on the second, and the reverse. The page on what a quantitative reasoning test measures sets out why one index can be read two ways.

Visual processing (Gv) covers the ability to generate, hold and transform images of shapes. Wai, Lubinski and Benbow (2009) cite Lohman's definition of spatial ability as the ability to generate, retain, retrieve and transform well structured visual images. In CHC terms the narrow abilities most often named for tasks like these are visualization (Vz) and speeded rotation (SR), which is how the Layer Rotation page describes them. The page on spatial intelligence covers how spatial ability is tested and trained.

Fluid reasoning (Gf) is the ability to solve novel problems that do not depend on learned content, including induction and general sequential reasoning. The page on fluid intelligence tests describes the task formats. Working memory (Gwm) holds and manipulates information over a short time, and processing speed (Gs) is how quickly simple tasks are done. Crystallized knowledge (Gc), the verbal and general knowledge a person has acquired, matters to engineers because standards, specifications and reports are written, but it is the least specific of the abilities here.

The third column is our description of where each ability plausibly shows up in engineering work, not a finding. The studies in the following sections test only some of these links, and they test them at group level. They do not say that an engineer with a lower score on one ability cannot do engineering.

5 What Does the Research on Spatial Ability and STEM Show?

Spatial ability is the ability most consistently linked to who goes on to advanced STEM credentials and occupations in the longitudinal data, and it adds prediction beyond mathematical and verbal scores. The strongest evidence is one large dataset analyzed in several ways. Wai, Lubinski and Benbow (2009), in the Journal of Educational Psychology, analyzed Project TALENT, a stratified random sample of US high schools in 1960 with roughly 400,000 students in grades 9 to 12, who were followed for 11 years. The authors aligned those findings with pre-1957 studies and with Graduate Record Examination data and the Study of Mathematically Precocious Youth.

Three of their findings matter for an engineer. The first is the gradient. Among those holding STEM doctorates, 45 percent had been in the ninth stanine, which they describe as the top 4 percent, on spatial ability 11 years earlier, and nearly 90 percent were in stanine 7 or above, so less than 10 percent were below the top quartile. About 30 percent of those with STEM terminal master's degrees and 25 percent of those with STEM terminal bachelor's degrees were in the ninth stanine. The authors conclude that the importance of spatial ability for STEM increases with more advanced credentials.

The second is incremental validity. When they entered mathematical and verbal ability first and then spatial ability into regressions predicting STEM versus non STEM groups, the increment for spatial ability was statistically significant and averaged .04, an additional 4 percent of criterion variance. That is a small effect in absolute terms, our reading of a .04 increment.

The third is the missed talent. Within the three composites in Project TALENT, 70 percent of the top 1 percent in spatial ability did not make the top 1 percent on either the math or the verbal composite, and over half of the top 1 percent in spatial ability were below the top 3 percent cut on both. The authors argue that talent searches based only on math and verbal scores miss these students. An earlier paper on the same sample, Humphreys, Lubinski and Yao (1993) in the Journal of Applied Psychology, examined engineering in particular. They found that spatial-math and verbal-math composites differentiated a variety of educational and occupational groups, and wrote that engineering and physical science disciplines appear to be losing many talented persons by restricting assessment to conventional mathematical and verbal abilities such as those of the SAT and the GRE.

Project TALENT also lets the authors compare degree fields on a composite of spatial, mathematical and verbal z scores, which they label a g level in an appendix note. They report that engineering doctorates averaged 1.73 and engineering bachelor's degrees 1.22 on it, close to math and computer science (1.75 and 1.18) and physical science (1.62 and 1.15) and above biological science (1.33 and 0.86). Those are averages of three z scores within the Project TALENT sample, so they are not on an IQ scale and we do not convert them to one.

Two later papers extend the picture. Lubinski (2010), in Personality and Individual Differences, reviewed more than 50 years of longitudinal research and argued that spatial ability is a neglected source of individual differences in complex learning and work, and that selecting students for advanced STEM learning without considering it might do harm. Kell, Lubinski, Benbow and Steiger (2013), in Psychological Science, followed 563 intellectually talented 13 year olds, identified by the SAT as in the top 0.5 percent of ability, for more than 30 years. The SAT mathematical and verbal subtests together accounted for 10.8 percent of the variance in patents and three classes of refereed publications, and adding spatial ability accounted for an additional 7.6 percent.

These are strong results, and they have limits that deserve the same plain statement. The samples are US students tested between 1960 and the late 1970s, the second is a highly selected talent search sample, and the outcomes are group differences in who earned degrees and patents. They show that spatial ability is informative about who enters and advances in STEM. They do not show that a particular engineer needs a particular spatial score to practice. The page on the mental rotation test covers one classic spatial task and how its free versions differ, and the sibling page on the Block Design test covers a construction format.

6 Can Spatial Skills Be Trained, and Does It Help Engineering Students?

Spatial skills improve with training, the improvement lasts and transfers to untrained spatial tasks, and one large engineering program found better course results for students who took a short spatial course. Uttal and colleagues (2013), in Psychological Bulletin, meta-analyzed 217 research studies. After removing outliers, the average effect size (Hedges's g) for training relative to control was 0.47 with a standard error of 0.04. Training effects were stable and not affected by delays between training and posttesting, and training transferred to other spatial tasks that were not directly trained. The authors conclude that spatially enriched education could pay substantial dividends in increasing participation in mathematics, science and engineering.

The engineering test case is Sorby, Veurink and Streiner (2018), in Learning and Individual Differences. A spatial test was given to more than 3,000 first year engineering students over five years. Students who scored below a pass or fail cutoff were assigned to a one credit spatial skills course, and those who passed formed the control group. A regression discontinuity analysis found a treatment effect on performance in a variety of introductory courses and on STEM grade point averages, particularly in Engineering Problem Solving and Analysis, and a positive effect on the retention of women in engineering. This design compares students just below and just above a cutoff, so it supports a causal reading for students near that cutoff and not for every engineering student.

One consequence follows for an engineer reading a spatial score. A spatial score is not a fixed trait to be accepted: the meta-analysis shows the ability responds to instruction, so a low score on a spatial index is a starting point and not a verdict. The page on reliability and validity covers why a single administration is a snapshot of performance on that day.

The practical reading is balanced. If an engineering student scores low on a visual spatial index, the research supports treating it as a skill with a known route to improvement, and the Sorby study is one example of such a route in engineering courses. If an engineer scores high, the research supports reading it as a strength relevant to the technical fields, not as proof of competence in any one of them.

7 How Should an Engineer Read a Quantitative Score?

A quantitative score has two possible meanings, learned mathematics and quantitative reasoning, and an engineer should check which one a test reports before reading the number. Engineers have usually studied mathematics for years, so a quantitative index can reflect that schooling as much as reasoning, and it can differ from a fluid reasoning index in the same person without any inconsistency.

The two ACIS quantitative subtests sit on the learned side and the working memory side of that line. The page on the Mathematical Achievement subtest describes tasks that emphasize learned quantitative knowledge and the application of school based mathematical content, and its instructions state 25 questions with a 35 minute total time limit, a reference table with formulas, and pen and scratch paper allowed but no calculators. The page on the Arithmetic subtest describes mental quantitative problem solving that is partly number knowledge and partly holding intermediate values in mind, and its instructions state that no calculators, notes, scratch paper or drawing tools are allowed and that the problems are delivered in audio. Both subtests are in the Full Scale form only, and both sit in the Quantitative Reasoning Index.

The ETS data in the earlier section show why this matters for interpretation. A Quantitative Reasoning mean of 160 for intended engineering graduate students on the GRE reflects years of coursework as much as reasoning ability, which is why the page on the math IQ test separates a math exam from a reasoning measure. An engineer who wants to know how their reasoning compares can look at fluid reasoning subtests, which do not rely on formulas, and at the profile across indices.

The page on mental math tests covers the mental calculation format as a sibling topic. Whether mental calculation matters for a given engineering job is a question the test cannot answer. The Arithmetic page describes the task as partly number knowledge and partly holding intermediate values in mind, so a score on it reflects both.

A last point is the ceiling. Engineers often ask whether a test can tell a high score from a very high one. That depends on the instrument's range and on the precision near the top, and the page on high range IQ testing covers the limits of tests near the top of the scale. The Standards, in Standard 2.14, say that where cut scores are used the standard errors of measurement should be reported in the vicinity of each cut score, which is the same point: precision is not constant across the scale.

8 What About Software, Electrical, Civil and Mechanical Engineers?

We found no study that reports a verified average IQ for software, electrical, civil or mechanical engineers separately, so the honest answer for each subfield is the general evidence plus what research on that field shows. The search phrases engineers type, such as "average IQ of a software engineer" or "IQ test for civil engineers," imply a subfield ranking that the sources we opened do not provide. Any page that states one number for a subfield without a sample, a test and a year is giving a guess.

For software, two small studies are worth knowing, and both point away from a single number. Bergersen and Gustafsson (2011), in the Journal of Individual Differences, studied 65 professional software developers from nine companies in eight European countries in a two day study. They found that the effect of working memory and experience on programming skill was mediated through programming knowledge, and that programming knowledge explained individual differences in programming skill to a large extent. The authors say low statistical power is a concern for some results. Piantadosi and colleagues (2023), in Empirical Software Engineering, tested 32 participants, mostly students, on bug fixing and feature tasks and measured attention and memory factors with psychometric tests. All the attention and memory factors had very low correlation with correctness and time, and years of programming experience was far more important than any of them. Together they suggest that for programming, accumulated knowledge and experience matter at least as much as the underlying abilities, and that neither study is a basis for an IQ expectation for a software engineer.

The Wolfram ranking does include a software adjacent group among its highest, IT business analysts, architects and systems designers, with the caveat that this is one unit group in one UK sample and the paper's intervals are wide. The Edwards and DeYoung paper reports that computer scientists had a high level of math-verbal ability. For a software engineer who wants a profile, the domains the evidence points to are working memory, quantitative reasoning and fluid reasoning. The ACIS Optimized form includes Digit Span and Alphanumeric Sequencing for working memory and Coding and Symbol Search for processing speed, and the speed index asks a different question from the reasoning indices.

For electrical, civil and mechanical engineering, the spatial evidence is the main field level link we could source: Project TALENT's engineering group, as a whole, scored high on the spatial, mathematical and verbal composite, and the Edwards paper finds engineers elevated on tech ability in the 1979 cohort, though that result did not survive correction. We do not break the group into subfields. An engineer in any of these subfields is better served by a profile than by a rumor about the field.

The siblings of this page cover neighboring audiences. The page on the IQ test for gamers discusses a different group with its own selection effects, and the page on the scientific IQ test covers the research side of the same STEM area.

9 What Does ACIS Measure for These Abilities, and What Does It Cost?

ACIS covers the abilities engineering draws on in its Full Scale form, and the Custom Test lets an engineer buy only the domain they want, at 5 dollars per subtest. The prices below were read on October 29, 2026 and can change. Every form is a one time payment with no subscription, with breaks allowed, progress saved between subtests and 30 days to complete.

Every form reports a Full Scale IQ with its 95 percent confidence interval, and the Custom Test returns "any domain you complete returns its full index score, and every subtest returns its scaled score." The Full Scale IQ needs the Full Scale battery, and at 10 subtests the Full Scale form costs the same. The 10 dollar and 15 dollar Custom Test prices are the 5 dollar subtest price times two and three.

Two of the visual spatial subtests are worth describing from their pages. Visual Puzzles asks whether separate pieces can be mentally combined into a finished whole without physically assembling them, a part-whole task. Layer Rotation asks the examinee to work out how each part has to be turned to reach a common orientation and imagine the parts combined, and the technical manual says it carries no words at all, so English proficiency matters less to it than to the verbal subtests. The manual also says verbal and knowledge tasks are affected by English proficiency and educational exposure, so an engineer who learned English later should read the verbal index with that in mind. The technical manual is the documentation to read for method, and this page quotes no statistic from it.

The free trial needs no card, and it allows one trial per person. Finished trial subtests are kept if the person later unlocks a form that includes them. The trial's own description says which subtests it covers.

10 How Should an Engineer Read the Score?

A score is an estimate with a margin, compared with a norm group, and for an engineer the profile across indices is usually more informative than the single total. The standard scale used by ACIS and by the major batteries has a mean of 100 and a standard deviation of 15, and scaled subtest scores run from 1 to 19 with a mean of 10 and a standard deviation of 3. The page on the standard deviation of 15 explains why that scale was chosen, and a percentile follows from the norm group's distribution.

Percentiles and scores are not interchangeable. The page on IQ score versus percentile explains why they cannot be averaged. A score converts to a percentile through the norm group's distribution, so the same raw performance can map to different percentiles in different norm groups, and an engineering student who compares a score from one test with a score from another has two different reference groups.

Two cautions apply to a high score. First, precision belongs to each reported score, and one index or one subtest should be read with more caution than the total. The Standards, in Standard 2.14, ask for conditional standard errors of measurement at several score levels, which is the formal version of that caution. Second, the confidence interval is part of the score. A total reported as 125 with an interval of plus or minus a few points means the person's true standing is plausibly anywhere in that range, so two engineers whose totals differ by two points have not been separated.

For engineers specifically, the profile matters. Differences between indices are normal: the page on index score differences explains that gaps between indices are common among people with typical profiles, and a gap between a quantitative index and a visual spatial index is not by itself a sign of a problem. The General Ability Index is a composite that sets working memory and processing speed aside, which is useful to read beside the Full Scale IQ when working memory and processing speed tell a different story from the reasoning indices.

The Edwards and DeYoung finding is relevant to this point. They report that specific abilities carry 30 to 57 percent of the importance of general intelligence for predicting education, income and occupational status, in their samples. That does not make a profile more important than the total in every case, but it is the reason this page recommends reading the indices and not only the Full Scale IQ.

11 What Do You Need for an Unsupervised Session?

An unsupervised session works best on a desktop or laptop in a quiet room with working sound and no interruptions, and an engineer should plan around the two features that differ from a classroom exam: audio items and the no calculator rule. The Standards say, in Standard 6.1, that test administrators should follow carefully the standardized procedures for administration and scoring specified by the test developer. In an online test the test taker is also the administrator, which makes following the stated conditions their responsibility.

The practical list is short:

• Sound. Digit Span, Alphanumeric Sequencing and Arithmetic present their items as audio and need working sound. Test the speakers or headphones before starting.
• No calculators on the quantitative subtests. The Arithmetic page states that no calculators, notes, scratch paper or drawing tools are allowed. The Mathematical Achievement page allows pen and scratch paper but no calculators, and provides a formula reference table.
• A quiet room and one sitting per subtest. Breaks are allowed between subtests and progress is saved, so a long form can be split across days within the 30 day window.
• Rested and unrushed. The scores describe performance in that session. A test taken after a night shift measures that night shift.
• No outside help. A second person solving items, a search engine or a notes window changes what the score means.

Engineers sometimes ask what separates a quiz from a normed index. A quiz returns a number with no stated reference group, and a normed index compares performance with a described norm group, which the page on the spatial IQ test explains for the visual spatial domain.

12 Where ACIS Sits, and What a Buyer Should Do

ACIS is one normed online option for an engineer who wants a profile across six domains, with real strengths and plain limits, and the decision about what to do with the score stays with the person. The strengths are concrete. The Full Scale form covers all six domains, including both quantitative subtests and Visual Puzzles, every form reports a Full Scale IQ with a 95 percent confidence interval and six indices with percentiles, adult norms cover ages 16 to 90, and it is a one time payment with no subscription (Quick 15 dollars, Optimized 30 dollars and Full Scale 50 dollars, read October 29, 2026). The refund policy, quoted literally, is: "Full refund within 5 days if the report does not deliver the features described at checkout or a technical issue prevents you from accessing your results." That is a conditional guarantee, not an unconditional money back one.

The limits are as concrete. ACIS is online and unsupervised, it is not a clinical or diagnostic instrument, it is English only, and it is not for school accommodations or admission to high IQ societies. ACIS is not for hiring. A score from it is not a credential for an engineering job, and no employer should be asked to accept one. The page on IQ tests for adults covers the wider category, and the page on the cognitive ability test describes the groups of tasks.

The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) are the yardstick this page tries to meet, not an endorsement of ACIS. Standard 1.1 says no test permits interpretations that are valid for all purposes or in all situations, Standard 1.4 says that if a score is interpreted for a use that has not been validated, it is incumbent on the user to justify that interpretation, and Standard 5.8 says norms should refer to clearly described populations that include the groups with which test users ordinarily wish to compare examinees. For an engineer this means one thing: use the score for the question it was built for, a profile of abilities compared with adult norms, and do not stretch it to say what a profession needs. The American Psychological Association is one of the three co-publishers of the Standards.

A buyer should do this, in order:

1. Decide the question. If you want a documented result for a clinical or educational purpose, a psychologist is the route. If you want a personal profile, an online normed battery is one option.
2. Pick the form that fits. Full Scale for all six domains, the Optimized form for a broader profile without the quantitative subtests, or the Custom Test at 10 dollars for the quantitative index or 15 dollars for the visual spatial index.
3. Start with the free trial. It needs no card and it shows what the format feels like before anything is spent.
4. Read the profile, not only the total. Keep the confidence interval in view and compare the indices.
5. Compare options. The supervised and online alternatives are in the first table on this page, and a psychologist remains the route for any clinical or educational question.

13 Sources Behind This Page

Every figure above is traceable to one of the following, and each is linked at the point where it is used. Prices, times and product facts were read on October 29, 2026 and can change. Any figure we derived ourselves is labeled in the text as our arithmetic.

• Wai J, Lubinski D, Benbow C P. Spatial ability for STEM domains: Aligning over 50 years of cumulative psychological knowledge solidifies its importance. Journal of Educational Psychology, 2009, volume 101, issue 4, pages 817 to 835.
• Humphreys L G, Lubinski D, Yao G. Utility of predicting group membership and the role of spatial visualization in becoming an engineer, physical scientist, or artist. Journal of Applied Psychology, 1993, volume 78, issue 2, pages 250 to 261.
• Lubinski D. Spatial ability and STEM: A sleeping giant for talent identification and development. Personality and Individual Differences, 2010, volume 49, issue 4, pages 344 to 351.
• Kell H J, Lubinski D, Benbow C P, Steiger J H. Creativity and technical innovation: Spatial ability's unique role. Psychological Science, 2013, volume 24, issue 9, pages 1831 to 1836.
• Uttal D H, Meadow N G, Tipton E, Hand L L, Alden A R, Warren C, Newcombe N S. The malleability of spatial skills: A meta-analysis of training studies. Psychological Bulletin, 2013, volume 139, issue 2, pages 352 to 402.
• Sorby S, Veurink N, Streiner S. Does spatial skills instruction improve STEM outcomes? The answer is "yes". Learning and Individual Differences, 2018, volume 67, pages 209 to 222.
• Wolfram T. (Not just) Intelligence stratifies the occupational hierarchy: Ranking 360 professions by IQ and non-cognitive traits. Intelligence, 2023, volume 98, article 101755.
• Edwards T, DeYoung C G. More than general intelligence: Cognitive abilities and class structure. Intelligence and Cognitive Abilities, 2026.
• Usher I, Hellyer P, Lee K S, Leech R, Hampshire A, Alamri A, Chari A. "It's not rocket science" and "It's not brain surgery": "It's a walk in the park": prospective comparative study. BMJ, 2021, volume 375, article e067883.
• Bergersen G R, Gustafsson J E. Programming skill, knowledge, and working memory among professional software developers from an investment theory perspective. Journal of Individual Differences, 2011, volume 32, issue 4, pages 201 to 209.
• Piantadosi V, Scalabrino S, Serebrenik A, Novielli N, Oliveto R. Do attention and memory explain the performance of software developers? Empirical Software Engineering, 2023, volume 28, issue 5, article 112.
• Educational Testing Service. Interpreting Your GRE Scores: 2026 to 27. Tables 1 and 2, test takers July 1, 2022 to June 30, 2025, read October 29, 2026.
• American Educational Research Association, American Psychological Association and National Council on Measurement in Education. Standards for Educational and Psychological Testing, 2014. Standards 1.1, 1.4, 2.14, 5.8 and 6.1 and their comments, read in the open access edition.
• Pearson Assessments. Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5) product page. Age range, completion times, qualification level and kit price, read October 29, 2026.
• American Mensa. Testing page. Regular testing prices and the qualifying percentile, read October 29, 2026.
• University of Wisconsin School Psychology Training Clinic. Fee schedule. Psychoeducational evaluation fee, read October 29, 2026.

14 Frequently Asked Questions