Spatial intelligence: the ability behind the term, the tests that measure it, and what training changes

0 The short answer

Spatial intelligence usually means the ability to represent and transform spatial information, including mental rotation, visualization and changes of viewpoint. In the CHC framework, many of these tasks fall under visual processing, Gv. Adult batteries sample the ability with several tasks rather than one puzzle. Spatial training research finds average improvements and transfer to other spatial tasks, but that does not guarantee a gain in an individual's IQ or prove that spatial ability is the only trainable domain. Read a score with its reference group, interval, testing conditions and other indices. ACIS's technical manual also flags an edition change affecting interpretation of its Visual Spatial Index.

1 What Spatial Intelligence Is: Visual Processing in the CHC Model

The phrase spatial intelligence names a real broad ability, but the measured ability is called visual processing, Gv, and it sits under the general factor rather than beside it as a separate intelligence. The phrase also appears in Howard Gardner's theory of multiple intelligences, which proposes several largely independent intelligences and which has its own chapter, by Jie-Qi Chen and Gardner, in the standard handbook of intellectual assessment, Contemporary Intellectual Assessment, fourth edition, published by Guilford in 2018. The chapter before it, by W. Joel Schneider and Kevin S. McGrew, sets out the Cattell-Horn-Carroll theory that testing actually uses, in which visual processing is one of about sixteen broad abilities correlated through a general factor. The page on types of intelligence explains why the evidence supports the second structure and not the first. The practical difference is a correlation. If spatial ability were an independent intelligence, a person's spatial score would say nothing about their verbal or reasoning scores. It says a good deal. It is often confused with fluid reasoning because both use figures, but Gf is defined by discovering a new rule from the item itself, as the page on fluid intelligence tests explains.

The empirical basis for Gv is John B. Carroll's 1993 survey, Human Cognitive Abilities, published by Cambridge University Press. Carroll reanalyzed more than 460 data sets from the factor analytic literature and proposed a three stratum theory: narrow abilities at the bottom, broad abilities above them and a general factor at the top. He devoted a chapter to abilities in the domain of visual perception, and the visualization and rotation factors described there contributed to the development of modern spatial assessment. Schneider and McGrew's 2018 chapter keeps the same broad ability, names the narrow abilities under it and gives the definitions that current test manuals use, which the next section tabulates.

The correlations are the evidence for the placement. In Project TALENT, the cohort of about 400,000 American ninth to twelfth graders described later on this page, the spatial composite correlated .61 with the mathematical composite and .59 with the verbal composite in the ninth grade cohort, while mathematical and verbal ability correlated .76 with each other, according to Wai, Lubinski and Benbow. Spatial ability was the most distinct of the three, and it was still far from independent. In ACIS the ability is reported as the Visual Spatial Index, VSI, one of the six primary indices described on the page on the six cognitive domains. The technical manual's historical confirmatory factor analysis, run on 2,750 complete records with Spatial Navigation in the battery, gives the VSI a loading of .906 on the higher order general factor, the second highest of the six after Fluid Reasoning at .922. The index correlates .813 with Fluid Reasoning, .795 with Quantitative Reasoning, .783 with Verbal Comprehension, .715 with Working Memory and .587 with Processing Speed. Those figures are the empirical form of the argument: the ability is distinct enough to deserve its own index, and general enough that a high spatial score usually travels with high scores elsewhere. The page on what an IQ test measures explains why a general factor summarizes shared variance across cognitive tasks, and the page on the g factor explains where a loading comes from.

2 The Narrow Abilities Under Gv, With the Task Formats That Measure Them

A spatial index is a sum of several narrow abilities, and knowing which one a task samples is the difference between reading a score and guessing at it. Schneider and McGrew's chapter lists the narrow abilities that factor analyses have repeatedly separated within visual processing, and the ACIS manual's CHC glossary maps its three visual spatial subtests onto the same names. The table below uses that vocabulary. For each narrow ability it gives the kind of task that measures it, described by format and never by content, because the items of any real test are protected. Linn and Petersen's 1985 meta-analysis in Child Development sorted spatial tests into three older families, spatial perception, mental rotation and spatial visualization, and that grouping still organizes the sex difference literature, so it appears in the prose that follows the table.

Read the table for one thing: the same word, spatial, covers tasks that make different demands. A speeded rotation item rewards a fast, accurate turn of a simple shape. A visualization item rewards holding many parts in mind while they change. An imagery item rewards moving oneself, in imagination, rather than the object. The ACIS manual says that agreement across its three subtests supports a broad visual spatial strength and that divergence between them can distinguish rotation, scanning, imagery and visual memory contributions, which is why the Visual Spatial Index is meant to be read subtest by subtest and not only as a total. The page on nonverbal IQ tests makes the related point that a matrix task and a rotation task both use figures and measure different things, and the page on the Matrix Reasoning subtest shows what the fluid reasoning version of a figural item asks for.

Linn and Petersen's three families overlap the table rather than matching it. Their mental rotation family is the speeded rotation row. Their spatial visualization family covers the visualization and flexibility of closure rows, the multi step tasks. Their spatial perception family, tasks that ask a person to judge a spatial relation against their own body or against gravity, has no single CHC row, and it is the family where the modern imagery and perspective taking measures come closest. Knowing which family a test belongs to matters because, as the section on sex differences shows, the size of the average difference between men and women depends on it.

3 Mental Rotation: The Experiment That Made Spatial Ability Measurable

Spatial ability became a laboratory quantity in 1971, when Shepard and Metzler showed that the mind turns an image at a measurable rate, and the tests built from their figures remain the most used spatial measures in research. Their paper in Science presented pairs of perspective drawings of three dimensional block figures and asked whether the two objects had the same shape. The time needed to answer rose as a linearly increasing function of the angular difference between the two orientations. It was no shorter when the difference corresponded to a rotation in the picture plane than when it corresponded to a rotation in depth. A person judging the pair was, in effect, turning one object until it lined up with the other, and the further it had to turn, the longer the judgment took. That result is why the term mental rotation exists, and why the cluster of tilted cubes at the top of this page is a spatial problem rather than a decoration.

Vandenberg and Kuse turned the figures into a group test in 1978. Their Mental Rotations Test, a paper and pencil instrument constructed from the Shepard and Metzler figures, showed an internal consistency of .88 by the Kuder Richardson formula and a test retest reliability of .83 in large samples. It correlated strongly with other tests of spatial visualization and showed virtually no association with tests of verbal ability, and it produced consistent sex differences over the entire range of ages investigated. Each item presents a target figure and a set of alternatives, some of which are the target rotated and some of which are mirror images, and the taker must find the rotated copies. That format, a target and its candidates under time pressure, is the template for most speeded rotation items in commercial and research tests since, including the eight option items of the ACIS Layer Rotation subtest described below.

Rotation is not the whole of spatial ability. Hegarty and Waller reported in Intelligence in 2004 two experiments that distinguished the ability to rotate a perceived object from the ability to imagine a change in one's own viewpoint, which they call perspective taking. The two abilities are highly correlated but separable. The separation did not depend on the method used to test people, and it generalized from perspective taking within a perceived small scale array to perspective taking within an imagined large scale environment. The perspective taking test used in that work asks the taker to imagine standing at one object in a drawn array, facing a second object, and to indicate the direction of a third, which is a change of the viewer and not of the object. ACIS samples this kind of viewpoint change in Spatial Comprehension, whose directional items use a compass convention defined on a knowledge sheet before the subtest begins, and it samples object rotation in Visual Puzzles and Layer Rotation. The manual asks that the three be read together for that reason: a person can be quick to turn a shape and slow to turn themselves, or the reverse, and the index alone does not show which.

The reason the distinction matters outside the laboratory is that everyday spatial tasks split the same way. Reading a map held in a fixed orientation, packing a car, assembling furniture from a diagram and judging whether a sofa will fit through a doorway are object transformations. Finding a route back through an unfamiliar city, pointing toward a landmark you cannot see and describing a route to someone facing the other way are viewpoint transformations. Both are spatial, both load on the same broad ability, and a test that samples only one of them measures only part of what the word covers.

4 How Adult Batteries Measure It: The WAIS-5 and the ACIS Visual Spatial Index

A normed battery does not report spatial ability as a pass or a fail; it reports an index on a scale with a mean of 100 and a standard deviation of 15, built from two or three subtests, with a standard error that the manual publishes. The Wechsler Adult Intelligence Scale, Fifth Edition, published by Pearson in 2024 for ages 16 years 0 months through 90 years 11 months, reports five primary index scales, and its Visual Spatial Index is built from Block Design and Visual Puzzles, according to Pearson's overview brochure reviewed on September 24, 2026. The same brochure states that the normative samples were collected in 2023 and 2024, that the ten primary index subtests take about 60 minutes and the seven subtest Full Scale IQ about 45 minutes, and that an Expanded Visual Spatial Index is among the ancillary index scores. The fourth edition had reported a single Perceptual Reasoning Index; the fifth reports Visual Spatial and Fluid Reasoning separately, and the page on what the WAIS-5 is sets out the full structure while the page on the WAIS-IV against the WAIS-5 explains what the split changed for scores. Block Design asks the examinee to reproduce a pictured design with blocks, a visualization task with a motor component. Visual Puzzles asks which three pieces would form a completed puzzle, a visualization task without one.

ACIS reports its Visual Spatial Index from three subtests, each with its own page. Every figure in this paragraph is from the technical manual, version 1.4. Visual Puzzles shows a completed figure and asks for the three pieces, out of six, that would reconstruct it; pieces may be mentally rotated but not reflected, and exactly three must be chosen, so a partial answer cannot be submitted. It has 25 puzzles, with 20 seconds for each of the first seven and 45 seconds for each of the remaining eighteen, a reliability of .850 as McDonald's omega and a standard error of 1.16 scaled score points. Layer Rotation shows several transparent layers of a square grid, each carrying a tab on one edge, and asks which of eight figures results from turning every layer until its tab is at the top and then stacking them; all layers stay visible for the whole item. It has 30 items, a 45 second limit on each with no shared time bank, and a discontinue rule after three consecutive errors. It replaced Spatial Navigation on September 7, 2026, and the manual states that its norms, and the tables that depend on the Visual Spatial and Nonverbal indices, will be issued with the next edition. Spatial Comprehension presents 35 directional and geometric items in 30 minutes, answered by multiple choice or by a typed number, using a compass convention defined on a knowledge sheet available before the subtest begins, so that performance reflects visualization with the terms rather than familiarity with the vocabulary. Its reliability is .870 and its standard error 1.08 scaled score points.

The published index from the earlier battery has a composite reliability, McDonald's omega, of .9488, and a standard error of 3.39 IQ points, which is 15 multiplied by the square root of one minus .9488, our arithmetic on the manual's coefficient. The manual is explicit that these index values belong to its documented edition, which was modeled with Spatial Navigation in the battery, and that Layer Rotation has no published values yet. A reader interpreting a Visual Spatial Index today should read that notice first. The page on reliability and validity explains what a composite omega is and why a standard error follows from it. Two related composite definitions appear in the manual, with the same need to retain the battery-era qualification: an expanded visual spatial composite that adds Visual Sequence, with an omega of .9508 and a standard error of 3.33, and a Nonverbal Index that includes Visual Puzzles and Layer Rotation among eight subtests carrying minimal verbal content, with an omega of .9629 and a standard error of 2.89.

5 How Does a Selection Test Differ From a Broad Spatial Index?

A selection test and a broad spatial index can differ in purpose, task coverage and reference group, even when their items look similar. A selection score may be compared with an applicant or occupational group, a broader norm sample or a stated criterion, depending on the instrument. An index within a cognitive battery describes performance on its contributing tasks using the publisher's scale. The page on the standard deviation of 15 explains one common scale. Both kinds of instrument can have norms and reliability evidence; the relevant question is what the particular score supports. Their scores may be related without being directly interchangeable.

One documented example is a dental admission measure. The Dental Admission Test, run by the American Dental Association, includes a Perceptual Ability Test of 90 items in six subtests: apertures, in which the taker decides whether a three dimensional object can pass through an opening; view recognition, in which the taker visualizes how an object appears from other angles; angle discrimination; paper folding, in which a folded and punched sheet is mentally unfolded; cube counting, in which the exposed faces of stacked cubes are counted; and form development, in which a flat pattern is matched to the solid it folds into, according to the ADA's 2026 candidate guide, updated August 4, 2026 and reviewed on September 24, 2026. The section is allotted 60 minutes. Scores on every DAT section are reported on a scale from 200 to 600 in 10 point increments, equated across forms, and each dental school sets its own use of them. These tasks sample a range of perceptual and spatial demands, which is why the test is a fair example of the format. What it is not is a norm referenced measure of the ability in the general population: its scale describes dental applicants, and a 500 on it cannot be converted into a Visual Spatial Index.

Employer spatial tests vary in their documentation and intended use. Ask which abilities are sampled, which reference group supplies the percentile, whether relevant reliability and validity evidence is available, and how the employer uses the result. A percentile need not describe only that day's applicants. The pre-employment assessment guide explains the validation question, and the job performance review distinguishes predictive evidence from a guarantee about a candidate. A score should be interpreted under the actual vendor's documentation, not assumptions about all selection tests.

The practice question deserves its own sentence. Because spatial performance responds to practice, a fact the training section documents, any selection test on a fixed format is open to preparation, and a prepared candidate's score partly measures preparation. A normed index has the same exposure, which is why the ACIS manual limits retakes and does not treat repeated practice attempts as equivalent to first valid completions, and why the page on how to prepare for an IQ test distinguishes preparation that helps a person from preparation that contaminates a score.

6 What Spatial Ability Predicts: Project TALENT and the Study of Mathematically Precocious Youth

Spatial ability predicts who enters and advances in science, technology, engineering and mathematics, and it does so beyond what mathematical and verbal scores already predict, which is the finding that half a century of longitudinal data agree on. The largest dataset is Project TALENT. Wai, Lubinski and Benbow's 2009 paper in the Journal of Educational Psychology drew on a stratified random sample of American high schools, grades 9 through 12, with about 400,000 participants tested in 1960 and tracked for more than 11 years, and aligned those results with findings published before 1957, with Graduate Record Examination data from 2002 to 2005 and with the Study of Mathematically Precocious Youth. The spatial composite was built from four tests whose formats the paper illustrates: a three dimensional visualization test of 16 items in which a flat piece is folded into a solid, a two dimensional visualization test of 24 items in which a figure is matched to a rotated copy rather than a mirror image, a mechanical reasoning test of 20 items and an abstract reasoning test of 15 items in the matrix format. The composite is a mixture of the visualization and speeded rotation rows of the table above, with a fluid reasoning component, and the paper says so.

The results are stated as proportions. Among participants who later held a STEM doctorate, 45 percent had scored in stanine 9, the top 4 percent, on spatial ability more than 11 years earlier, and nearly 90 percent had scored in stanine 7 or above, which means fewer than 10 percent of STEM doctorate holders had been below the top quartile in spatial ability as adolescents. The share in the top 4 percent was about 30 percent among those whose highest STEM degree was a master's and about 25 percent among those whose highest was a bachelor's, so the importance of spatial ability rose with each further credential. The paper's second result is about who gets missed. Some programs for talented youth require scores in the top 1 percent on mathematical or verbal measures. Within the three composites, 70 percent of the top 1 percent in spatial ability had not reached the top 1 percent on either the mathematical or the verbal composite, and more than half were below the top 3 percent on both, so they would not have been invited to a modern talent search. That group nevertheless earned STEM and visual arts degrees and entered STEM and visual arts occupations well beyond base rate expectations. The page on IQ and academic achievement reviews how general ability predicts schooling; this is the specific case in which one domain predicts something the general score does not.

The second dataset is smaller and more select. Shea, Lubinski and Benbow's 2001 twenty year study, also in the Journal of Educational Psychology, followed 563 participants who had been identified at age 13 through the Study of Mathematically Precocious Youth as scoring in the top 0.5 percent for their age on the SAT and who were assessed on spatial ability at identification in the late 1970s. Educational and occupational outcomes were collected 5, 10 and 20 years later. Those who chose mathematics and science as their favorite courses, earned undergraduate and graduate degrees in STEM and were in STEM careers 20 years on had shown higher spatial ability at 13 than those who went into the humanities and other fields, and discriminant function analyses at all three points showed that spatial ability added a statistically significant amount of variance beyond SAT mathematical and verbal scores. Kell, Lubinski, Benbow and Steiger followed the same 563 people for more than 30 years and reported in Psychological Science in 2013 that the two SAT subtests jointly accounted for 10.8 percent of the variance in who held patents and who had published refereed work in three classes, and that adding spatial ability accounted for a further 7.6 percent, a statistically significant increase. Spatial ability, in a sample already selected for high mathematical reasoning, still identified the people who went on to produce technical and creative work.

Three cautions keep those findings in proportion. They are group level associations: they say that people who become engineers had, on average, higher spatial scores, not that a given spatial score makes anyone an engineer. The SMPY findings come from the top 0.5 percent of mathematical reasoning and describe incremental validity within that range, which is a stronger claim than a simple correlation and a narrower population than the general one. And the abilities were measured in adolescence, when spatial scores predict a future that has not happened; an adult reading a Visual Spatial Index is reading a description of the present, and the page on whether IQ changes with age explains how stable a domain standing is across adulthood.

7 Spatial Reasoning Jobs and Careers: What the Record Supports and What Circulates

The occupations that the record ties to spatial ability are the ones whose training and admission systems already test for it, engineering, the physical sciences, dentistry, architecture and the visual arts, and the tie is probabilistic rather than a requirement. The evidence base is the previous section. STEM doctorate holders in Project TALENT were overwhelmingly drawn from the top quartile of adolescent spatial ability, and the spatially talented adolescents whom mathematical and verbal cutoffs missed entered STEM and the visual arts well above base rates. Wai, Lubinski and Benbow quote Richard Snow's 1999 observation that spatial ability relates to specialized achievement in architecture, dentistry, engineering and medicine and that admissions testing had barely used it. Dentistry is the profession that took the advice: the Perceptual Ability Test described above is a required section of the admission test for every applicant to a United States dental school. Sorby, Veurink and Streiner's 2018 engineering study, discussed in the training section, is the clearest record of what happens when an engineering program measures the ability at entry and teaches it to those who lack it.

What circulates is different. Career pages and quiz sites list spatial intelligence careers, often with a claimed minimum score attached, and none of the lists this page reviewed cited a norm sample, a test or a study for the numbers. The studies reviewed here do not establish a universal minimum Visual Spatial Index for these occupations. It contains averages and proportions: engineers and physical scientists score higher on spatial tests on average than people in the humanities, the distributions overlap, and a person below an occupation's average is not excluded from it by the average. The page on average IQ by profession shows how wide those overlaps are for general ability, and the page on IQ and success explains why group associations should not be treated as individual career assignments.

For a reader choosing a direction the useful sentence is the one about what is missed. The evidence says that a high spatial score with an ordinary verbal or mathematical score is a real and common profile, that it predicts success in fields that most selection systems do not screen for it, and that an adult who has never been tested on the ability may never have seen the strength on paper. A Visual Spatial Index does not assign a career. It shows a person whether the profile applies to them.

8 The Sex Difference in Mental Rotation, Stated Carefully

Men score higher than women on mental rotation tests on average, the difference is moderate, it is smaller or absent on other spatial tasks and it responds to practice, which is why it should be stated with its size and its limits rather than as a slogan. Linn and Petersen's meta-analysis in 1985 established the pattern: large sex differences were found only on measures of mental rotation, smaller differences on measures of spatial perception, and, when differences were found, they could be detected across the life span. Voyer, Voyer and Bryden's 1995 meta-analysis of 286 effect sizes, published in Psychological Bulletin, confirmed the pattern and put figures on it, with an average difference in favor of males of 0.56 standard deviations on mental rotation, 0.44 on spatial perception and 0.19 on spatial visualization, as reported in that analysis. They also found that the size of the difference varied from test to test within each family, partial support for the idea that the differences had decreased in recent years, and an age of first appearance that depended on the test used.

What those numbers mean is easy to misstate. A difference of 0.56 standard deviations means that a randomly chosen man would outscore a randomly chosen woman on a mental rotation test about 65 percent of the time, our illustrative arithmetic assuming normal distributions with equal variances; with no difference the figure would be 50 percent. A difference of 0.19 on spatial visualization gives about 55 percent. The distributions overlap far more than they separate, and a large share of women score above the male average on every one of these families. The differences are also specific to the task. They are largest on speeded rotation of three dimensional figures in the Shepard and Metzler format, the kind Vandenberg and Kuse's test uses, and they shrink on untimed and multi step visualization tasks, which is the family that most closely resembles the folding and assembly items of the tests in this page's tables.

The developmental record has been rewritten since 1995. Lauer, Yhang and Lourenco's 2019 meta-analysis in Psychological Bulletin examined 303 effect sizes for sex differences in mental rotation among 30,613 children and adolescents. A small male advantage first emerged during childhood and then increased with age, reaching a moderate effect size in adolescence. Procedural factors, the characteristics of the task and of the stimuli, accounted for variability in the reported differences even after age was controlled. A difference that grows through adolescence and depends on how the task is presented is consistent with a mixture of causes, and the meta-analysis does not settle which. Neither does this page.

Two findings bear directly on the practical question. Feng, Spence and Pratt reported in Psychological Science in 2007 that 10 hours of training on an action video game produced substantial gains in spatial attention and in mental rotation, that women benefited more than men, and that a control group playing a non action game showed no improvement. Sorby, Veurink and Streiner's engineering intervention, discussed next, improved the retention of women in engineering. A difference that moves with 10 hours of practice and narrows after a one credit course is evidence that current performance is not wholly fixed, and the evidence supports the possibility of improvement without guaranteeing an individual outcome. The page on left brain and right brain claims covers the popular explanation that the evidence does not support.

9 What Does Spatial Training Change?

Spatial training can improve performance on practiced and related spatial tasks, but its average effect is not a guaranteed gain for every person or every outcome. Uttal and colleagues' 2013 meta-analysis combined 217 studies. After excluding outliers, its average training effect relative to control was Hedges's g = 0.47, with standard error 0.04. The synthesis found evidence of durability across the observed delays and transfer to untrained spatial tasks. Its scope was spatial skill; it did not establish that this is the only trainable cognitive domain or that spatial practice raises a broad Full Scale IQ.

An effect size is expressed in the variability of the study outcome, not in an IQ test's published score units. Multiplying 0.47 by 15 produces approximately seven, but that arithmetic does not demonstrate a seven point gain on ACIS or on a Wechsler index. The outcome tasks, participants, interventions and comparison conditions differ. Converting the average to an individual before-and-after promise would discard exactly the information needed to know what improved.

Two intervention designs illustrate the range of evidence. Feng, Spence and Pratt's experiment examined action video game training and changes in spatial attention and mental rotation. Sorby, Veurink and Streiner studied spatial skills instruction in an engineering setting and reported benefits for selected educational outcomes. These studies concern particular tasks, instructional settings and participants. They are evidence that useful improvement is possible, rather than a universal prescription that any game or drawing course will produce the same result.

For someone choosing a practical goal, define the target skill before choosing the activity. Reading a technical drawing, mentally assembling a three dimensional object and planning a route overlap in spatial demands without being identical. Practice that helps one may provide limited information about another. Instruction should offer feedback on the target performance and opportunities to apply it to new examples. Improvement is more persuasive when it appears on a task that was not simply memorized during practice.

The distinction between near and farther transfer matters. Becoming faster at a familiar rotation format is useful if that format resembles a real requirement. It is a narrower achievement than improving an unrelated academic outcome or a broad index. The page on whether IQ can be improved compares different kinds of evidence, and the working memory testing guide explains why gains on a practiced task should not automatically be interpreted as gains in a broad ability.

A training study also needs an appropriate comparison. A group can improve because it receives instruction, because it becomes familiar with the testing procedure, or because its later session differs in motivation and conditions. A well designed comparison helps separate those explanations. Outside research, a person can track practical progress without claiming to have isolated the cause of every score change. If the goal is easier navigation, fewer route errors may be a more direct outcome than another IQ label.

An initial measurement can be helpful when it serves the goal, but a paid cognitive assessment is not required before someone learns a spatial skill. Likewise, repeatedly taking the same protected test is not a suitable training plan. Follow the instrument's retest guidance and keep assessment items separate from practice materials. The preparation guide distinguishes learning the instructions and arranging good conditions from rehearsing secure item content.

10 How Should a Visual Spatial Index Be Read?

Read a spatial score as an estimate from a particular instrument and edition, then examine its relation to the broader profile. The reference group tells you what comparison the number makes. The interval communicates its reported precision. The subtests show which performances contributed. The guide to IQ score interpretation explains why all three pieces belong with the number, and the score versus percentile guide explains the conversion of scale units into relative standing.

Consider a hypothetical spatial index of 118 with an interval of 111 to 125. On the mean 100, SD 15 normal model, 118 is about the 88th percentile, our arithmetic. The interval extends across a meaningful range of relative standings. It does not establish that the person has one exact ability level, and it should not be used to describe every spatial activity outside the test. This is an illustrative reading example, not a current ACIS report reconstructed from unpublished norms.

The distinction matters for the manual edition available with this article. Its published VSI reliability of .9488 and related factor estimates come from the edition whose battery included Spatial Navigation. The manual states that Layer Rotation replaced that subtest on September 7, 2026 and that the dependent tables will be issued with the next edition. A calculation using the older coefficient produces a historical SEM of about 3.39 points, but it is not independent validation of the replacement index. Current interpretation must preserve that qualification.

Now compare the index with other scores. A spatial index above a verbal or fluid reasoning index may suggest a relative strength, but the observed gap needs the instrument's appropriate comparison method. The simple square root of the sum of two squared SEMs assumes uncorrelated measurement errors. That assumption should not be silently imposed on every pair of scores. Statistical significance, the frequency of a gap in the norm sample, score validity and practical usefulness are different questions. A difference can be uncommon without establishing a diagnosis.

Read the subtests with the same care. Better performance on rotation than assembly may suggest a hypothesis about task demands. It does not prove a specific mechanism from a pair of scores. Per item time limits, scanning strategy, working memory demands and familiarity with the format can contribute. The ACIS manual offers interpretive guidance for Visual Puzzles, Layer Rotation and Spatial Comprehension; that guidance should be considered alongside the edition notice and the testing conditions.

An uneven profile does not automatically make Full Scale IQ meaningless. The broad composite summarizes level across its contributing tasks; the profile adds information about direction and relative emphasis. Whether a particular summary is useful depends on the assessment evidence and the question being asked. It is possible to retain a broad summary while describing a spatial strength that the summary alone would not communicate.

Finally, keep the practical problem in view. A student learning engineering drawing may benefit from instruction whether a spatial score falls just below or just above a chosen band. A job applicant must satisfy the employer's actual selection procedure. A person considering clinical concerns needs an assessment designed for that purpose. The index is evidence about performance on sampled tasks, not a universal answer to all three questions.

11 How Can Progress Be Checked Without Confusing Practice With Ability?

A useful progress check measures the skill you intended to improve and uses examples that were not simply rehearsed. Begin with a concrete outcome, such as interpreting a new drawing accurately or explaining a route from a different viewpoint. Record the conditions and errors that matter to that task. If speed is relevant, track it alongside accuracy, because a faster answer with more errors is not the same improvement as a faster correct solution.

The next step is to vary the examples. A learner who has memorized the orientation of a particular figure may appear to have acquired a general rotation skill when the gain is item specific. New figures with the same underlying demand provide a better check. More distant tasks, such as a practical assembly problem, ask whether the improvement generalizes beyond the exercise format. Both outcomes can be useful, but they answer different questions and should be named separately.

Progress should also be assessed over a reasonable interval. An immediate post-practice score captures a different question from performance retained after a delay. A person can keep a simple record of what was practiced, for how long, and what happened on later unfamiliar tasks. That record does not become a controlled experiment, but it supports a more careful conclusion than comparing two unrelated online quiz totals.

The same caution applies to personal explanations. A disappointing test sitting is not proof that someone cannot learn a spatial skill, and a large improvement is not proof that a broad cognitive trait rose by the same amount. Look for changes in the activities that matter, acknowledge the limits of the comparison, and consult the relevant professional when the assessment question involves a clinical or formal educational decision.

12 Where ACIS Sits, and What a Reader Should Do

We sell a paid online assessment that reports a Visual Spatial Index, so treat this section as a disclosure and check it against the rest of the page. ACIS is a self administered battery of 20 subtests across six domains, scored against an English speaking adult reference frame of 3,243 records aged 16 to 90, with a technical analysis set of 2,750 complete records used for the factor analysis and the reliability tables. Its Visual Spatial Index is built from Visual Puzzles, Layer Rotation and Spatial Comprehension, with the edition-specific measurement qualifications explained above, and every index is reported with its percentile and its 95 percent confidence interval. The prices, read on the home page on September 20, 2026 and subject to change, are one time: the Quick form of 6 subtests at 15 dollars and about 45 minutes, which does not include the visual spatial subtests; the Optimized form of 13 subtests at 30 dollars and about 110 minutes; the Full Scale form of 20 subtests at 50 dollars and about 175 minutes, which reports the Full Scale IQ and all six indices; and a Custom form that lets a buyer choose the exact subtests, priced per subtest on the order page, with any fully completed domain returning its index. Five subtests can be taken free without a card, a purchased form stays open for 30 days, and there is a 5 day quality guarantee.

The limits are the ones the page has applied to everyone else, plus one of its own. ACIS is unsupervised and online, so a score depends on integrity controls, retake limits and completion rules rather than on a proctor. Its English speaking adult reference frame and technical analysis set are described in the manual; their counts alone do not establish a sampling design. Its report is not a clinical diagnosis, is not signed by a psychologist, and is not accepted by Mensa for admission. The limit specific to this page is the notice in the technical manual: Layer Rotation replaced Spatial Navigation on September 7, 2026, its norms and the tables that depend on the Visual Spatial Index and the Nonverbal Index will be issued with the next edition, and the index reliability figures quoted here belong to the edition modeled with the retired subtest. A reader who buys the three visual spatial subtests today should read that notice on the manual before interpreting the index, and should read the subtest scores beside it.

What a reader should do follows from the question they are asking. If the question is whether a job or a school will accept a spatial score, the answer is the selection test that institution names, and this page's table shows why a normed index is not a substitute. If the question is whether spatial ability is a personal strength that a verbal or mathematical record has hidden, a suitably supported assessment can provide relevant evidence, and the free trial shows the format before any payment. If the question is how to improve, the training section describes relevant evidence; a paid assessment is not a prerequisite for learning a spatial skill. In every case the reading of the index does not change: it is a position in a described reference frame, a few points wide either side, best understood next to the other five indices and next to its own three subtests.

13 Sources Behind This Page

The sources below support the research findings and the task descriptions; ACIS coefficients retain the edition qualification explained in the body. ACIS reliability, standard error, loading, correlation, timing and item figures are from the technical manual, version 1.4, with its notice of September 7, 2026; the standard error of 3.39 for the Visual Spatial Index, every confidence interval, percentile, standard error of a difference, overlap percentage and the conversion of an effect size into points are our arithmetic on published figures and are labelled as such where they appear. Product structure, admission test rules and fees were read on the pages cited on September 20, 2026 and will change.

• Shepard R N and Metzler J. Mental rotation of three-dimensional objects. Science, 1971, volume 171, issue 3972, pages 701 to 703.
• Vandenberg S G and Kuse A R. Mental rotations, a group test of three-dimensional spatial visualization. Perceptual and Motor Skills, 1978, volume 47, issue 2, pages 599 to 604.
• Linn M C and Petersen A C. Emergence and characterization of sex differences in spatial ability: A meta-analysis. Child Development, 1985, volume 56, issue 6, pages 1479 to 1498.
• Carroll J B. Human Cognitive Abilities: A Survey of Factor-Analytic Studies. Cambridge University Press, 1993.
• Voyer D, Voyer S and Bryden M P. Magnitude of sex differences in spatial abilities: A meta-analysis and consideration of critical variables. Psychological Bulletin, 1995, volume 117, issue 2, pages 250 to 270.
• Shea D L, Lubinski D and Benbow C P. Importance of assessing spatial ability in intellectually talented young adolescents: A 20-year longitudinal study. Journal of Educational Psychology, 2001, volume 93, issue 3, pages 604 to 614.
• Hegarty M and Waller D. A dissociation between mental rotation and perspective-taking spatial abilities. Intelligence, 2004, volume 32, issue 2, pages 175 to 191.
• Feng J, Spence I and Pratt J. Playing an action video game reduces gender differences in spatial cognition. Psychological Science, 2007, volume 18, issue 10, pages 850 to 855.
• Wai J, Lubinski D and 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.
• Kell H J, Lubinski D, Benbow C P and 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 and 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 and Streiner S. Does spatial skills instruction improve STEM outcomes? The answer is 'yes'. Learning and Individual Differences, 2018, volume 67, pages 209 to 222.
• Lauer J E, Yhang E and Lourenco S F. The development of gender differences in spatial reasoning: A meta-analytic review. Psychological Bulletin, 2019, volume 145, issue 6, pages 537 to 565.
• Schneider W J and McGrew K S. The Cattell-Horn-Carroll theory of cognitive abilities. In Flanagan D P and McDonough E M, editors, Contemporary Intellectual Assessment: Theories, Tests, and Issues, fourth edition, chapter 3. Guilford Press, 2018.
• Pearson. Measure more with WAIS-5: overview brochure, with the primary index scales and their subtests, publication date, age range, completion times and norming years. pearsonassessments.com, reviewed September 24, 2026.
• American Dental Association. Dental Admission Test (DAT) 2026 Candidate Guide, updated August 4, 2026, with the Perceptual Ability Test specifications, timing, score scale and fees. ada.org, reviewed September 24, 2026.

The interpretive framework is the Standards for Educational and Psychological Testing, published jointly by AERA, APA and NCME in 2014, and APA guidance on testing and assessment. Score interpretation must retain its intended use, uncertainty and applicable edition.

14 Frequently Asked Questions