A number series test, explained: what it measures, where it appears and how to take a normed one

0 The short answer

A number series test asks you to infer the rule that generates an ordered list of numbers and supply the next or missing one, which makes it a test of rule induction, a fluid reasoning (Gf) ability, that also draws on number knowledge (Gq). Most pages titled "number sequence test" online are practice material for employer screens and exam preparation, not intelligence tests, so look for adult norms, a stated score scale and a confidence interval before you read a result as an IQ signal. ACIS offers one normed version, Visual Number Series, with 20 typed answer series of up to 45 seconds each, inside the Full Scale form (50 dollars, about 175 minutes, prices read October 28, 2026) or alone in the Custom Test for 5 dollars. The decisive limit is that one subtest is a noisier measure than a battery, and ACIS is online, unsupervised and not a clinical instrument.

1 What does a number series test ask you to do?

A number series item shows a short ordered list of numbers with one missing and asks for the number that completes it, so the whole task is to discover the rule that links the terms and apply it once more. Practice publishers describe the same layout. One preparation site says a series usually consists of four visible numbers plus one missing number, with the gap most often at the end and sometimes at the beginning or the middle (Psychometric Success, page read October 28, 2026). Another says number sequence questions usually show four to seven numbers with one missing, or two or three missing on harder items (JobTestPrep, page read October 28, 2026). The format is small, which is why it appears in so many batteries: a compact item can sample a person's ability to see structure.

A short original example shows the process. This series was written for this page, is deliberately easy and is not taken from any test: 2, 3, 5, 8, 12, then a blank. The gaps between neighbors are 1, 2, 3 and 4, so the next gap should be 5 and the missing number is 17. Notice what happened in your head. You did not look up a procedure. You compared neighbors, noticed that the comparison itself followed a pattern, and projected the pattern one step forward. Harder series stack more of those moves: a rule that alternates between two operations, two interleaved sequences that must be separated before either makes sense, or a rule applied to the differences of the differences.

Two features of the response format matter. In a multiple choice version the answer options constrain the problem: a person can test each option against the series, working backward from the answers. In a produced response version, the person must state the rule precisely enough to apply it, because nothing on screen suggests the answer. The ACIS technical manual makes this point about its own subtest, noting that because the answer is typed, the examinee cannot work backwards from options. Neither format is wrong. They simply place different loads on the person, and a score from one should not be read as if it came from the other.

The scientific literature has a name for the task family, series completion, and a model of what happens inside it, which a later section reviews. For now the practical point is that a number series is an induction problem dressed in numerals. Visual pattern problems such as matrices and figure sequences ask for the same kind of discovery with shapes, and the sibling page on pattern recognition tests covers that family and how it differs from number formats.

2 What do number series actually measure?

Number series measure inductive rule discovery applied to numbers, so they sit between fluid reasoning (Gf) and quantitative knowledge (Gq), and which side dominates depends on how demanding the arithmetic is. In the Cattell-Horn-Carroll framework, described on the page on the CHC model of intelligence, Gf is the ability to reason with novel material that cannot be solved by recall alone, and Gq is acquired mathematical knowledge and procedural skill. A series item draws on both. The examinee must detect a transformation among terms, which is the reasoning component, and must carry out the arithmetic that expresses it, which is the knowledge component.

Publishers have classified the task both ways. The Woodcock-Johnson V technical abstract from Riverside Insights (2025) assigns Number Series two broad abilities, Gq and Gf, and describes the task as seeing a series of numbers with one missing and determining the underlying rule to supply the missing number (Riverside Insights, 2025). The ACIS technical manual places its own Visual Number Series in Gf with Induction as the narrow ability, and describes the content as numeric while the intended construct is not simple arithmetic calculation. The ACIS subtest page makes the same point in plainer words: in CHC terms the task sits between Gf and Gq, because the examinee must discover relations, compare quantities and apply a rule to a novel structure.

The manual defines Induction as deriving rules from specific instances and detecting regularities and generating hypotheses about underlying structure. It defines Quantitative Reasoning as reasoning with quantities, relations and numerical constraints, and Gq as acquired mathematical knowledge and procedural skill. Read side by side, those definitions show why a series item is hard to classify cleanly. Seeing that each term is the previous one plus a growing amount is induction. Computing 12 plus 5 is not. A series whose rule requires an unusual operation shifts weight toward knowledge; a series whose arithmetic is trivial but whose structure is hidden shifts weight toward reasoning.

The same logic explains how number series differ from neighboring tasks. A matrix reasoning item asks for induction over shapes, so it needs no arithmetic at all. Figure Weights asks for reasoning about quantitative relations among balanced shapes and requires no number notation. A logic puzzle of the kind discussed on the page about Einstein's riddle asks for deduction under constraints, a stepwise operation, rather than for a rule to be inferred from examples. The broader question of what each of these abilities contributes is answered on the page about what IQ measures, and the page on fluid intelligence tests covers Gf as a whole, which a single number format cannot represent.

3 Where do number series appear in intelligence and ability batteries?

Number series appear in individually administered clinical batteries, in school group ability tests and in online batteries, and publishers do not agree on whether they belong in the core fluid reasoning measure. The clearest documented case is the Woodcock-Johnson family. Riverside Insights' technical abstract for the Woodcock-Johnson V (McGrew, Mather, LaForte and Wendling, 2025) says the system covers ages 4 through 90 and older. It lists Number Series among the cognitive tests of the Extended set, classified Gq and Gf, and it describes the new Matrices test as replacing Number Series as the primary measure of fluid reasoning in the general ability score (Riverside Insights, 2025). The same document says the WJ IV general intellectual ability cluster had used Number Series as its primary fluid reasoning measure, and that the WJ V fluid reasoning cluster is completely new. ACIS summarizes the clinical context of that battery on its page about the Woodcock-Johnson V.

The reason Riverside gives matters more than the reshuffling. According to the abstract, WJ IV user feedback and published and unpublished research suggested that including Number Series in the general ability score may have confounded the measurement of general intelligence for people with deficits in mathematics skills. The abstract adds that matrices tests are well supported measures of fluid reasoning that may be less influenced by cultural and linguistic factors than other types of Gf tests. That is a publisher's rationale, not a verdict from an independent study, and it is worded with "may". Still, it is a documented decision by the maker of a major battery to treat the number format as a useful but less clean measure of general ability, and a reader choosing a number series test should know about it.

Number series also appear in school ability testing. Riverside's CogAT test descriptions list Number Series as Test 6 in the Quantitative Battery, which also contains Number Analogies and Number Puzzles. For levels 10 through 17/18, the page says each question contains a series of numbers that follows a pattern and that students identify the rule the numbers follow; for lower levels it describes strings of beads that make a pattern (Riverside Insights, CogAT test descriptions, page read October 28, 2026). The level names run from 5/6 to 17/18, which signals a school age battery and not an adult one. The page does not name grades or norm groups, so no claim about either belongs here.

The Wechsler scales are the other major adult reference point, and the picture there is different. ACIS's own subtest comparison, on the page covering the WAIS-5 subtests, lists Visual Number Series among the 11 of 20 ACIS subtests that have no WAIS-5 namesake. That is a statement about subtest names, not a claim about every Wechsler task. It does mean a person who has a WAIS-5 report from a psychologist will not find a subtest of that name to compare against, whereas a person with a Woodcock-Johnson report may. In ACIS, the format is a subtest of the Full Scale form only, where it enters the Fluid Reasoning Index with Matrix Reasoning, Figure Weights, Logic Grid and Complex Relations, and two composites the technical manual describes: the Inductive Reasoning Index and the Complete Quantitative Reasoning Index.

4 What does research say about how people solve a number series?

Research describes series completion as four steps, detecting a relation, discovering periodicity, completing a pattern description and extrapolating, and the first step appears to draw on stored knowledge of numbers. The model goes back to the early work of Simon and Kotovsky on how people acquire concepts for sequential patterns (Simon and Kotovsky, 1963, Psychological Review, volume 70, issue 6, pages 534 to 546). LeFevre and Bisanz summarize it in their paper on number series. Detection of relations means examining the series and generating a hypothesis about the relation between adjacent elements, such as a constant addition. Discovery of periodicity means finding where a repeating unit starts and ends. Completion of the pattern description means producing a rule that accounts for every position within and between periods. Extrapolation means generating the numbers that fill the blanks.

LeFevre and Bisanz also describe what the earlier work implied about difficulty: a more complex pattern description puts more load on working memory. They report that Holzman, Pellegrino and Glaser found qualified evidence of a link between working memory capacity and number series performance in children in grades four and five, but no such evidence in adults, and that the type and the size of the arithmetic operation also contributed to item difficulty (Holzman, Pellegrino and Glaser, 1983, Journal of Educational Psychology, volume 75, issue 4, pages 603 to 618, as summarized by LeFevre and Bisanz). A careful reader should draw a modest conclusion from that summary: working memory plausibly matters for hard patterns, the adult evidence was weaker, and the arithmetic operations themselves shape difficulty.

The LeFevre and Bisanz study went one step further and asked what makes the first step succeed. They proposed three procedures for detecting relations among numbers: recognition of memorized numerical series, calculation and checking. In the first experiment people verified whether sequences formed rule based series. Memorized counting series, such as counting by fives, were verified more quickly than unfamiliar noncounting series, and partial counting information in false series helped people reject them. In the second experiment, people who scored well on a standardized number series test differed reliably from people who scored poorly in recognition of memorized sequences and in calculational efficiency (LeFevre and Bisanz, 1986, Memory and Cognition, volume 14, issue 4, pages 287 to 298). The authors concluded that this provides a basis for understanding how individual differences in knowledge influence performance on problems often used to assess inductive reasoning.

Those results do not say that number series are bad measures of reasoning. They say that reasoning here is intertwined with fluency in numbers, which is exactly what the Gq and Gf labels in the previous sections imply. They are also decades old, based on experiments with simple series and not on modern norm referenced batteries, so they explain the mechanism and not any particular test's accuracy. The ACIS technical manual draws a practical consequence for its own subtest: if Visual Number Series is low while the other fluid tasks are adequate, the issue may be symbolic sequence tracking, quantitative notation or difficulty inferring transformations from ordered symbols.

5 Why are most online number sequence tests really job screening practice?

Most pages that rank for number series searches are preparation material for employer or admission screens, and the pages I read present them as practice for numerical reasoning tests, not as intelligence tests with adult norms. A live search for the plain query "number series test" on October 28, 2026 returned a first page dominated by employer aptitude practice sites, exam and school test preparation pages and free quizzes, with only a few research or survey resources and one paper on convergence tests in mathematics. That describes one search on one day, not a count of the whole web, but it shows what a person looking for a number sequence IQ test is likely to meet first.

Two of those pages state their positioning plainly. JobTestPrep says number series questions are an integral part of almost all numerical reasoning tests, and its page never presents them as an IQ measure. Psychometric Success describes numerical reasoning questions that require only a basic level of education, notes that such tests are usually power tests in which candidates have enough time to finish, and gives around 15 to 20 questions in 20 to 30 minutes as the usual timing for numerical reasoning as a whole (both pages read October 28, 2026). Neither page states a price for its number series material, and neither describes an age referenced standard score or a confidence interval. They sell preparation for a test an employer will choose, which is a legitimate product with a different purpose from an ability measure.

The difference matters when a person reads the result. A practice test can tell you whether you can find rules quickly and whether your arithmetic holds up under time pressure, and the practice itself can improve both. It cannot tell you where you stand against adults of your age on reasoning ability, because that requires a reference sample, a score scale and an estimate of measurement error. The page on how IQ scores are normed explains what that reference sample does, and the page on the best online IQ tests compares the products that do report age referenced scores. If you want something free first, the page on a free IQ test describes what a no cost trial can and cannot show.

The sibling page on the math IQ test separates quantitative achievement from reasoning, which is the same distinction that number series straddle. Treat any result from a free quiz as a practice score, and treat a claim of the form "your IQ is" on a page with no stated norm group with the skepticism that the missing norm group deserves.

6 Which number series test should you take? The real options compared

The right option depends on whether you want a clinician's report, practice for an employer screen or an age referenced score you can obtain at home, and the table sets the four honest choices side by side with their limits. Every figure comes from a page opened on October 28, 2026, or from the publisher document named in the earlier sections. Where a page does not say, the cell says "not stated" instead of guessing.

Read the table row by row. The clinical route gives the richest interpretation, because a psychologist can combine a number series score with observation, history and other tests, but the publisher pages I read do not state a price or a timing, so a buyer must ask the clinic. The school route is not designed for an adult. The practice route is useful if an employer has named a test, and the page on where to take an IQ test covers the full range of in person and online routes if you are not sure which one you need. The ACIS route is the only row in which an adult can obtain a normed number series result without an appointment, and it is also the row with the most constraints you should know about, which the last section lists.

Three questions help a reader choose. First, who will read the result? A hiring manager reads a screen score, a clinician reads a battery and you alone read a home result. Second, does the format matter, or does the ability matter? If you only want to know how you reason, a single number format is the weaker answer, and the broader IQ test for adults page describes batteries that sample more abilities. Third, do you expect to score near the top of a scale? People who expect that can read the page on high range IQ testing before spending money on a short test, because a single subtest has less room to separate strong results from very strong ones.

7 What makes a number sequence test hard?

A number sequence test is hard when the rule is hidden behind several steps, when it needs more numbers held in mind than a person can hold, or when the arithmetic is unfamiliar, and a time limit then turns a hard item into a missed one. The research reviewed above supports each source of difficulty. LeFevre and Bisanz report that more complex pattern descriptions are associated with greater working memory demand, and that the kind of arithmetic operation and the size of the numbers in an operation contributed to item difficulty in the work they cite. They also found that familiar counting series, such as counts by a constant step, were verified faster than unfamiliar ones, which means that the same structure can feel easy or hard depending on what a person has stored.

The ACIS technical manual describes its own range in similar terms: series run from single step arithmetic progressions to alternating and compound patterns. A single step progression asks for one relation. An alternating pattern asks the person to notice that two relations take turns. A compound pattern asks the person to combine relations, for example a change whose size itself changes. As the number of required discoveries grows, so does the demand on working memory, and so does the chance of pursuing a wrong hypothesis for too long. The ACIS subtest page sets 45 seconds per item for 20 series, which means the subtest cannot take longer than 15 minutes of item time (our arithmetic: 20 times 45 seconds is 900 seconds).

A second kind of difficulty is spurious. A series can be made hard by requiring an obscure operation, such as one a person met only in a specialized course, and then the item measures what the person studied. A test builder who wants a reasoning score tries to keep the operations within general schooling and make the structure the challenge. That is a design intention, and a reader of a free quiz cannot check it, which is another reason the quiz that advertises itself as "hard" tells you less than a quiz that tells you its norm group. If you want to see the difference between knowledge and reasoning on quantitative material, the page on arithmetic as a subtest describes a task that is deliberately about learned numerical procedure presented aurally, so it contrasts instructively with a series.

Time also changes what is measured. A series solved in 20 seconds and a series solved in 70 seconds may reflect the same rule discovery with different fluency, and a time limit converts that difference into a score. Speeded simple tasks form a separate construct, processing speed, which a number series is not, even though it is timed.

8 Can you practice for a number series test, and what about answer keys?

Practice and instruction raise number series scores, so a result obtained after studying a test or after seeing its answers measures preparation as well as ability and cannot be read like a first sitting. An experiment published in the Journal of Intelligence compared a group that watched a number series video tutorial (58 people) with a control group that watched an irrelevant tutorial (52 people) before both worked on number series items. The tutorial group scored higher (d = .48), had more items with written notes (d = .41), and the number of items with notes correlated .66 with the sum score in the tutorial group and .75 in the control group. The effect of the tutorial on the sum score was statistically mediated by the number of items with notes (Schneider and Sparfeldt, 2021, Journal of Intelligence, volume 9, issue 4, article 55). The authors present note taking as a candidate mechanism, and the design supports the conclusion that a tutorial raised scores in that sample, not that notes are the only route.

That finding has two practical consequences. First, a method can be taught: look at the differences between neighbors, look at the ratios, test whether two interleaved sequences exist, check the differences of the differences, and work backward from a hypothesis to see whether it reproduces every visible term. A person who learns that checklist becomes faster, and a score after learning it is a score of a person who knows the checklist. Second, the ACIS subtest page states that no external aids are allowed, including calculators, notes, scratch paper and drawing tools. Given the correlation between written notes and scores in the experiment above, that rule matters: the same series can be easier on paper than in the head, so a score earned without notes should not be compared with a practice score earned with them.

Retesting is a separate effect. A meta-analysis of 174 samples from 122 studies, with a total of 153,185 participants, found significant retest effects across cognitive ability tests, with a standardized gain of 0.33 from the first to the second administration and no further gain after the third administration. Alternate forms produced smaller retest gains than identical forms, by 0.15 for the first to second comparison (Scharfen, Peters and Holling, 2018, Intelligence, volume 67, pages 44 to 66). That is a result across many test types, not number series alone, and it concerns average gains, not any one person.

What about "number series test with answers"? An answer key is useful for learning a method and harmful for measurement. If you work a practice set and then read the keyed solutions, you learn the rules the author used, which is training. If you read the key before you sit a test you intend to count, you contaminate your own score, and if the key circulates widely the test stops measuring anything. The responsibility to protect test content falls on test users and publishers in the professional standards discussed in the last section, and it is the reason ACIS describes its subtest without publishing items. Learn from explained practice sets that are labeled as practice, and keep your serious sitting clean. The page on reliability and validity explains why a score that has been inflated by prior exposure to the same items loses its validity as an ability estimate.

9 Is a low number series score weak reasoning or weak math?

A low number series score has at least three readings, weak rule induction, weak number fluency, or a poor fit between the format and the person, and only a profile of related scores can tell them apart. This is the same problem Riverside described when it reconsidered Number Series in its general ability score: the concern was people with deficits in mathematics skills, whose number series results may understate their general reasoning. A group statement like that is not a diagnosis of any reader. A low result on one subtest does not show a learning difficulty, and ACIS is not a diagnostic instrument; anyone who suspects a specific difficulty with numbers should ask a qualified professional for an evaluation.

The way to separate the readings is comparison. The ACIS technical manual gives two comparisons for its own subtest. If Visual Number Series is high while Arithmetic is lower, the profile may show induction that is stronger than quantitative manipulation under working memory load, because Arithmetic is delivered aurally and holds the terms in memory while the answer is computed. If Visual Number Series is low while other fluid tasks are adequate, the problem may lie in symbolic sequence tracking, quantitative notation or difficulty inferring transformations from ordered symbols. Both comparisons use scores within the same battery, which is the point: one subtest is read beside its closest neighbors.

A reader can apply the same logic outside ACIS. If a person is slow on number series but strong on a figural task such as matrix reasoning, the sensible hypothesis is that number notation, not rule discovery in general, is the obstacle. The sibling page on the logic IQ test describes tasks that test reasoning with constraints and no numerals, and the page on abstract reasoning tests places figural rule tasks beside general intelligence. Someone who wants to avoid arithmetic altogether will find the trade offs on the sibling page about an IQ test without math, and the page on quantitative reasoning tests explains why a quantitative score and a reasoning score can disagree.

None of this makes the number format a bad choice. A person who is strong on both numbers and figures gains little from either alone, and a person whose scores diverge learns something that a single composite hides. The caution is only that the label on a result, "fluid reasoning" or "quantitative", describes the intended construct, and the individual score reflects whatever mix of ability and knowledge that person brought to it.

10 How should you read a number series score?

Read a number series result as a scaled score with a margin of error, set beside the person's other fluid and quantitative scores, and never average or convert it by hand into an IQ. In ACIS, every subtest returns a scaled score from 1 to 19 with a mean of 10 and a standard deviation of 3, while indices and the Full Scale IQ use the familiar scale with a mean of 100 and a standard deviation of 15, reported with percentiles and a 95 percent confidence interval. The page on the 15 point standard deviation explains why that scale looks the way it does, and an IQ percentile calculator turns a standard score into the share of the reference group scoring at or below it. Scaled scores and percentiles are not interchangeable, which is why the page on IQ score versus percentile warns against averaging them.

Precision depends on the score. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014), Standard 2.3, says that for each total score, subscore or combination of scores that is to be interpreted, estimates of reliability and precision should be reported, and its comment adds that total scores can be highly consistent while subscores have unacceptably low reliability depending on how they are defined and used (AERA, APA and NCME, 2014). The practical translation is that a single subtest such as number series is a rougher estimate than the battery it belongs to, and that the margin of error on one scaled score is wider than the margin on a composite. The ACIS technical manual is the place to look for the instrument's own documentation, not the page you are reading.

ACIS also defines predefined composites that include this subtest. The technical manual says the Inductive Reasoning Index focuses on Induction through Visual Number Series, Matrix Reasoning and Complex Relations, and describes it as useful when the question concerns rule inference and abstraction across novel stimuli. It says the Complete Quantitative Reasoning Index combines Mathematical Achievement, Figure Weights, Arithmetic and Visual Number Series. The General Ability Index uses a larger reasoning and knowledge set that includes this subtest. According to the manual, a composite is reported only when its subtests are available, and creating a new composite after looking at the profile is a misuse the manual warns against.

Finally, interpret differences with care. Scores on related tasks differ in most people, and the size of a gap that deserves attention depends on the precision of both scores. The page on index score differences explains how common gaps are, and the General Ability Index is a composite that reduces the weight of memory and speed tasks. Treat a surprising number series result as a prompt to look at neighboring scores, not as a conclusion.

11 What should you prepare before an unsupervised session?

A number series subtest asks for typed whole numbers under a timer without external aids, so the practical preparation is a keyboard, a quiet room, no calculator and a plan for breaks across the longer forms. The ACIS subtest page states the instructions: enter only whole numbers, 45 seconds per item with auto advance when time expires, and no calculators, notes, scratch paper or drawing tools. That is a useful format for unsupervised use, because a typed numeral is unambiguous and cannot be guessed from options. Because the clock runs while you type, a comfortable keyboard and a stable connection protect the seconds that count.

Several conditions apply to the battery rather than to this subtest. Progress is saved between subtests, breaks are allowed and a person has 30 days to complete a form, so the Full Scale form of about 175 minutes does not have to be taken at one sitting. Some subtests present items as audio and need working sound: Digit Span, Alphanumeric Sequencing and Arithmetic do. The subtest page for number series does not mention audio, so do not assume either way. The instructions and the surrounding battery are in English, and the technical manual says verbal and knowledge tasks are affected by English proficiency and educational exposure, so a number format does not remove every language demand even though its items are numerals.

An unsupervised session also changes what a score means. Nobody confirms who took the test, whether notes were out of reach or whether the person was interrupted. That is a limitation of every online instrument, ACIS included, and it is a reason the page keeps repeating that results are for personal understanding and not for decisions made by others. Sit when you are alert, silence notifications and do not look at the series from a previous attempt if you are retaking; the retest research reviewed earlier suggests exposure alone can move scores. The other abilities a full session samples are useful context for deciding how much of the battery you want to take.

12 Where ACIS sits, and what a buyer should do

ACIS offers a normed adult number series subtest that can be bought alone or as part of a full battery, and it is an online, unsupervised, English only instrument that is not clinical, not diagnostic and not a substitute for a professional evaluation. The subtest is Visual Number Series: 20 series, a typed answer, 45 seconds per item. It is available in the Full Scale form, which costs 50 dollars and takes about 175 minutes, or inside the Custom Test at 5 dollars per subtest, where every subtest returns its scaled score and any domain you complete returns its full index score. The Quick form at 15 dollars and the Optimized form at 30 dollars do not include it. All prices were read on October 28, 2026 and can change. Payment is one time, with no subscription. The guarantee reads: "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."

There is a free trial with no card, limited to one per person: the site refuses a second trial from the same email address or the same browser. Trial results are kept if a person later unlocks a form that includes those subtests. The page on an adult normed cognitive ability test places this subtest in the wider battery.

Be clear about what ACIS is not. It is not for hiring, for school accommodations or for admission to high IQ societies, and its results are not a clinical or diagnostic finding. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) frame why those limits matter. Standard 1.1 asks that a test developer set out clearly how scores are intended to be interpreted and used and what construct the test is meant to assess; Standard 1.12 says that when an interpretation rests on premises about cognitive processes, theoretical or empirical evidence for those premises should be provided; Standard 6.10 asks that interpretations released to a test taker describe in plain language what the test covers, what scores represent, how precise they are and how they are meant to be used; and Standard 9.0 makes test users responsible for knowing the validity evidence behind an interpretation and for protecting the security of test content. The American Psychological Association's Guidelines for Psychological Assessment and Evaluation (approved March 2020) add that psychologists should select tools with sufficient validity evidence and score reliability (Guideline 6), treat test performance as one piece of an assessment drawn from multiple sources (Guideline 7), and attend to appropriate normative comparison and standardized administration (Guideline 8) (APA, 2020). An unsupervised online administration departs from standardized conditions by design, and the right response is to say so, which this page does.

The next step depends on the reader. If an employer has named a numerical reasoning test, buy practice material and learn the methods, keeping in mind that the result is not an IQ score. If you want a normed number series score and nothing else, a Custom Test with Visual Number Series costs 5 dollars. If you want a profile that places the score beside matrices, figure weights, logic and complex relations, the Full Scale form includes all five fluid subtests and the composites above, and a Custom Test with all five also returns the index, and the technical manual documents how scores are reported. If a clinical or educational question is driving the search, a psychologist is the right route.

13 Sources Behind This Page

Every figure on this page comes from one of the sources below, each opened or checked on October 28, 2026, and every ACIS fact comes from the published site or its technical manual, with prices subject to change. DOIs were checked against Crossref. Publisher and standards documents were read in full or searched for the passages cited. Vendor pages are cited only as market facts about practice material, never as evidence about intelligence.

• Simon, H. A., and Kotovsky, K. (1963). Human acquisition of concepts for sequential patterns. Psychological Review, 70(6), 534 to 546. https://doi.org/10.1037/h0043901 (The four component model is described from LeFevre and Bisanz, 1986.)
• Holzman, T. G., Pellegrino, J. W., and Glaser, R. (1983). Cognitive variables in series completion. Journal of Educational Psychology, 75(4), 603 to 618. https://doi.org/10.1037/0022-0663.75.4.603 (Findings are cited as summarized by LeFevre and Bisanz, 1986.)
• LeFevre, J., and Bisanz, J. (1986). A cognitive analysis of number-series problems: Sources of individual differences in performance. Memory and Cognition, 14(4), 287 to 298. https://doi.org/10.3758/BF03202506
• Schneider, B., and Sparfeldt, J. (2021). How to get better: Taking notes mediates the effect of a video tutorial on number series. Journal of Intelligence, 9(4), 55. https://doi.org/10.3390/jintelligence9040055
• Scharfen, J., Peters, J., and Holling, H. (2018). Retest effects in cognitive ability tests: A meta-analysis. Intelligence, 67, 44 to 66. https://doi.org/10.1016/j.intell.2018.01.003
• Riverside Insights (2025). Woodcock-Johnson V Technical Abstract (McGrew, Mather, LaForte, and Wendling, 2025). Technical abstract PDF, read October 28, 2026.
• Riverside Insights. CogAT test descriptions, Quantitative Battery, Number Series. CogAT test descriptions, page read October 28, 2026.
• American Educational Research Association, American Psychological Association, and National Council on Measurement in Education (2014). Standards for Educational and Psychological Testing. Washington, DC: American Educational Research Association. Open access edition, read October 28, 2026 (Standards 1.1, 1.12, 2.3, 6.10 and 9.0).
• American Psychological Association (2020). APA Guidelines for Psychological Assessment and Evaluation, approved by the APA Council of Representatives, March 2020. Guidelines PDF, read October 28, 2026 (Guidelines 6, 7 and 8).
• Psychometric Success. Numerical series tests FAQ. Numerical series tests, page read October 28, 2026 (market fact about practice material).
• JobTestPrep. Number series practice page. Number series, page read October 28, 2026 (market fact about practice material).

14 Frequently Asked Questions