The IQ test for couples: how two partners take a real test and compare the scores fairly

0 The short answer

A real IQ test for couples does not exist as a product: two partners each take the same individually scored, normed test in a separate session and then compare the results, and online "couples IQ tests" are usually quizzes about how well partners know each other. ACIS is one concrete option for the real comparison, with three one time prices per person, Quick at 15 dollars (about 45 minutes), Optimized at 30 dollars (about 110 minutes) and Full Scale at 50 dollars (about 175 minutes), read on October 28, 2026, and it is online, unsupervised, not clinical and in English only. The decisive step is reading the gap: with the partner correlation reported in the research and the usual normal curve assumptions, about one couple in three differs by 15 points or more, so a difference between two scores is often ordinary. Compare confidence intervals and index profiles, not a single headline number, and do not read either score as a verdict on the relationship.

1 What Is a Couples IQ Test, and What Do the Online Ones Measure?

No standardized IQ instrument is designed for couples, and most pages that use the phrase offer a relationship quiz that does not measure cognitive ability. An IQ score comes from one person working through a standardized set of tasks and being compared with a defined reference group. Nothing in that procedure involves a second person, so the useful form of the question is how two people can each take a real test and compare the results without fooling themselves. This page answers that question for adults. The pages on what an IQ test actually measures and IQ testing for adults cover the instrument itself.

A search on October 28, 2026 for "couples iq test free" returned mostly pages of the other kind. One example is a free quiz titled "Are We Intellectually Compatible" on Marriage.com, which we opened on that date. The page states that you do not need to measure your IQ or compare your educational degrees to find out whether you are intellectually compatible. It lists 15 questions about shared pastimes, cultural tastes, how partners talk about articles and politics, conflict resolution and shared values, and it does not explain how the answers are scored. That is a conversation prompt about compatibility, and the page itself says it is not an IQ measurement. It is not a criticism: a quiz about shared values can be a good evening. It is simply a different product from a test with a norm group.

Two other kinds of page turn up under the same search, and neither answers the question. Some pages offer a single free general IQ screener and suggest that two partners each take it. Others sell a test branded for couples whose items concern the relationship, not reasoning, memory or vocabulary. If a page promises a "couples IQ score" that combines two people into one number, ask what it is combining, because a pair does not have an IQ in the sense that an individual has one. The research on partners treats each person's score separately and then asks how the two scores relate across many couples.

Two neighbour pages own questions that this page leaves out. The page on the partner similarity evidence reports the findings, including the 0.44 and 0.23 correlations discussed later here. The page on index score differences covers how large a gap between two index scores inside one person's profile must be before it deserves attention. Here the focus is the gap between two people, which is a different statistical object, and the protocol that makes the comparison fair.

2 Which Tests Can Two Partners Take, and What Does Each Cost?

Two partners have six realistic routes, and only three of them return a score that both people can compare: a clinician giving a Wechsler scale, a normed online battery and a short screener of unknown quality. The table sets out what each route measures, who the score is compared with, whether a person supervises, how long it takes, what it costs for one person on the date read, and what the person receives. For two partners, multiply by two (our arithmetic).

The WAIS-5 row comes from the publisher's product page, which calls the scale an individually administered clinical instrument and lists ages 16:0 to 90:11. The clinic price comes from one university psychology clinic, whose page lists a gifted assessment based on the Wechsler scales at 100 to 400 dollars, computed on a sliding scale by income and by the type and complexity of the assessment, with half due at the first meeting. That is a single clinic, and its page does not say who administers the test or whether a written report is included. The full range of clinical prices is covered on the page about what an IQ test costs and in the sibling page on the cheapest IQ test.

The Mensa row matters most for couples, and it is the one most people overlook. The American Mensa testing page states that a qualifying score means you have tested at or above 98 percent of the general population, that the test takes one to two hours, and that it cannot provide a detailed report with scores, percentile ranks or your IQ score. A partner who passes and a partner who does not therefore hold a yes and a no, not two numbers. That can be the right route for membership, and it cannot support a comparison of two scores. The page about where to take an IQ test compares the routes in more detail.

For a comparison of two partners, the practical choice is a clinician for a documented evaluation or a normed online battery for a scored profile of each person at lower cost and with no scheduling. The rest of the page assumes the online route and uses ACIS as the worked example, because its three forms make the same-form rule easy to follow. The same logic applies to any other normed instrument you choose.

3 How Should Two Partners Take the Test? Separate Sessions, the Same Form

Each partner should take the test alone, in a separate session, on the same form and under similar conditions, because the comparison is only as clean as the least standardized session. A score is meaningful when the conditions resemble those under which the norms were collected. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014), which are open access on testingstandards.net, state in Standard 6.4 that the testing environment should furnish reasonable comfort with minimal distractions to avoid construct-irrelevant variance. In a home test, the second person in the room is one of the distractions.

The protocol below is our synthesis of that standard and of common sense, not a finding of any study.

• Same form. Both partners choose the same form. A comparison of a Quick score with a Full Scale score compares two Full Scale IQ values built from different subsets of subtests, and the six domain indices exist only where both took the domain.
• Separate sessions. Take the test on different days or at different times, alone, with the other partner out of the room. A partner who sits beside you, reads the screen or offers a hint turns an individual score into a joint one.
• Different devices and addresses. ACIS grants one free trial per person, and the site refuses a second trial from the same email address or the same browser. Each partner therefore starts with their own email address and their own browser or device. Each partner also needs a separate purchase if they unlock a form.
• No item talk until both are finished. Do not describe tasks to each other between sessions. The point is not secrecy for its own sake. A partner who knows what a task looks like starts with a practice advantage that the first partner did not have.
• Comparable states. Take it rested, not after an argument, a long shift or a late night. A score from a bad evening compared with a score from a good morning measures the evening and the morning.

The practice advantage is real in the research on repeated testing. A 2018 meta-analysis of retest effects in cognitive ability tests, Scharfen, Peters and Holling, pooled 174 samples from 122 studies and found significant retest effects, with a standardized gain of 0.33 for the first repetition and no further score gains after the third administration. Those are people who took the same or a parallel test again, not partners who heard about the tasks. That a partner who has been told what the tasks look like gains similarly is our reasoning, not a result of the study. It is a reason to hold the conversation until both sessions are complete. For the same reason, do not retake a form to close a gap.

The order of the two sessions is not something the evidence addresses. Agreeing to finish both in the same week and to keep the tasks private until then is enough.

When the two sessions are done, each partner should open their own report first, alone. A report belongs to the person who took the test, and nobody owes the other person a number. The Standards list test takers' rights to access their results and to be protected from unauthorized use of them as a cluster in chapter 8, which is a reasonable principle for two adults to apply to each other. Whether and when to share is a decision for the relationship, and nothing in the evidence on this page tells anyone how to make it.

4 Which ACIS Form Should Both Partners Choose?

Both partners should choose the same form, and the right one depends on how many domains you want in the comparison, because the three forms differ in length, price and the indices they can return. All prices below are one time, with no subscription, and were read on the published site on October 28, 2026; they can change. The two partner column is our arithmetic, two purchases at the single price.

Every form reports a Full Scale IQ with its 95 percent confidence interval. Seven subtests appear only in the Full Scale form: Information, Arithmetic, Mathematical Achievement, Visual Number Series, Complex Relations, Visual Puzzles and Visual Sequence. That is the reason a couple that wants the quantitative reasoning index, or a view of all six domains, needs Full Scale for both partners. A couple that wants a quick look at verbal, reasoning and memory can use Quick. Optimized adds the visual spatial and processing speed domains. The different lengths buy different amounts of detail, and the form you choose for both partners should match the detail you want.

There is also a Custom Test at 5 dollars per subtest, in which the buyer picks any subtests. In the product's own wording, "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. For two partners who care about one ability, such as verbal comprehension, Custom Test is a route to comparing one index without the rest. It does not return a Full Scale IQ.

Each partner can begin with the free trial. It needs no card, it is limited to one per person, and a trial started from the Full Scale form opens a set of subtests whose finished results are kept if the person later unlocks a form that includes them. The trial gives a feel for the format of the tasks before any money is spent. A partial trial is not a full battery, so it cannot be compared with a purchased result, and the page on a free IQ test says what the free route returns.

Breaks are allowed, progress is saved between subtests, and a person has 30 days to complete a form. That flexibility matters for couples with different schedules, since the two sessions can happen in different weeks. It also means the two people can finish on different days, which is fine, as long as each session is taken under similar conditions, as section 3 describes. To preview what a report contains before paying, the site publishes example reports for each form.

5 Compared With Whom? Why Both Partners Are Measured Against the Same Adults

The comparison between two partners is valid only because each score is placed against the same adult reference group, not against the other partner. An IQ score is not a count of right answers. It is a position in a distribution: the same raw performance is converted to a standard score with a mean of 100 and a standard deviation of 15, using a norm group of adults. The page on how IQ scores are normed explains the conversion, and the page on why the standard deviation is 15 explains what a point is worth.

ACIS reports Full Scale IQ and the six primary indices on that scale. It reports scaled subtest scores from 1 to 19, with a mean of 10 and a standard deviation of 3, and it reports percentiles and a 95 percent confidence interval. Its adult norms cover ages 16 to 90. For two partners, the useful consequence is that each is placed against adults, so the two scores are on a common scale even though the people may differ in age. What the report does not do is tell you how one partner compares with the other. That is the reader's step, and it uses ordinary subtraction on the standard scale.

Two habits keep that step honest. First, subtract IQ points and index points, not percentiles. A percentile is a rank, and ranks are not equally spaced: the same 10 points covers very different numbers of people at different places on the scale. The page on why you cannot average IQ and percentile shows the trap. Second, if you want to translate a score into rarity, use the calculators for percentile from a score and for how rare an IQ is, after the comparison, not before it.

The table below shows how the same 10 point gap sits at three places on the scale. It uses the normal curve with a mean of 100 and a standard deviation of 15, and the percentiles are our arithmetic, rounded to one decimal.

The same 10 points is a distance of nearly 25 percentile points in the middle of the scale and fewer than two at the top. A couple who discusses "who is ahead" in percentile terms will see a dramatic gap in the middle and an almost invisible one near the top, although the underlying difference is identical. The score gap on the standard scale is the fairer unit, and it is the one that the next section puts error bars around.

6 How Big Must a Gap Between Two Scores Be Before It Is Real?

A gap between two partners' scores is real only if it exceeds what measurement error alone would produce, and because two scores each carry error, the error band of the gap is about 1.41 times wider than the band of either score. The Standards for Educational and Psychological Testing address the principle in Standard 2.4: when a score interpretation emphasizes differences between two observed scores, reliability and precision data, including standard errors, should be provided for such differences. The comment to that standard adds that the precision of change scores can be much lower than the reliabilities of the separate scores involved. The standard speaks of two scores of one individual and of two group averages. Applying the same logic to two different people tested on the same instrument is our extension, and it is a conventional one.

The arithmetic is short. If each person's score has a standard error of measurement of s points, and the two errors are independent, the standard error of the difference is s times the square root of 2, about 1.41 times s. Take an illustrative standard error of 3 points per person. This is a teaching number and not an ACIS figure. The standard error of the difference is then about 4.2 points, and the 95 percent band for the difference is about 8.3 points. A gap of 8 points or less could easily be error alone, and a gap well above it is unlikely to be.

Intervals add a second lesson. With that illustrative error of 3 points, each person's 95 percent interval is about 5.9 points either side of the score. Two such intervals stop overlapping only when the gap exceeds about 11.8 points. Yet the gap becomes larger than error alone would produce at about 8.3 points. A gap between 8.3 and 11.8 points can therefore leave the two intervals overlapping and still be a real difference. Overlap is a quick check that errs toward calling things equal, and the square root rule is the sharper one. The page on reliability and validity explains where the standard error comes from.

Three cautions apply to the arithmetic. It assumes the intervals in the report are built from a standard error of measurement and that the two errors are independent. It describes the measurement of a score on one form, not the stability of the score between days. And it concerns the reliability of the gap, not its meaning. A real gap of 12 points says that two scores differ by more than error. It does not say why they differ, whether the difference matters in anyone's life, or which partner had the better day.

A fourth caution bears on conditions. An unsupervised session adds sources of variation that a norming study may not have contained: noise, interruption, fatigue and one person trying harder than the other. Those are not in the standard error that a report prints. The commentary in chapter 6 of the Standards says that technology has made it possible to administer tests in settings where conditions may not be strictly controlled, and that those who allow less standardization are responsible for providing evidence that it did not affect scores. For a couple, the practical reading is to treat a gap as more fragile when the sessions differed in conditions, and to repeat nothing, since retaking invites the practice effect described earlier.

7 How Big a Gap Should Partners Expect? What the Research and the Arithmetic Say

Partners are alike in measured IQ on average, yet the typical couple still differs by more than most people expect: at a partner correlation of 0.44, about 1 couple in 3 differs by 15 points or more. The partner correlation figures come from the page on IQ and relationships, which reports a meta-analytic partner correlation of 0.44 for IQ scores across 17 samples, from the 2023 meta-analysis of partner correlations by Horwitz, Balbona, Paulich and Keller, and 0.23 for fluid IQ in the UK Biobank, where the measure was a brief 13 item test. We keep those figures and their discussion on that page. The arithmetic below is ours and starts from them.

Assume two partners' scores are each normally distributed with a mean of 100 and a standard deviation of 15, that the two are related through a bivariate normal distribution with the stated correlation, and that scores are measured without error. The standard deviation of the difference between the partners is then 15 times the square root of 2 times (1 minus the correlation). At 0.44 that is about 15.9 points. The expected size of the gap, ignoring direction, is about 12.7 points. About 53 percent of couples differ by 10 points or more, and about 34 percent differ by 15 points or more. At 0.23 the share differing by 15 points or more is about 42 percent. If partners were paired at random, with no correlation, it would be about 48 percent.

These shares are our arithmetic under the assumptions above. Real measurement error widens every gap, so the true shares would be somewhat larger. The 20 point shares at 0.23 and 0.44 match the table on the page about IQ and relationships. Two points follow. A correlation of 0.44 is substantial for a psychological trait, and it still leaves a typical couple with a gap larger than the 8 point error band in the previous section. And a gap of 10 or 15 points between two partners is ordinary in the statistical sense: it is what a population of real couples looks like, not an anomaly in one of them.

A study of newlyweds adds a different detail, about which abilities partners resemble each other on. In Watson and colleagues (2004), the newlyweds showed moderate similarity in education and verbal intelligence and little similarity in matrix reasoning, and the similarity reflected initial assortment, meaning similarity at the time of marriage, rather than convergence with time. If that pattern holds more widely, two partners could be closer on a verbal index than on a fluid reasoning index. That is a reason to compare the profile and not only the headline number, which the next section takes up. It is one sample of newlyweds, and a single study cannot say how any couple will look.

8 Why Compare Index Profiles Instead of One Number?

Two partners can share a Full Scale IQ and differ on every index, and two different Full Scale IQs can hide a shared pattern, so the profile says more about how two people think than the headline number does. In the Cattell-Horn-Carroll framework that ACIS follows, a Full Scale IQ summarizes performance across several broad abilities. The page on the CHC model explains those broad abilities: comprehension-knowledge (Gc), fluid reasoning (Gf), quantitative knowledge (Gq), visual processing (Gv), short-term working memory (Gwm) and processing speed (Gs). The ACIS report shows a score for each domain a form covers.

Take three examples at the level of the subtest pages. Verbal comprehension (the Gc index) draws on tasks such as Vocabulary, which the subtest page describes as asking how precisely a person understands words and noting that word knowledge grows with years of reading, schooling and conversation. Fluid reasoning (Gf) draws on tasks such as Matrix Reasoning, where a person studies a pattern of shapes and infers the missing element from the rule structure, without needing verbal knowledge. Working memory (Gwm) draws on tasks such as Digit Span, which asks how many items a person can hold in immediate awareness and, in harder versions, reorder. A verbal index and a reasoning index are not the same ability, and the Watson result in the previous section shows why that matters for couples.

The reading load of each task differs, and for partners who differ in first language or schooling, that matters. The ACIS technical manual states that verbal and knowledge tasks are affected by English proficiency and educational exposure. If one partner learned English as a second language, a lower verbal index may measure English and not verbal ability in the person's strongest language. The sibling pages on the vocabulary IQ test and the general knowledge IQ test explain what word knowledge and general information tasks measure, which is the part of a profile most tied to schooling and language exposure.

Comparing six indices for two people creates 6 comparisons between partners and 15 pairs of indices inside each profile (our arithmetic). With that many comparisons, some gaps will look striking by chance. The Standards say, in Standard 1.14, that when interpretation of subscores, score differences or profiles is suggested, the rationale and relevant evidence in support of such interpretation should be provided. The page on how large a gap inside one profile must be works through the critical values, and the sibling page on the IQ profile test explains when a weak index is noise. Decide in advance which one or two comparisons you care about, such as verbal against reasoning, and treat the rest as texture.

A profile also invites a story about complementary strengths, and the studies reviewed on this page do not test whether complementary profiles help a couple. The only defensible use of a profile in a couple is descriptive: it shows, for each person, which abilities the test sampled and where the person's scores fall. For a short extra look at one domain, the page on the processing speed test describes an index that takes about 10 minutes, and its subtests, Coding and Symbol Search, are the two in the ACIS processing speed domain.

9 Does a Similar Score Mean a Better Relationship? What Couple Studies Found

The couple studies reviewed here find that partners are alike on many traits, and the one study that tested IQ compatibility against satisfaction directly found no significant relationship. The 2023 meta-analysis by Horwitz and colleagues pooled partner correlations across 22 traits, using 480 correlations from 199 peer reviewed studies, and added 133 trait correlations from up to 79,074 couples in the UK Biobank. Its abstract reports the highest correlations for political and religious attitudes, educational attainment and some substance use traits, and lower but positive correlations for psychological and anthropometric traits. It also reports high between sample heterogeneity for most traits. The page on IQ and relationships covers the IQ rows in detail.

Three further studies bear on satisfaction.

• Watson and colleagues (2004), studying newlyweds, report that marital satisfaction was primarily a function of the rater's own traits and showed little relation to spousal similarity.
• Luo and Klohnen (2005) studied 291 newlyweds with a couple centered approach. In the abstract, similarity was positively associated with marital quality for personality related domains but not for attitude related domains, with attachment characteristics most strongly predictive, and profile based similarity predicted better than absolute difference scores.
• Gignac and Zajenkowski (2019) studied 218 couples in a paper titled People tend to overestimate their romantic partner's intelligence even more than their own. People overestimated their own IQ by about 30 points, their partner's by 38 points (women) or 36 points (men), measured against Advanced Progressive Matrices. The estimates tracked tested IQ with some accuracy, and the degree of intellectual compatibility, objective and subjective, failed to correlate significantly with relationship satisfaction for both sexes.

The first two studies report similarity across many traits and domains, and their abstracts do not single out an IQ gap as a predictor of satisfaction. Only the third tested measured IQ compatibility against satisfaction directly, in one sample, and it found no significant relationship. Our reading of the first two is that the answer to whether similarity matters depends on what is compared and how, which is a reason not to take a single study as settled. Every one of these is an observational study of groups, so none can show what a gap does inside one couple. Two further limits apply. A study of newlyweds says nothing about couples who have been together for decades, and a study of couples who volunteered for research may not describe couples who did not.

The estimation result also has a practical meaning for anyone about to compare real scores. On average, people in that study estimated their partners' IQ about 36 to 38 points above what a matrix test showed. A real score can therefore differ from the picture each partner holds, in either direction, and that is not a defect in the test or the relationship. It is the ordinary distance between impression and measurement, and it is a reason to read the number as a measurement of one session, not as a revelation about a person.

10 What Should Partners Not Do With the Two Numbers?

Treat the two scores as two measurements of two people on one occasion, and avoid every use that turns them into a ranking, a verdict or a decision. The list below states the misuses that the evidence on this page rules out, and each one is a statement about measurement, not advice about anyone's relationship.

• Do not rank the partners. A gap that passes the error test, as in section 6, says that two scores differ. It does not say that one person is more intelligent in any sense a person would recognize in daily life, and the profile may favor each partner on different indices.
• Do not generalize from a group average. Group averages describe populations. The page on average IQ by gender reports small gaps between group averages and a wider spread among males, which predicts nothing about which partner of a given couple will score higher. The protocol and the arithmetic are the same for any two adults, whatever their gender.
• Do not compare unlike forms or unlike sessions. A Quick score beside a Full Scale score, a rested session beside a tired one, or a session with the other person present beside one alone, each mixes the difference in people with a difference in procedure.
• Do not retest to close a gap. The retest evidence cited in section 3 shows gains from repetition, which are practice and not growth in ability.
• Do not decide anything on the pair of numbers. The APA Guidelines for Psychological Assessment and Evaluation, written for psychologists, state in Guideline 7 that individual performance on psychological tests is only one piece of assessment, set in context and secured from multiple sources. A casual home comparison has far less than that.
• Do not read a low score as a deficiency or a diagnosis. A normed online test is not a clinical instrument, and a score does not diagnose anything. A person who has concerns about attention, learning or mood belongs with a qualified professional, not with a partner and a spreadsheet.

A last point concerns the number itself. IQ scores sit on a scale that people read as a ranking of worth, and nothing in the scale supports that reading. The page on the IQ score chart lays out percentiles and ranges, which describe a position in a distribution and nothing more. If one partner's score is far from the other's, the useful questions are about the measurement: was the session comparable, is the gap larger than error, and does the profile show the same pattern as the headline. The answer to whether the relationship is a good one lies elsewhere.

11 What Practical Conditions Do Two Unsupervised Sessions Need?

Each unsupervised session needs the same basics: a quiet room, working sound, uninterrupted time and a private screen, because an online test has no examiner to enforce them. The commentary in the Standards says that unmonitored settings, including the Internet, should meet standardized administration conditions, and that otherwise the report on scores should note that standardized conditions were not guaranteed. An online test cannot supervise a home session, so meeting the conditions is the test taker's job.

• Sound. Digit Span, Alphanumeric Sequencing and Arithmetic present their items as audio and need working sound. The first two are in every form, and Arithmetic is in the Full Scale form only. In a shared home, headphones keep the audio from leaking to the other partner, which is our suggestion, not an ACIS requirement.
• Time. Plan for about 45 minutes for Quick, about 110 for Optimized and about 175 for Full Scale. Breaks are allowed and progress is saved between subtests, and a form has to be finished within 30 days. Use the breaks between subtests, not during one.
• Privacy. A quiet room with the door closed and the phone away. The other partner should not be in the room.
• Speed tasks. Coding and Symbol Search reward steady, quick responding, and a distracted session lowers them first. Do not take them at the end of a long day.
• Separate accounts. Use a separate email address and a separate browser or device for each partner, because ACIS grants one trial per person.
• Language. The test is in English only. A partner whose first language is another one should expect the verbal and knowledge subtests to reflect English exposure.

The evening before is not the time to practice, and neither partner should look for what the tasks contain. The purpose of a normed test is a fair measurement, and rehearsing for it changes what is being measured. A relaxed session on an ordinary day gives the most usable number.

12 Where ACIS Sits, and What a Buyer Should Do

ACIS is a normed online battery that can give two adult partners comparable scores at a low one time price, and it is not a clinical instrument, not a supervised test and not a basis for any decision about a relationship. It is online and unsupervised, in English only, and its verbal and knowledge tasks are affected by English proficiency and educational exposure, as the technical manual says. It is not for hiring, for school accommodations or for admission to high IQ societies. Its strengths are the ones this page has used: one time prices for three forms, a Full Scale IQ with a 95 percent confidence interval on every form, six indices when the form covers them, a free trial with no card, and norms for ages 16 to 90. The documentation is in the technical manual, and this page quotes no figure from it.

The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) are the frame for reading the two scores. Standard 2.4 asks for reliability and precision data on score differences, which is why section 6 puts an error band on the gap. Standard 1.14 asks for rationale and evidence when profiles or score differences are interpreted, which is why section 8 limits itself to a descriptive profile. The commentary on unmonitored settings explains why section 11 puts the conditions in the hands of the test taker. The APA Guidelines for Psychological Assessment and Evaluation (2020) point the same way: Guideline 6 asks psychologists to select tools with sufficient validity evidence and score reliability, and Guideline 7 treats test performance as one piece of an assessment. Neither document gives a home comparison between partners any authority beyond the evidence behind the instrument.

The guarantee reads, literally: "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." It is a conditional guarantee, not an unconditional money back one, and prices were read on October 28, 2026 and can change.

What to do next is short.

1. Decide whether a normed score is what you want, or whether a conversation or a relationship quiz would serve better.
2. Pick one form for both of you. Quick is the low cost start, and Full Scale is the complete one.
3. Each partner starts the free trial with a separate email address and browser, or buys a form.
4. Take the sessions alone, on separate days, under similar conditions, and do not discuss the tasks until both are finished.
5. Read your own report first. Compare the two scores in IQ points, put the error band of about 1.41 times one interval on the gap, and compare profiles, not only the headline number.

13 Sources Behind This Page

Every third party fact and figure on this page comes from the documents below, opened or verified on October 28, 2026, and any figure we derived ourselves is labeled in the text as our arithmetic. Prices and policies on commercial and clinic pages change, so check the page before you pay.

1. American Educational Research Association, American Psychological Association, and National Council on Measurement in Education (2014). Standards for Educational and Psychological Testing. Washington, DC: AERA. Open access edition: https://www.testingstandards.net/open-access-files.html
2. American Psychological Association (2020). APA Guidelines for Psychological Assessment and Evaluation. Approved by the APA Council of Representatives, March 2020. https://www.apa.org/about/policy/guidelines-psychological-assessment-evaluation.pdf
3. Horwitz, T., Balbona, J., Paulich, K., and Keller, M. (2023). Evidence of correlations between human partners based on systematic reviews and meta-analyses of 22 traits and UK Biobank analysis of 133 traits. Nature Human Behaviour, 7(9), 1568 to 1583. https://doi.org/10.1038/s41562-023-01672-z
4. Watson, D., Klohnen, E., Casillas, A., Nus Simms, E., Haig, J., and Berry, D. (2004). Match makers and deal breakers: analyses of assortative mating in newlywed couples. Journal of Personality, 72(5), 1029 to 1068. https://doi.org/10.1111/j.0022-3506.2004.00289.x
5. Luo, S., and Klohnen, E. (2005). Assortative mating and marital quality in newlyweds: a couple-centered approach. Journal of Personality and Social Psychology, 88(2), 304 to 326. https://doi.org/10.1037/0022-3514.88.2.304
6. Gignac, G., and Zajenkowski, M. (2019). People tend to overestimate their romantic partner's intelligence even more than their own. Intelligence, 73, 41 to 51. https://doi.org/10.1016/j.intell.2019.01.004
7. 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
8. Pearson. Wechsler Adult Intelligence Scale, Fifth Edition (WAIS-5), product page, read October 28, 2026. https://www.pearsonassessments.com/en-us/Store/Professional-Assessments/Cognition-%26-Neuro/Wechsler-Adult-Intelligence-Scale-%7C-Fifth-Edition/p/P100071002
9. University of South Florida, Psychological Services Center. Clinic policies, read October 28, 2026. https://www.usf.edu/arts-sciences/departments/psychology/psychological-services-center/clinic-policies/
10. American Mensa. Testing, read October 28, 2026. https://www.us.mensa.org/join/testing/
11. Marriage.com. Are we intellectually compatible (quiz), read October 28, 2026. https://www.marriage.com/quizzes/are-we-intellectually-compatible

14 Frequently Asked Questions