Science
Personality and relationships
Are Love Languages Real? What the Research Says
The CQ Team · 8 min read ·
Love languages have weak research support. Studies of up to 954 couples find matching on a primary love language does not make couples happier.
Love languages are a real idea with weak evidence behind their main claims. Gary Chapman's 1992 book says each person has one primary way of feeling loved, that there are five such ways, and that couples are happiest when each partner "speaks" the other's. Studies since then, including a preregistered study of 499 people and a study of 954 couples together an average of 27 years, find little support for any of the three. What does hold up is plainer: people who receive a wide mix of care from a partner are more satisfied.
Chapman named five: words of affirmation, quality time, receiving gifts, acts of service and physical touch. His quiz, the Love Language Personal Profile, has 30 forced-choice items. Each item makes you pick one of two options, such as "we hold hands" or "my loved one gives me a gift". More than 30 million people have taken it, according to the main scientific review of the framework by Impett, Park and Muise (2024), led by Emily Impett, a psychologist at the University of Toronto Mississauga.
That review sets out the three claims the theory rests on and checks each against the studies. Later work has tested them with larger samples.
| Chapman's claim | What the research finds |
|---|---|
| Each person has one primary love language | People rate all five highly, and fewer than half have a clear favourite |
| There are exactly five | The five overlap heavily, and models with 7 to 10 types fit better |
| Couples who speak each other's language are happier | Couples who match on a primary language are no happier than couples who do not |
Most people do not. In a preregistered study of 499 people in long-term, live-in relationships, fewer than half had an identifiable primary love language (Flicker, Sancier-Barbosa and Impett, 2025, reviewed in Impett, 2026). The same research found that people with a sharply defined primary language reported lower relationship quality and more doubt about whether their partner cared.
When researchers let people rate each language on its own scale, instead of forcing a choice, people score all five at about 4 out of 5 (Impett et al., 2024). Which language counts as "primary" then depends on the quiz format. Gifts came out on top for 0 to 4% of people on Chapman's forced-choice quiz, and for over 50% of people in one study that used the highest rating-scale score.
The forced choice also fails to agree with people's own ratings. In Polk and Egbert's (2013) study of 83 couples, the language people picked on the forced-choice quiz did not predict which language they rated highest on a separate scale.
If your quiz result surprised you, try rating each of the five on its own from 1 to 5. Most people find more than one near the top.
Quality time, in the one couples study we found that reports the split for both sexes. Mostova, Stolarski and Matthews (2022) gave the forced-choice quiz to 100 heterosexual couples of 31 nationalities, with an average age of 27.
| Primary love language | Men | Women |
|---|---|---|
| Quality time | 42% | 41% |
| Physical touch | 28% | 30% |
| Words of affirmation | 16% | 11% |
| Acts of service | 14% | 14% |
| Receiving gifts | 0% | 4% |
This was a convenience sample recruited through social media. No published study we found gives the share for a representative population, and we found no published data from the official quiz.
The five overlap too much to be five separate things. Ratings of the different languages correlate at r = .54 to .75 (Egbert and Polk, 2006, cited in Impett et al., 2024). In a study of 954 couples, some pairs of languages correlated as high as r = .795 (Chopik et al., 2025). A correlation near .8 means people who score high on one language almost always score high on the other, so the two scales are largely measuring the same thing.
Starting from scratch gives a different count. Flicker and colleagues asked three US samples of adults in long-term relationships (696, 500 and 499 people) how they express and receive love, then let the data sort the answers. Models with 7 to 10 love languages fit better than Chapman's five and predicted relationship quality better (Flicker et al., 2025, reviewed in Impett, 2026). One behaviour the five miss is bringing a partner into your wider social circle. In the preregistered study of 499 people, support during difficult times, encouragement for a partner's own pursuits and accountability were among the strongest predictors of relationship quality.
Impett's 2026 review adds a reason the list is narrow: Chapman built it from his counselling work with married, religious, mixed-gender couples with traditional family values, and that setting shaped what he counted as love.
None of the studies we found that compared matched and mismatched couples found the matched couples happier.
The main study pointing the other way measures something broader than a primary language.
Mostova and colleagues (2022) measured the gap between how much of each language a person wants and how much their partner gives, then added up the gaps across all five. In their 100 couples, a bigger gap went with lower relationship satisfaction (r = -.36 for men, -.30 for women), and with lower satisfaction in the partner too (r = -.40 and -.31). Correlations of that size are moderate: they matter, but most of what makes a couple satisfied lies elsewhere. The authors note the design cannot show cause, since unhappy partners may also give less, and they recommend dropping the idea of one primary language.
Read together, the null studies are the stronger evidence. They are larger, one was preregistered, and they test Chapman's actual claim: that your primary language matters more than the others. The positive result fits a simpler account that Impett's team proposes: any loving behaviour helps, whichever "language" it belongs to.
Impett's 2024 review offers a different metaphor: love works like a balanced diet, where people need a range of nutrients, and no single food covers it. The review points to responsiveness, a partner noticing and meeting your needs, as the process that decades of relationship research back. Asking "what is your love language?" can start that conversation. Impett's 2026 review cites work suggesting couples who use the framework to talk about what makes each of them feel cared for may benefit, even without sticking to a partner's "preferred" language.
The 954-couple study also shows where to look instead. Personality traits and attachment styles explained the outcomes that love-language matching did not. On CQ you can read what agreeableness and neuroticism measure, and how attachment styles describe the way each partner handles closeness. For the question of whether two personalities fit, see what research says about personality compatibility, and for more on couples, family and friends, the personality and relationships hub.
If you want to see where you sit on the traits, the CQ Big Five test is a short test in the style of the Big Five framework, not the official instrument.
The love language quiz sorts people into five boxes, but the couples studies show that the partners who do best are the ones who give and notice care in many forms. Asking your partner which of the five matter to them, and how much, gets closer to what the research supports than finding one word for each of you.
How we checked: we read Impett, Park and Muise (2024) and the Mostova, Polk and Egbert, and Bunt and Hazelwood papers in full, the Chopik (2025) abstract, the three Flicker abstracts, and Impett's 2026 review. Further reading: the University of Toronto's summary of the 2024 review.
It has not been shown to be. The official quiz forces a choice between pairs of options, so it cannot show that you value several languages equally, and in one study of 83 couples the language people picked did not match the one they rated highest on a separate scale. We found no published reliability data for the official quiz.
Chapman's quiz measures how you prefer to receive love. Research that measures both finds the gap matters: in Mostova's 100 couples, satisfaction was higher when what a partner gave was closer to what the other person wanted, summed across all five languages.
Nobody knows yet. Almost every love-language study measures people once, and we found no study that has tracked the same people's preferences over years.
The research so far studies romantic couples only. We found no peer-reviewed study testing the five love languages between friends or family members.
Curious where you sit? The MBTI test is free, takes about 3 minutes, and needs no sign-up.
Take the MBTI test →