what predicts a good roommate
what predicts living well together is the alignment of concrete living habits, schedules, expectations, and communication, not personality type and not demographics. that is the corrected reading of the research: the intuitive idea that similar personalities make better roommates did not survive scrutiny, and what held up instead was whether two people's actual habits and the way they communicate fit. it is a less romantic answer and a more checkable one, because habits are things you can ask about before you sign. the honest qualifier comes first, because the rest of the page leans on it: this evidence is correlational and mostly drawn from first-year college students, so the language stays "predicts" and "associated with," never "causes."
ask what someone does, not what they are. habits are checkable before you sign; a personality is not.
the corrected reading of the roommate-compatibility research
what does the research say predicts a good roommate
it points at habits and communication, in that order, over personality. the most direct study here looked at what tracked roommate relationship satisfaction and found it was the alignment of concrete habits and communication patterns, not global personality similarity (niu & brown, 2023). a separate line of work found that perceived similarity in communication patterns, not personality, was what was associated with roommate satisfaction (martin & anderson, 1995). so the predictor is not "are we alike as people," it is "do our daily defaults fit, and can we talk about it when they do not." the figure below is the practical version of that distinction: the predictors the research points at are the ones you can check before move-in, and the one it deflates is the one a first meeting feels like it tests.
| signal | what the research says about it | checkable before the lease? |
|---|---|---|
| concrete living habits (cleanliness, noise, sleep schedule) | aligned habits track satisfaction (niu & brown, 2023) | yes — by asking specifics |
| expectations (guests, money, quiet hours) | shared, explicit expectations are part of fit | yes — by asking specifics |
| how someone communicates about friction | communication-pattern fit is associated with satisfaction (martin & anderson, 1995) | partly — you can read for it |
| conscientiousness (tidy, reliable, follow-through) | the one trait that keeps mattering — because it is a habit proxy | partly — via the habits it stands for |
| global personality similarity / 'good vibes' | did NOT hold up as a predictor of satisfaction | feels checkable, mostly is not |
Framing of Niu & Brown (2023), Martin & Anderson (1995), and Carli (1991); see sources below.
did personality similarity not predict roommate satisfaction
no, and this is the finding worth correcting first, because it is the one most people assume. the intuitive claim, that matching two similar personalities predicts a satisfying roommate relationship, was refuted across the verification we hold this hub to. when satisfaction was measured against aligned habits and communication versus global personality similarity, the habits-and-communication side is what carried it (niu & brown, 2023). the personality-matching story is the loser here. it makes a tidy quiz and a poor prediction. the corrected version is the one above: living well together is mostly a habits-and-communication question, and personality enters in a narrower, more specific way than the matching story claimed. it is worth marking the size of the study that anchors this, because "the research says" should come with its sample.
why does conscientiousness still matter when other traits do not
because conscientiousness is the trait that operationalizes habits, so it is really a habit wearing a personality label. of the big personality traits, it is the one that consistently keeps mattering for roommates, and the reason is mechanical: conscientiousness shows up as the concrete, checkable behaviors, being tidy, reliable, following through, paying on time. those are exactly the habits the rest of the research says predict living well together. so conscientiousness survives as a predictor not because personality matching works after all, but because this one trait is a stand-in for a pattern of behavior. read another way, it is the same answer: the thing that predicts a good roommate is a reliable habit pattern, whether you measure it as a habit or as the one trait that tracks habits.
| the trait reads as… | …the checkable habit underneath |
|---|---|
| conscientious | dishes actually get done; the place stays at an agreed baseline |
| reliable | plans get followed through; the chore that was agreed to happens |
| responsible with money | the rent and the shared bills go out on time |
| organized | shared supplies get tracked and restocked, not left to one person |
Framing of Niu & Brown (2023); conscientiousness as a habit proxy.
is more togetherness better, or is privacy part of compatibility
privacy is part of it, and the research frames a good fit as a balance rather than a maximum. the picture from a review of shared-living research is a fine balance of sociability and privacy, not the maximization of togetherness (clark et al., 2018). a compatible household protects each person's autonomy and private space about as much as it makes room for shared time. this is a synthesized framing, the evidence describes a balance, it does not prescribe an optimal ratio, so the honest version is a direction, not a number: a roommate who needs a lot of quiet and one who needs a lot of company are not incompatible by trait, they are incompatible only if the apartment has no agreed room for both. the balance is the thing to design for, not a single "more social is better" default.
why does communication keep coming up
because how a friction gets handled is its own predictor, separate from whether the friction happens. two strands of the research are about this. one found that roommate conflict is shaped by attributions, how a roommate reads the intent behind a behavior, more than by the objective disagreement, and that integrative, problem-focused communication was associated with better resolution than avoidant or win-lose styles (sillars, 1980). the other found that a person's own communication competence was the strongest predictor of the intention to end a roommate relationship, the lever was more about how you communicate than about the other person (bahns, 2013). note the careful version of that last one: the outcome measured was the desire to dissolve the relationship, not an actual move-out. together they say the same practical thing as the habits findings from a different angle: aligned habits lower how often you collide, and good communication decides what a collision costs. and friction is the normal case to plan for, not the exception, in the population this is best measured on.
what should you actually screen for, then
screen for the concrete habits, schedules, and expectations the research points at, and read for how someone communicates, because that is what predicts the rest. the perceived-similarity trap is why this matters: perceived similarity, how alike you think you are, predicts liking more strongly than actual measured similarity does (carli, 1991), which means two people can both call themselves "easygoing" and mean opposite things and still screen each other as a match on the word. the fix is to ask what someone does, not what they are, and to ask for specifics: when the kitchen gets cleaned, what time the place needs to be quiet, how guests work, how money is handled, and what they do when a roommate is doing something that annoys them. the questions to ask a potential roommate and the breakdown of what makes roommates compatible go deeper on the specific questions; the point here is that the things the research says predict a good roommate are exactly the things a first meeting does not reveal. it lines up with how roommates actually fight from the other direction: aligned habits lower how often you collide, and your own communication decides what a collision costs.
this piece is a synthesis of the published roommate-compatibility and similarity literature, not a first-party study. every number traces to a named source in the list below: the n=479 figure is the sample of niu & brown (2023), the strongest single roommate study here; the n=30 figure is carli's (1991) lab sample for the perceived-versus-actual-similarity finding, which is not roommate-specific and is treated as a similarity-research principle applied to roommates; and the ~48% conflict figure is HERI/UCLA's Your First College Year survey (2007), a sample of roughly 31,500 first-year students (sometimes cited as Liu/Sharkness/Pryor 2008, the same survey artifact, not a second source). the two comparison figures are conceptual maps of the research, sorting signals by whether they are checkable before move-in or translating one trait into the behaviors it stands for. they do not assign effect sizes; the underlying studies are correlational and do not rank these predictors numerically, so the figures are deliberately drawn without numbers to avoid implying a precision the evidence does not have.
it can say, with reasonable consistency, that aligned habits and communication are associated with living well together and that global personality similarity is not, with conscientiousness surviving as a habit proxy. it cannot promise a good roommate or claim any of this causes anything, these are correlational findings and the language stays "associated with" and "predicts" on purpose. the largest caveat is the sample: most of the strongest roommate studies were run on first-year college students (and the conflict-prevalence figure here, ~48% of first-year students reporting frequent or occasional conflict, is explicitly that population, per HERI/UCLA's Your First College Year survey, 2007), so the transfer to adult renters is reasonable and consistent with broader similarity and communication research but has not been measured the same way on them. the perceived-versus-actual-similarity finding is from a small, non-roommate lab sample (carli, 1991), so it travels as a principle, not a roommate measurement. and even a clean screen predicts the likely collisions, not the lived outcome: living with the person is still the part that decides it.