They say the toughest walk you can take is that last trip down Death Row. But at least there you know what to expect. On August 22, 1986, I hadn't a clue. At the end of a hallway, in a corner room, on the top floor of U-Hall 4, was my destiny. Whatever I found behind that door would have a profound effect on my first year on the Hill, which would set the stage for my entire college experience, which would be my springboard toward adulthood. It was like playing "Let's Make A Deal" without options. The prize could be lifelong friendship--a soulmate through the hurdles of higher education. Or it could be a guy who collects toenail clippings and calls himself Captain Calculus. Behind that door was my freshman roommate.

Anonymous, Class of '02: "I'm from India, and I come all the way over here, and they give me this roommate who's so not like me. He's from Indiana. I listen to rap, hard rock, and heavy metal. He listens to all this weird opera stuff. You walk into our room, and on one side you see paintings by Monet and Van Gogh. On the other side, you see pictures of half-naked women and beer."

The roommate phenomenon is a petri dish of potential personality clashes, a situation personified by tongue-in-cheek websites with names like "100 Ways to Confuse Your Roommate." (No. 46: When your roommate comes in, pretend that you are on the phone, screaming angrily and shouting obscenities. After you hang up, say, "That was your mom. She said she'd call back.") It's an environment where the potential for conflict is high and the tolerance threshold can get remarkably low. How low? From the Class of 2001: "I just told him, I'd appreciate it if you don't make my bed anymore." From 2002: "He did yoga every night in the middle of the floor. So while I was typing a paper or something, he was making weird breathing noises." Or this, from 1989: "That's part of why she was so awful. She wasn't mean, and she wasn't particularly inconsiderate. She never gave me anything to retaliate against. She just annoyed."

There are the roommates who barely speak to each other, ones who talk deep into the night--and the rest, who fall somewhere in between. There are the dynamic duos, like the inseparable twins in The Lord of the Flies who aren't Sam or Eric but Samneric. There are the odd combinations: the East Side artist with the Alabama engineer, the offensive tackle from Iowa with the defensive émigré from Russia, the health food nut with the Twinkie junkie, the Deadhead with the ROTC, or--and this has happened--the guy who used to be a camp bully with the poor kid he used to pick on.

It's nothing so whimsical as chance or fate that determines your roommate; it's the Cornell Housing Office, which happens to be in the midst of a first-year experiment of its own. In the past, Cornell had focused on honoring building and location requests (now more difficult in light of the initiative to house all freshmen on North Campus by fall 2001) and asked an occasional lifestyle question on its housing application, usually something along the lines of, "Are you a smoker?" But in 1998, Campus Life decided to take a different tack. Patrick Savolskis, who was hired as manager of the housing office in June 1998, polled universities traditionally competitive with Cornell about their roommate assignment procedures. "The one thing I got from talking to my colleagues," he says, "is there's no perfect answer. Everyone has a different system."

Consider Princeton, whose housing form asks nearly a dozen multiple-choice questions, including this: What six words accurately describe your ideal roommate? "People typically use terms like friendly, outgoing, considerate, sometimes words like spiritual or liberal--things that probably describe themselves," says Adam Rockman, Princeton's coordinator of undergraduate housing. "If one student lists their ideal roommate as ultraconservative and another lists it as ultraliberal, we're going to steer clear of putting them together."

Princeton uses a computer program to find students with similar interests, then human staffers make the room assignments. Stanford's process is similar, but the final decisions are made by a pair of upperclassmen. "We kind of control for some things. We don't let people of the same ethnicity live together, and we also don't let people from the same region of the country live together," says Chris Walton, one of this year's decision-makers. "And we probably wouldn't put someone who likes Garth Brooks with someone who likes hard-core gangsta rap."

Anonymous, Class of '55: "I grew up in a little town in Virginia. I was anything but sophisticated. I come to Cornell and meet my roommate, and he's the suave, sophisticated guy from New York City. He knew what bleu cheese was. He knew what a martini was. I had never even heard of one. But our parents both happened to be in the paint business, so they immediately became friends. And you know what? So did we. It's turned into a nearly fifty-year relationship."

Some schools have the luxury of spending more time on fewer roommate pairs. Davidson College in North Carolina, for example, goes beyond the standard housing preference questionnaire. The residential life staff uses a Myers-Briggs Type Indicator to identify personality types and even scours application essays for pertinent details. But Davidson averages only 460 freshmen a year, compared with 3,228 in Cornell's Class of 2003. Most colleges are limited by size, which is why Dartmouth asks four lifestyle questions on its housing application and Penn asks only three, one of them being, "Do you keep kosher?"

Campus Life concluded it was time for Cornell to ask some lifestyle questions of its own. Given that Cornell's freshman class is almost three times the size of Dartmouth's and Princeton's and nearly twice that of Stanford's, it was impractical to use human judgment in assigning roommates. In the end, Savolskis chose to keep the questions to a minimum, primarily to assure that the computer software performed well. "The more parameters you have, the harder it is to match," he explains. "So we kept it simple." The fickle finger of freshman fate has been boiled down to these eight queries:

• Do you prefer a single-gender or co-ed residence hall/floor/wing?

• Do you prefer a single room?

• Do you smoke?

• Would you room with a smoker?

• Can you sleep with background noise or a light on in the room?

• Can you study with background noise in the room?

• Are you an early bird or a night owl?

• Is a neat room important to you?

"Depending on those answers, we'll try to match you up on our computer system with someone who's answered the questions in a like manner," says Savolskis. "We've essentially ranked the questions. For example, smoking is a high one. They can't really negotiate that. That one we try to match one hundred percent." However, those preferences hardly describe a freshman in full. Two neat, night-owl, non-smokers could be worlds apart in every other way. And there isn't a housing questionnaire yet that asks how loud you snore, how often you hit the snooze button, or why you insist on listening to the same album over and over.

But even perfect questions in no way guarantee perfect answers. Two freshmen may appear to be perfectly matched--until it turns out that somebody's parents answered the questions. Even if the student filled out the questionnaire, once that high school senior enters college you can just about throw those answers away. Says Savolskis: "Often, a couple of weeks into the school year, they're different people." The non-smoker might start smoking. The early bird can develop a yen for after-hours parties and 11 a.m. classes. Or perhaps someone who expects to stay up late decides to join crew. Suddenly, he's going to bed at ten and waking at five.

Anonymous, Class of '81: "When he started, he was very studious, like I was. He was very smart, a straight-A kind of guy. But all of a sudden, three or four weeks into school, he stopped studying. He started concentrating more on his friends and his comic books. When I first met him, he had very short hair. But he never cut his hair for the length of the school year--not once. I never would have guessed that when I first saw him."

Cornell's housing application declares, "On the basis of this information and available accommodations, the Housing Office will do its best to match you initially to roommates who have listed similar preferences." The key word there is initially. The hope is that better matches will decrease the number of roommate conflicts. But room-change requests are inevitable, and there's no ideal system for handling them. Penn allows two one-week room-change periods, one in October and one in December; a lottery determines which students get to relieve their self-described suffering. Dartmouth forbids any changes until winter term, the second in its three-term system, and then it's up to the student to find a classmate willing to switch. At Princeton, it's possible to switch rooms, but it's strongly discouraged. And Harvard? "They take a lot of information from incoming freshmen. But once the students get the assignment, there's no switching," Savolskis says. "They basically say, 'Tough. You're lucky to get in.'" Cornell tries to be flexible. "For the most part, the assignment has to be concrete just because of sheer numbers, not because we're insensitive to the needs of parents or students," says Savolskis. "Schools guarantee housing, but I've never heard of a school that guarantees your preferences."

Still, every summer and fall, before and after room assignments are made, the Housing Office fields hundreds of phone calls from uneasy students and parents. Once the school year has begun, there is a two-week moratorium on room changes, so the Housing Office can determine who enrolled, who didn't, and how many vacant beds are available. Savolskis then makes offers once a week based on available space. But first, he and his staff play couples counselors to the embattled roommates, in the hope of forging peace. "Part of this education," says Savolskis, "is trying to teach students how to confront their peers, not in a bad way, but to say, 'Hey, when you get up at five o'clock in the morning, you make a lot of noise. Can we do something about this? Can you take your clothes out the night before? Can you stop your crewmates from pounding on the door to make sure you're awake?' They don't often have that discussion, so the easier thing is just to move."

Over the years, in various housing offices, Savolskis has heard his share of complaints, some nitpicky and some valid. He's dealt with students unplugging a roommate's alarm clock, rearranging the furniture, deleting messages from answering machines, changing computer passwords, even hiding a roommate's letters out of spite. Often, students keep their complaints to themselves because, as one member of the Class of 2002 explains, "You have to live with this person the rest of the year, so you don't want any tension there."

Anonymous, Class of 1991: "She was basically an insomniac, and we had a fan in the room because it helped her sleep. Every single night, when she couldn't sleep, she'd turn on the fan. The noise annoyed me, so I'd soon get up and turn off the fan. Then later, she'd get up and turn it on. We'd go back and forth all night. And we never once discussed it."

A study by the Association of College and University Housing Officers International has found that nearly 90 percent of freshmen now arrive on campus having never shared a bedroom, compared to perhaps five percent twenty years ago. It's not surprising, then, that throughout Cornell and the rest of the nation there appears to be an increasing demand for single rooms. "A blind date," says Savolskis. "That's really the best way to describe getting a roommate." And, just as in a blind date, first impressions are paramount--even on paper. One member of the Class of '81, an Ithaca native, remembers receiving notice during the summer that his roommate was from Queens. "I thought, great, here's an obnoxious guy from New York. And he thought I was some hick from the sticks," he recalls. "But we bonded. He understood me, and I understood him." Often, however, those first impressions stick. As a '93 grad recalls of her first days on the Hill, "I was willing to give everyone a chance to be my best friend--except my roommate."

Cornell, like most universities, suggests its incoming students contact their assignedroommate over the summer, if only to prevent a room being cluttered by two TVs, two answering machines, and two mini-refrigerators. The university also tries to honor specific roommate requests, such as two freshmen who met in high school or at an orientation. Such arrangements may work out fine--or the familiarity of a dorm room may breed contempt. "Sometimes you get exactly what you ask for," says Savolskis, "and it's not what you want."

So maybe it's a crapshoot, a flip of the coin, the luck of the draw. Or perhaps there are more profound psychological factors at work. "In study after study, propinquity--being physically close in ways that facilitate interaction--has been shown to be positively associated with liking," says Dennis Regan, associate professor of psychology, who teaches Introduction to Social Psychology at Cornell. That is, friendships are more likely to form between people living near each other.

But one classic study, published in 1976, illuminates the precarious nature of such close quarters. Using interviews with residents at an Irvine, California, condominium complex, three researchers at the University of California, San Diego, examined the relationships between physical distance, frequency of face-to-face contacts, and the probability that individuals would be liked or disliked. While they found that people are more likely to become friends--and tend to be better friends--the closer they live to one another, they also discovered that proximity appeared to be an even more important factor in disliking.

"But the psychology of that is different," Regan explains. "The psychology of liking is about interaction. Propinquity sort of stimulates having conversations and getting to know each other. But the researchers attributed disliking to what they called environment spoiling. That is to say, somebody's obnoxious taste in music, for instance, is much more likely to impact on you and make you dislike them if they're in your room or the room next door than if they're way down the hall or in another building."

So what does it all mean in the grand scheme of first-year fates? Perhaps merely that sharing space can be an intense experience--one way or the other. "If things go well, there's a good likelihood that you can be close friends," says Regan. "But if things go badly, you can be bitter enemies. Come to think of it, that's roughly how I recall my freshman year."

Brad Herzog '90 shared his freshman dorm room with his roommate and his roommate's pet alligator.

Published in Cornell Alumni Magazine. Used by permission.