Although Harvard professes to spend a great deal of time and effort in matching up roommates, there is great speculation that, in reality, the College uses a very large dart board located in the bowels the Freshman Dean's Office.

Try as it may to set up compatible groups, Harvard's careful placement system is far from flawless. In fact, graduates of the Yard have also speculated that Freshman Dean's Office uses such criteria as first names, bicep measurements and predisposition to caffeine addiction in determining rooming groups. Officials, however, contend that such random attributes usually aren't given consideration, even though coincidental similarities might surface.

Actually, arranging freshman rooming groups is a month-long project in social engineering. While most colleges rely on a random computer selection process. Harvard's plan is much more personal. Using gut instinct, information from the "Application for Rooms" submitted by each student, and a host of other assorted admissions data, six ordinary, subjective human beings spend the month of July pairing, tripling and quadrupling roommates.

"You want some diversity, but not so much diversity as to make roommates impossible to live with," says Thomas E. Hassan, senior adviser for the Union dorms, who is running the room assignment prooess this summer. Smoking habits, preferred size of room, level of sloppiness, extracurricular interests and other tidbits listed on freshman housing applications receive consideration, Hassan says. Even background records--which include your hometown, college board scores, ethnicity an financial status--will work their way into the process.

The process begins in June when a computer spews out the names of the 1,595 future froshies into six randomly selected groups. The six groups correspond to a chunk of the Yard dormitories--North, South, East, West, Union and Canaday. In cases where disabled students, single-room seekers and other special cases are placed in an unsuitable area, "trades" are made by exchanging names. This year, 25 such exchanges were made.

For the rest of the month, the half dozen freshman senior advisers examine student records more closely The first step is to separate smokers from non-smokers. From then on, the process becomes more subjective, almost artistic. "It's a very handcrafted process with a lot of individual judgement involved," says Associate Dean of Freshmen Susan W. Lewis. The goal of the project is to achieve Harvard's motto, "diversity," without creating a situation that would create havoc inside the dorm room come September, officials say.

Hassan, who has arranged freshmen in Union dorms for the past three years, says the tricky thing after attaining some diversity is to "make sure people share some interests in common."

Senior advisers also look at room-size requests and desired level of neatness. Geography becomes important at this point, too. Officials say a trend in conjoining roommates is placing non-Bay Staters with locals so those new to the Boston area can cultivate a home away from home. "We wouldn't put two Californians, an Alaskan and a Swiss in one room," says Lewis.

Besides just grouping rooms together, the six advisers, who each place about 250 freshmen, must also dollop the freshmen into floors and entryways. Most freshmen will end up in double and quad blocks. The rest are grouped into triples, quints and few groups of six are even formed.

Despite all the planning, problems and minor conflicts arise. But Hassan is quick to point out that the fault may not lie with his crew. In some cases, students are not particularly accurate when filling out their housing applications.

Yet, officials say that except for medical excuses, very few requests come in for room changes. In fact, Harvard stresses that freshmen learn to work out dorm room squabbles on their own, Lewis says.

A few "safe singles" are reserved in Greenough and Canaday, mainly for freshmen with mobility problems such as broken legs and with serious illnesses. But, they also have been known to be open to some students who just need a vacation from their assigned suite. Freshman officials, however, stress that roommates should stay the course before resorting to mental sabbaticals.

In the long run, statistics show that the system, for the most part, works. Figures have not been updated lately, but a study conducted several years ago found that 70 percent of upperclassmen roomed with at least one of their freshman roommates.

It wasn't until the influx of Harvard computers six years ago that freshmen received roommate addresses during the summer enabling them to write to or call their future roommates, says Lewis. Freshmen will see the results of the process in the middle of August, when the dean's office sends a computer print-out listing the names and addresses of future roommates.