In a significant shift of opinion, house masters yesterday agreed in principle to reform the housing lottery after reviewing College statistics that show a lack of academic diversity among the houses.

While still disagreeing over the extent of change, masters unanimously concluded that grade point averages, academic concentrations, numbers of honors students and extracurricular activities vary significantly by house and that the discrepancies should be corrected, masters and Dean of the College L. Fred Jewett '57 said after a meeting yesterday.

The masters decided to investigate several new reform options and set a January deadline for a decision, pending further debate among themselves and with students, Jewett said. Any drastic lottery overhaul will be postponed until the following academic year, he said.

"I feel it is now quite clear we are going to have to investigate changes in the system," said Eliot House Master Alan E. Heimert '49, who until now was a leading opponent of house assignment reform. "The data...helped persuade me that there are indeed some problems. There is a situation that calls for some changes."

Heimert, however, said any changes should be sharply defined. "Can we deal with the problem with surgical instruments rather than a bulldozer?" he asked.

Masters attributed their change of heart to new statistics detailing house-by-house proportions of honors graduates, private school graduates and scholarship students. The figures, drawn mainly from the past two years and released to the masters Monday, include house breakdowns of grade point average, extracurricular activity, academic concentration and race.

"Academics is the central concern, and everything revolves around it," said North House Master J. Woodland Hastings. "It was a unanimous feeling that there is a problem."

According to Leverett House Master John E. Dowling '57, wide discrepancies in academic performance and concentration distribution exist between houses at each extreme.

Socioeconomic differences between houses received less attention from the masters, and racial distribution generated little controversy, masters and the dean said.

Student representatives from the Undergraduate Council on the Committee on House Life will receive the statistics in outlined form today and tomorrow, they said.

The masters' new consensus removes a major stumbling block from the plan to change the lottery from its present form, which maximizes student choice. Last year, opposi- tion from some masters helped scuttle aproposal to reserve 25 percent of the places ineach house for random assignment.

Four broad options for changing the lotteryhave emerged, Jewett said. One is thestudent-generated proposal for "non-orderedchoice," in which students select four houses andare randomly assigned to one of them. Anotherwould modify "partial randomization" to allowstudents partial access to a second- orthird-choice house if their first choice meets itsquota, Jewett said.

A third option, comparable to the ratios nowplaced on gender distribution in the houses, wouldset limits on house assignments based on academicconcentration or group ranking. Totalrandomization is a final option.

A "small majority" of masters favor some degreeof randomization, Dunster House Master Karel F.Liem said.

The masters have agreed to proceed slowlythroughout the next several months while reviewingadditional public input, Dowling said.

Heimert has asked the dean to conduct furtherstatistical evaluations to test the impact ofvarious combinations of changes, Heimert andJewett said.

Other masters could not be reached or declinedto comment last night after the meeting