(FORTUNE Magazine) – Asked why he is so active on the philanthropic front, Walter Annenberg once | said, ''It gives me a good feeling.'' Asked about his recent supermunificent ($365 million) contribution to three universities and a prep school, Walter said, ''Good heavens, education is the key to the future of our country.'' Not asked to defend the quality of education at the Annenberg School for Communication (which got $120 million of the loot), he can count himself lucky. As we were all reminded when the latest grants were announced, the Annenberg School, based at the University of Pennsylvania, is not like other academic journalism centers. Instead of training reporters and filmsters, it emphasizes theoretical and critical studies of the media. Penn President Sheldon Hackney, currently onstage as Bill Clinton's choice to run the National Endowment for the Humanities, crowed that the grants would enable the school ''to maintain its quality and creativity in perpetuity.'' Just what we were afraid of.

The school is forever landing on the front pages with studies laboring dopily to prove that television is an instrument of social oppression. Back in 1977 it garnered huge headlines with a report complaining inter alia that Archie Bunker kept dominating Edith in All in the Family. A few years later it was back with a dizzying array of statistics demonstrating that women had only 25% or 30% of the roles in TV dramas, vs. the 51% they were theoretically entitled to, and Alaskan natives did worse. Later still was an Annenberg School report making many of the same points but with radicalized rhetoric: The systematic ''undervaluation'' of women and minorities was now being characterized as an exercise in ''symbolic annihilation.'' We claim to have been the only journalist in medialand who tried at the time to ascertain whether the report's raw data supported all this gloom and doom, and in the process made a few stunning discoveries. One was that the underlying data showed nonwhites and women as far more likely than white males to be cast as ''good guys.'' Another showed that on a scale where researchers were estimating the characters' degree of success or failure in life, blacks and women were more likely than white men to be getting ahead. Oh? You don't really find that stunning? The school's latest big outpouring was produced by professor George Gerbner, formerly dean of the Annenberg School, and published just a few days before Walter's new grants were announced. This study was based on data assembled by the school's notorious (around our house) Cultural Indicators research team, and was commissioned by the Screen Actors Guild and the American Federation of Television and Radio Artists, both said to be interested in getting the media to ''reflect more fairly'' the presence in our lives of women, ethnic groups, old folks, and the disabled. The study scales new heights of silliness. Intent on incorporating a kvetch about children's cartoon programs, but also determined to measure TV characters' social and ethnic characteristics, the CI folks ran into the problem that ''anthropomorphic animals and other creatures are not easily classified.'' Two out of ten cartoon characters are conceded to be ageless and more than half unclassifiable by race, yet the report solemnly broods over the fact that the shows introduce children to few Asian/Pacific Americans. In adult TV dramas, the report concedes that men are more likely to be ''bad'' than women, but then worries about the fact that the pattern is reversed among the elderly. Another weird worry: Among the mentally ill, there are ''actually more villains than heroes.'' Actually. Especially alarming is the fact that people of color get only 13% of the prime-time network roles, even though they are ''estimated to reach a majority in America by the year 2000.'' Huh? The U.S. Census Bureau expects blacks, nonwhite Hispanics, Asian Americans, and Native Americans to constitute about 18% of the population seven years from now. A nice question is whether the $120 million will reduce or augment the fatuity rate.