There is prima facie evidence to support Putnam's accusation that television is a major culprit in the killing of social capital in America, but certainly no open-and-shut case. As Putnam himself observes elsewhere in his book, "The astounding series of poor predictions about the social consequences of the telephone is a deeply cautionary tale." Caution is to be exercised here by sticking to a basic rule of evidence: We need to cross-examine television with careful attention to other sources of social capital. So, for instance, by every statistical clue, level of education is a salient source of social capital; it has shown itself to be a virtual philosopher's stone for converting leisure time into groups joined, friends visited, and countless other kinds of association. Tossing out education by ignoring or statistically controlling it would be like tossing out the testimony of the police in a court case.

Figure 1 displays Voluntary Association by level of TV viewing for each of three levels of education. Clearly, more TV viewership is associated with less group membership, as indicated by the downward tilt of all three lines to the right. Even clearer, though, is the size of the separation between the lines. In statistical terms, education is over ten times as powerful in inflating Voluntary Association as television is in deflating it.

Figure 2 performs the same operation for Informal Socializing with weaker results. The downward tug of TV-watching on network interaction is barely perceptible, but the upward push of education is unmistakable. Even accepting for the moment that TV-watching does erode both Voluntary Association and Informal Socializing, the rapidly rising educational levels in the United States would seem to be a much more powerful countertrend building them back up. Besides, more educated folks tend to watch less television.

Putnam's concern is that Americans across the board are, in effect, blowing off the PTA meeting to stay home and watch The Simpsons. If true, this means TV is driving out social capital, which does fit the data pattern in Figures 1 and 2. Also consistent with the data, however, is a mental picture that has people deciding not to go to the meeting, then flicking on The Simpsons for something to do. If this is the true picture, the causal arrow goes from social capital decisions to TV as a time-filler rather than a time-stealer. At this point, the evidence for either case is pretty circumstantial.

In advance of a million-dollar study to settle this question, two points provide the proper perspective. First of all, the TV-as-a-cause theory can only apply to the more discretionary, spare-time kinds of social capital; nobody seriously contends that people are turning down marriage proposals or work hours so they can watch more television. Second, notice that the top lines in each figure represent college-schooled folks, the most rapidly expanding of the three educational levels. Even if all of the downward tilt in the lines is indeed people passing up meetings to free up prime-time, educational trends will exert significant upward pressure on association in America.