racial attitudes as of 1965 or those articulated down the road. It is whether or not one was
surrounded by peers who felt this way that mattered.
Looking across time, influence from the local climate is still evident decades later as the
erstwhile high school senior ages. The coefficient drops from .41 to .25 between 1965 and 1973,
and is .18 as of 1997. Notice that it dips down to a statistically insignificant .05 in 1982, during a
time of heightened resistance to school integration and the government’s role in helping
minorities. Inferentially, a strong secular tide was undercutting the residues of school context
effects. Parental influence, less potent from the beginning, has lost its trace by 1997.
Turning to model III and the question of reinforcement effects, we observe some
evidence that racial attitudes reinforced by parents and school climate in 1965 were more likely to
be carried forth to 1973. Both coefficients are substantively significant (.09 for parents, .18 for
school climate), though the school climate effect is of marginal statistical significance (t=1.63).
These reinforcement effects are not evident past 1973, unlike what we found for party
identification. The baseline stability levels are much lower for racial attitudes than for party
identification as well; the coefficient on lagged opinion for the 1973 model was .37 in the case of
party identification and is .15 in the case of racial attitudes. Some of this greater instability might
be tied to differences in measurement error across the variables. But it surely also reflects the
true volatility of racial attitudes in the late 1960s and early 1970s for those coming of age in that
interim, a period which saw more race rioting, the growth of the black power movement, the
assassination of Martin Luther King, and the emergence of the Democratic party as the party most
advancing black interests.
Political Tolerance
As with racial issues, the extent to which the high school seniors expressed tolerant
attitudes was more closely tied to the climate of opinion within their schools than the attitudes
expressed by their parents.
In 1965, the coefficients were .36 (t=4.13) and .05 (t=1.74)
10
The parent tolerance variable is formed from only two of the three component variables used when
forming the youth and school climate indicators, as discussed earlier. This, however, does not account for
the relatively weaker association observed. The same pattern holds when we use indicators formed from
17
Convention