6—
Age, Education, and Dealignment

Our analysis up to this point has been largely static. What we found about 1988 or 1976 was equally true of the 1950s. Nothing changed very much, except that Pure Independents have even less civic virtue than they did a generation ago. Having made our case that leaners are essentially partisans who have nothing in common with the neutral Pure Independents, we turn now to describing the people most responsible for the growing numbers of Independents, a trend that we will call dealignment without prejudice to our argument about the leaners' partisanship.

We begin with the familiar and plausible proposition that independence, always more common among the young, was particularly appealing to those who reached voting age at the time of the Vietnam War. Was dealignment confined to the young, or did it also reflect abandonment of party identifications by some older Americans? A second question concerns the persistence of these independent identifications. Are the baby boomers becoming more partisan as they reach middle age? In other words, how general was dealignment, and how durable?

A third line of inquiry concerns the next generation, to whom Vietnam is history and Eugene McCarthy as remote as Joseph McCarthy. Are they as disinclined to identify with a

party as their immediate predecessors, or do they have the more modest inclination to independence of those who first could vote before the death of President Kennedy?

This is not unexplored territory, but as we will see, differentiating leaners from Pure Independents makes possible a clearer description of trends in party identification in the past twenty-five years. The distinction is central to our analysis of education, the concluding topic of this chapter and the only other demographic variable helpful in understanding dealignment.

Age and Partisanship

Most scholars agree that the increase in Independents was a reaction to "shocks" that seemed sharpest in the decade that began in the mid-1960s: the failure of political institutions, including the two parties, to achieve satisfactory solutions to the country's most salient problems, the Vietnam War and racial conflict.^[1] We will examine these interpretations directly in the next chapter. Here we concentrate on those citizens who seemed most susceptible to such shocks. Young people are widely believed to be more politically labile. Whether because they are more impressionable than their elders or because their political "anchors" are weaker, their reaction to dramatic political events is sharper and more durable than that of older citizens. It seems reasonable, then, to expect that young people would be found "at the cutting edge" of dealignment, that is, more inclined to be Independents.^[2]

[1] The most meticulous tracing of declining party identification is in Philip E. Converse, The Dynamics of Party Support (Beverly Hills, Calif.: Sage, 1976). He found the first marked drop in "mean party identification strength" occurred between 1964 and 1966; a second from 1971 through 1974. Converse did not measure an increase in Independents as such, however. His measure of party identification strength assigned a score of 3 to strong partisans, 2 to weak partisans, 1 to leaners, and 0 to Pure Independents.

[2] M. Kent Jennings and Gregory B. Markus, "Partisan Orientations over the Long Haul: Results from the Three-Wave Political Socialization Panel Study," American Political Science Review 78 (December 1984): 1015.

The American Voter reported that even in the 1950s "young people, just entering the electorate, are more likely than any of the older age groups to call themselves Independents."^[3] If Independents were progressively scarcer in older age groups, this might be a life-cycle effect : "party identifications strengthen with age."^[4] The mechanism is thought not to be age as such; rather, age is a surrogate for the number of years an individual has been affiliated with a party. "Perhaps older citizens are more staunchly partisan because the repeated use of party labels to interpret and understand the political world through the years reinforces and strengthens one's partisan ties."^[5] This interpretation "fits very well a more general thesis that group identification is a function of the proportion of a person's life he has been associated with the group."^[6]

But any analysis of age and partisanship must account for conflicting interpretations of a relationship observed at a single point in time or over a brief period. If older respondents in the 1950s were likelier to be strong partisans, this might reflect not so much their place in the life cycle as the fact that their political perspectives were formed in an era when partisanship was more enthusiastic, widespread, and unquestioned. This is a generation effect , defined by Philip Converse as "a kind of early imprinting which . . . will always leave its characteristic mark" distinguishing the affected age cohort from other generations not so affected by the imprinting

[3] Angus Campbell et al., The American Voter (New York: John Wiley and Sons, 1960), 161. This seemed to be the accepted position, but there were dissenters. See William H. Flanigan and Nancy H. Zingale, Political Behavior of the American Electorate , 4th ed. (Boston: Allyn and Bacon, 1979), 68.

[4] Philip E. Converse, "The Concept of a Normal Vote," in Angus Campbell et al., Elections and the Political Order (New York: John Wiley and Sons, 1966), 18n.

[5] Gregory B. Markus, "Dynamic Modeling of Cohort Change: The Case of Political Partisanship," American Journal of Political Science 27 (November 1983): 721.

[6] Campbell et al., The American Voter , 163.

event or condition.^[7] The defining characteristics of a generation effect are its limitation to an age-defined group and its durability. A period effect , on the other hand, makes its mark irrespective of age.

These formulations are somewhat problematic. No matter how much one may believe in the susceptibility of the young, it is difficult to imagine an event that would imprint one generation for life without leaving a discernible mark on at least some older people. Moreover, the effect of a gripping event may linger for only a few years before the familiar life cycle of strengthening partisanship resumes. If so, one must be chary about judging any relationship to be evidence of a generation effect. One might also discern an apparent period effect limited to people beyond the first flush of youth, but still short of middle age. As we will see, these caveats are illustrated by evidence on the party identification of different age groups from the 1960s through 1990.

We begin by examining the proportions of various age groups in each of the three categories of independent identifi-

[7] Converse, The Dynamics of Party Support , 74. For evidence of a strong life–cycle effect in the "steady–state period" from 1952 to 1964, see ibid ., chap. 3. A number of studies have shown that for at least a decade, beginning in the mid–1960s, party identification did not strengthen as people grew older. See ibid.; Paul R. Abramson, "Generational Change and the Decline of Party Identification in America: 1952–1974," American Political Science Review 70 (June 1976): 469–78, "Generational Replacement and Partisan Dealignment in Britain and the United States," British Journal of Political Science 8 (July 1978), and "Developing Party Identification: A Further Examination of Life-Cycle, Generational, and Period Effects," American Journal of Political Science 23 (February 1979): 78–96. See also M. Kent Jennings and Richard G. Niemi, "The Persistence of Political Orientations: An Over-Time Analysis of Two Generations," British Journal of Political Science 8 (July 1978): 333–63; Norman H. Nie, Sidney Verba, and John R. Petrocik, The Changing American Voter , enlarged ed. (Cambridge: Harvard University Press, 1979), 59–73.

These studies seldom distinguished among types of Independents. The exceptions are Abramson and Converse, who used the coding scheme described in note 1.

cation. These data for white respondents in each presidential election year from 1952 through 1988 are arrayed in table 6.1. The only one of these groups that is truly neutral between the parties—Pure Independents—had no special youth appeal in the 1950s. Not until 1964 were people under the age of 29 significantly more likely than the entire population to be Pure Independent. In 1968 a monotonic relationship between youth and pure independence emerged. Since then, the facts have matched the conventional wisdom that young people are especially disposed to independence and older people are not.

Much the same is true of Independent Democrats, although here reality matched the stereotype more readily. In 1952 everyone up to the age of 65 was equally likely to be an Independent Democrat. This category did not have any special youth appeal until 1956, when a consistent pattern began. Consistency is the last term one would apply to the relationship between youth and Independent Republicans. They are the only group to display a satisfying monotonic relationship in the 1950s, but this disappeared in 1960 and 1964. For a dozen years thereafter, the Independent Republican identity also seemed to appeal more to young people. In the 1980s, the only age group notably less attracted to this label was people 65 and over.

By 1968, the three varieties of independent identification were consistently most popular among people in their twenties. What is more, increasing proportions of youngsters reaching voting age were eschewing outright partisanship in favor of one or another sort of independent status.

Did this trend continue unabated through the 1970s and 1980s? Was the popularity of independence limited to the youngest set, or did older citizens come to share it? And what happened to the children of the sixties as they entered the second decade of their political maturity? Did they continue the disavowal of partisanship with which they entered political life? To answer these questions, we have made use of not only the simple cross-tabulations in table 6.1, but also analysis of eight age cohorts, which let us track party identification

in each age group as it moved through the life cycle. In order to reduce sampling error by accumulating larger numbers of cases, we used eight-year cohorts, the youngest of which in 1960 were those respondents who had not been old enough to vote in 1952. We began with 1960 because it preceded the "shocks" to party identification described by Converse; 1968, of course, was the year of peak political tumult. Because of the ratification of the 26th Amendment in 1971, the group too young in 1976 to have voted eight years earlier had an age span of eleven years and our youngest cohort in 1984 had an upper age limit of 25 rather than 28. The availability of 1990 NES data shortly before this book went to press let us take one more look at the earlier cohorts and also at a new cohort of those aged 18 to 23. Our oldest cohort takes in everyone over 74 in 1990. The cohort data are presented in tables 6.2, 6.3, and 6.4.

All three varieties of independence increased, in a different pattern for each category. The proportion of Pure Independents grew most among young people, but we did find a limited amount of change among some older cohorts. There was no change among people who were at least 45 in 1960 (Cohort VII) and a four-percentage-point increase in Pure Independents from 1960 to 1968 in the next two cohorts. Thereafter, the only shift in these three cohorts was a gradual decline in their proportion of Pure Independents. The oldest group in which party identification declined sharply was Cohort IV, who came of age in the Eisenhower era and weathered the Johnson administration without gaining Pure Independents. But in 1976, when Cohort IV was 37 to 44, approaching middle age, its proportion of Pure Independents nearly doubled, to 19 percent. Just 8 percent of them were Pure Independents in 1984 and 11 percent in 1990, which is about where they were in 1960 and, for that matter, where everyone 45 and older was in 1960.

Among the prime candidates for disillusionment are those who reached political maturity in the 1960s, our Cohort III. Eighteen percent were Pure Independents in 1968, at least

half again as many as any other cohort in 1968. Cohort III were at the same place in 1976, when 19 percent of them eschewed any partisan affiliation. By 1990 this dropped to 10 percent, the norm for everyone but the youngest cohort thirty years earlier.

In 1976 the next generation, Cohort II, had spent their entire brief political lives in a time of disorder, discontent, scandal, and apparent bipartisan failure. Twenty-two percent of them were Pure Independents in 1976, the highest proportion for any cohort in any year. By 1990 this figure had dropped by half, to 11 percent, the same as the two preceding cohorts in 1990.

Next is Cohort I, those aged 18 to 25 in 1984, whose political perspectives were largely developed in the era after Vietnam and Watergate. Fourteen percent of them were Pure Independents in 1984. This is substantially lower than the reading for the first appearance of the two previous cohorts and is a statistically insignificant two percentage points higher than the proportion of Pure Independents in Cohort IV at its debut in 1960. Six years later Cohort I's share of Pure Independents rose from 14 to 16 percent. This may be nothing more than a blip, which we consider more plausible than taking it as a portent of another round of dealignment.

Finally there is the truncated Cohort O, aged 18 to 23, who made their debut in 1990. Their relative youth may explain the 18 percent of this group who were Pure Independents.

In 1976 the three youngest cohorts had about the same one-fifth proportion of true partisan neutrals. In view of their age range, from 18 to 44, we consider this evidence of a limited period effect that stilled whatever relationship between age and increasing partisanship would usually be associated with movement through the life cycle. The three cohorts were also fairly similar eight years later, but at a sharply diminished level of independence. And by 1990 Cohorts II–IV had virtually the same proportions of Pure Independents as their counterpart age groups in 1960. In other words, after its suspension in the time of troubles, the familiar life cycle of strengthening partisan identification seems to have resumed in the (temporarily at least) more reassuring atmosphere of the Reagan administration. With the advantage of more time for hindsight, we can see that Paul Abramson was premature in concluding that the Vietnam-Watergate cohorts were fated to remain unusually withdrawn from partisan commitment:

The overall weight of evidence for nearly a quarter of a century casts doubts on the life-cycle formulation. Instead, the overall results show considerable stability in partisan strength within cohorts as they age, suggesting that once formative socialization occurs persons tend to

retain their level of partisan strength as they move through the life cycle.^[8]

Table 6.2 also displays the sharp differences betwen Cohorts II–IV and older citizens, those who were at least 29 in 1960 and therefore were well into their thirties when the time of troubles began. This generation gap can be seen most clearly in the 1976 column. In that year, one in five respondents under 45 were Pure Independents, double the proportion in the rest of the population. The proportion of Pure Independents among Cohorts V, VI, and VII was scarcely greater in 1976 than in 1960; differences for single cohorts observed in table 6.2 are below the level of statistical significance.

This conclusion is important enough to warrant repetition: There was no increase in partisan neutrality among Americans who had reached their forties before the nation was convulsed by Vietnam and racial conflict. The increase in Pure Independents was concentrated entirely among those who attained voting age in the 1960s and 1970s. In this conclusion we differ from Converse, who reported that strength of partisanship declined among the elderly and middle-aged as well as the young; and Warren Miller, who saw the disruptions and frustrations of the era creating "a period effect felt throughout the electorate."^[9]

[8] Abramson, "Developing Party Identification," 91. Later studies showing a revival of partisanship include Warren E. Miller and J. Merrill Shanks, "Policy Directions and Presidential Leadership: Alternative Interpretations of the 1980 Presidential Election," British Journal of Political Science 12 (July 1982): 308–9; and J. Merrill Shanks and Warren E. Miller, "Policy Direction and Performance Evaluation: Complementary Explanations of the Reagan Elections," British Journal of Political Science 20 (April 1990): 160.

[9] Converse, The Dynamics of Party Support , 95–97. In 1968 Converse's three cohorts were aged 25–40, 41–56, and 57 and older. Miller, "The Electorate's View of the Parties," in The Parties Respond , ed. L. Sandy Maisel (Boulder, Colo.: Westview Press, 1990), 103. Miller found particularly significant contributions to dealignment by the young. His findings were based on the ratio of strong partisans to Pure Independents among four-year cohorts. Using a differ-ent measure, Miller and Shanks reported a decline in partisanship among people who came of age as early as the 1940s; see "Policy Directions and Presidential Leadership," 308.

In contrast to Pure Independents, who continued to gain among young people in the 1970s, Democratic leaners reached their peak with this constituency in 1968. They lost ground by 1976 and continued to do so in the 1980s. In 1968, 20 percent of all whites under 29 (Cohort III) were Independent Democrats. Their proportion of this cohort fell to 14 percent in 1976. It fell again to 9 percent in 1984 and then rose to 14 percent in 1990. Cohort II, who arrived on the scene in 1976, were 16 percent Independent Democratic in that year, 13 percent in 1984, and 14 percent in 1990. The proportions were similar for Cohort I in 1984 and 1990 and Cohort O in 1990.

Democratic leaners became more numerous in the 1970s not because young people were more attracted to this identity, nor as a result of a broadly-based shift in the same direction. Instead, the increased proportion of Democratic leaners reflected the baby boom, which generated many more young citizens, more likely than older people to be Independent Democrats, but not more so in the 1970s than previously.

In all older cohorts (IV–VII) the increase in Independent Democrats from 1960 to 1976 was barely perceptible (and short of statistical significance); thereafter, there was a trace of movement in the other direction. Even more than with Pure Independents, increases in Independent Democratic identification were confined to the younger crowd.

Republican leaners resemble the other two types of Independents in one respect: their increase was greatest among younger cohorts. (In saying this we acknowledge the mild swings of Cohort VI as well as the sharp increase among Cohort V. After registering no change from 1960 to 1984, this latter group showed 17 percent Independent Republicans in 1990, when most of its members were in their sixties.) Cohort IV, who came of age in the Eisenhower administration, were just 5 percent Independent Republicans in 1960. This proportion doubled by

1968. The next cohort was 16 percent Republican leaners in 1968, and the three succeeding cohorts displayed about the same proportion when they made their debuts in 1976, 1984, and 1990.

In contrast to the other Independents, however, Republican leaners did not become scarcer as the 1970s passed into history; a decline from 1984 to 1990 was notable only in Cohorts VI and III. And as we just observed, when the next three cohorts arrived, they included as many Republican leaners as had their counterparts in past years.

How can we summarize this welter of findings? First, it is

clear that most of the increase in all three types of Independents came from those who entered the electorate from the middle 1960s through the 1970s. Fully 54 percent of Cohort III, aged 21 through 28 in 1968, were one or another type of Independent in that year. The same was true of 52 percent of Cohort II in 1976. In addition, Cohort IV became somewhat more independent. The older cohorts seemed scarcely affected. There was no increase in Independents among people who were at least 30 in 1960, and only a modest amount among those eligible to vote in that year. This finding is consis-

tent with the proposition that longtime party identification builds resistance to short-term influences.

Second, it is equally clear that the cohorts largely responsible for the increase in Independents are becoming more explicitly partisan as they age. The familiar life-cycle effect of the steady-state period is reemerging. This trend is most complete with respect to Pure Independents, who in 1990 were a much smaller proportion of Cohorts II–IV than in 1968 or 1976.

Third, youngsters reaching voting age after 1976, Cohorts I and O, were less averse to identifying with a party than their counterparts in 1968 and 1976, but still were a bit more inclined to be Independents than were Cohort IV, those arriving on the scene in 1960. It is too early to tell how much of this difference reflects their lower age. These new additions to the electorate, socialized primarily in the era of Carter and Reagan, contribute importantly to maintaining the proportion of Independents at a higher level than in the steady-state period. The other dynamic factor here, of course, is the passing of the oldest, quite partisan, generation.^[10]

What do these findings suggest about the future stability of the political system as the dealignment generation enters the age of maximum participation? The leaners are partisans. Thus what results from the distribution of partisan Independents as the population ages is not likely to have much to do with volatility, party responsibility, or similar issues. Wanting to feel independent, as leaners do, may well have psychological implications, but our concern, and the concern of most observers who discuss dealignment, is its political implications. In this regard, leaners do not differ substantially from outright partisan identifiers.

The shift toward Independents coincided with the entry into the electorate of the baby-boom generation and the enact-

[10] For similar findings see Shanks and Miller, "Policy Direction and Performance Evaluation," 160.

ment of the 26th Amendment. Those people most inclined to be Independents became a much larger share of the voting-age population. The extent of this phenomenon can be seen in table 6.5, which shows the proportion of people under the age of 29 in each type of Independent and the white population as a whole. In 1960, young people accounted for just 12 percent of all whites and anywhere from 7 to 18 percent of the three kinds of Independents. In 1972, young people were 25 percent of the entire white sample and 32 to 42 percent of Independents. By 1990 these proportions had fallen to 21 percent of the sample, 25 percent of the leaners, and 32 percent of the Pure Independents.

These data provide further reason for thinking dealignment is fading. Young people are a smaller proportion of the electorate, relatively fewer of them are without partisan affinities, and more people in the older dealignment cohorts are moving from independence to partisan identification.

Education and Independence

As we noted in chapter 3, the earliest conception of the Independent was an ideal citizen, one who could "give close and constant attention to public affairs."^[11] In large part, the rise of the Independent was "an opposition of intellectuals to political machines—and indeed, to parties as such."^[12] The view that Independents were educated citizens continued through the first half of the twentieth century. This image, developed largely from observations by scholars and pundits, was affirmed by early survey research. In 1948 George Gallup reported that "the higher the voter is in educational scale, the

[11] James Bryce, Modern Democracies , vol. 1 (New York: MacMillan, 1929), 47–48.

[12] Howard Penniman, The American Political Process (Princeton, N.J.: Van Nostrand, 1962), 39.

more likely he is to be independent."^[13] This conclusion was based on these figures:

The American Voter 's oft-cited low assessment of Independents' civic virtue included no reference to their educational attainment, except this remark: "There is no significant relationship between strength of party identification and formal education."^[14] A few years later Walter Dean Burnham's suggestion of a "new breed" of Independent was based on 1964 NES data showing that the "peak share of independents falls among high school graduates who have had some college education."^[15] Burnham concluded that the "political parties are progressively losing their hold upon the electorate" and that the losses "have largely been concentrated among precisely those strata in the population most likely to act through and in the political system out of proportion to their numbers."^[16]

Burnham was right to wonder about the apparent inconsistency between the 1964 data on Independents' education and the low levels of interest, knowledge, and participation attributed to them in The American Voter . But as we will see, his explanation of the anomaly and the implications he derived from it miss the mark. Nevertheless, well into the 1980s many journalists and scholars noted Independents' relatively high education, and some cited Burnham's new-breed hypothesis. Others thought that rising education contributed directly to a

[13] Cited in Dayton David McKean, Party and Pressure Politics (Boston: Houghton Mifflin, 1949), 183.

[14] Campbell et al., The American Voter , 479.

[15] Walter Dean Burnham, Critical Elections and the Mainsprings of American Politics (New York: W. W. Norton, 1970), 123.

[16] Ibid., 129, 130. For a more recent exposition of the new-versus-old thesis, see Frank Sorauf, Party Politics in America , 6th ed. (Boston: Little, Brown, 1984), 173–74.

putative loosening of party ties. These two passages exemplify many others:

The growth of Independents has come particularly in those persons with sufficient education to permit freedom from party cues.^[17]

Better educated Americans are less apt to affiliate with parties, which they perceive as lacking substance, and are more apt to identify themselves as free-thinking independents.^[18]

Implicit in these statements is a belief that education reduces the individual's need for the cognitive and perceptual road map provided by party identification.^[19] This notion is not timebound; it should be as applicable to the 1950s as to the contemporary scene. One simple way to explore this proposition is to examine relationships between respondents' education and their identification as Independents. Table 6.6 displays the percentages of people at various levels of schooling who were Independent Democrats, Pure Independents, and Independent Republicans in five of the ten presidential-election years from 1952 through 1988. These five years exemplify all ten.

Table 6.6 shows that the more educated people are, the lower the probability they will be Pure Independents and the higher the probability they will be leaners. In 1952, 7 percent of the least educated were Pure Independents as opposed to only 2 percent of college graduates. In 1988, the figures were 15 percent as opposed to 6 percent. The only exceptions were in 1960 and 1964, when there was no relationship between education and identification as a Pure Independent. The story is just the opposite for leaners. In 1952, 20 and 13 percent of

[17] Gerald M. Pomper, Voter's Choice (New York: Harper and Row, 1975), 34.

[18] Richard T. Saeger, American Government and Politics (New York: Scott, Foresman, 1982), 228.

[19] For a formal statement of this proposition, see W. Phillips Shively, "The Development of Party Identification among Adults: Exploration of a Functional Model," American Political Science Review 73 (December 1979): 1039–54.

college graduates were Independent Democrats and Republicans, respectively, as opposed to 10 and 6 percent of the least educated. This greater tendency for the better educated to be leaners is found throughout the period, with occasional modest exceptions that do not disturb the generalization.

Following trends from 1952 to the 1970s and 1980s reveals a considerable increase in the proportion of uneducated people who were Pure Independents, no change in their inclination to be Independent Democrats, and a modest rise in the percentage who were Independent Republicans. The picture for the college educated is very different. In the 1970s there was a modest increase in their inclination to be Pure Independents that reversed in the next decade. By 1988 the proportion of college-educated people who were Pure Independents had reverted to the level of the early 1960s. On the other hand, educated people in 1988 were much more likely to be leaners than their counterparts in the late 1950s and early 1960s.

Because middle-aged and older Americans did not contribute to dealignment, we examine, in table 6.7, the relationships between education and independent identification of respondents under 29 and those aged 29 to 44. These more focused cross-tabulations elaborate our earlier findings and

demonstrate the striking divergences in the party identification of young people according to their educational attainment. It is clear that the appeal of genuine partisan neutrality is inversely related to education. Pure Independents are most common at the lower end of the educational distribution and become progressively scarcer as the level of schooling in-

creases. On the other hand, Pure Independents are scarce among younger college graduates, who are inclined toward partisan independence.

These differences are most pronounced among young citizens at the opposite ends of the educational spectrum—college graduates and those who had not graduated from high school. This can be seen more easily in table 6.8 where we show the proportion of the least- and most-educated respondents under the age of 29 who were leaners and Pure Independents in each election year from 1968 to 1988. In 1968, fully 41 percent of the unschooled young were Pure Independents. The proportion subsequently fell to a little over a quarter, except in 1980 when 45 percent of all the uneducated young disavowed any partisan identity. In contrast, just 13 percent of young college graduates were Pure Independents in 1968. The proportion scarcely rose above this level for the next three elections and then dropped sharply in 1984 and 1988 to 5 and 4 percent.

The popularity of partisan independence has had a very

different pattern; 23 percent of the less educated were leaners in 1968, and 27 percent in 1972. This number scarcely changed thereafter. On the other hand, 44 percent of young college graduates were leaners in 1968 and 42 percent were in 1972. Thereafter, the number fell unevenly to just 24 percent in 1988. In that year, only 28 percent of the best-educated younger generation were any sort of Independent, in striking contrast to 1968 when this was true of 57 percent.

We now can understand Pure Independents' declining voting rate. A growing proportion of them are near the bottom on the two demographic variables that are strongly related to turnout—age and education.^[20] Controlling for these variables smoothes out the turnout variation among the seven party-identifier categories but does not erase Pure Independents' last-place ranking or leaners' relative civic virtue.

Other demographic differences between the varieties of Independents and outright partisans are without theoretical meaning or are a function of the age and education relationships we have just examined.^[21] In any event, these differences seldom are very substantial. As we would expect, poorer people are more likely to be Pure Independent, while the better-off in particular favor the Independent Republicans. None of the findings reported in this chapter is consequentially altered if all Southern respondents are removed from consideration.

Summary

We have found one thing that all three types of Independents have in common: All are most appealing to young people. This was most clearly the case from 1968 through 1976 when over half of all whites under the age of 29 were Independents. The effect of this development was enhanced by the baby boom and the 26th Amendment, which were responsible for

[20] Raymond E. Wolfinger and Steven J. Rosenstone, Who Votes ? (New Haven: Yale University Press, 1980), chaps. 2, 3.

[21] For example, Pure Independents are most likely to be found among "nontraditional" religious believers such as Mormons, Quakers, and Jehovah's Witnesses.

an unusually large addition to the electorate in the period when young people were most drawn to independence. By 1990 Cohorts II and III had scarcely more Pure Independents than their counterparts in 1960. One might say that the life cycle of strengthening party identification resumed after its apparent suspension. If this trend continues, the dealignment cohorts increasingly will resemble older citizens. This convergence argues against diagnosing their youthful independence as a generation effect in the sense of a permanent imprinting resulting from a profoundly moving event. New arrivals in the electorate since 1976 are neither as independent as their immediate predecessors nor as given to explicit partisanship as the oldest cohorts. At the other end of the age spectrum, Americans born before 1930 made no contribution to dealignment and those who reached voting age in the 1950s became notably more independent only by 1976. This in-between cohort then reverted to its pre-Vietnam pattern, except for an unusually large proportion of Republican leaners.

Our major theme about the fundamental differences between leaners and Pure Independents was revisited when we examined relationships between education and party identification. Better educated people were more likely to be leaners, less likely to be Pure Independents. These relationships strengthened over the years and among young people.

The tendency for most highly educated Independents to be leaners is not helpful to the notion that the high educational attainment of Independents suggests a considered and deliberate rejection of the American two-party system. As we saw, particularly in chapter 5, leaners show no signs of rising above the parties. And as we have seen and will see again, the only group that seems without affective ties to the parties, the only one without inhibitions about flocking to a new party, is the largely apathetic Pure Independents. These correspond to Burnham's "old independents" with respect to interest and participation, and in other respects seem unlikely to provide a constituency for either realignment or a new party. As we will see in the next chapter, Pure Independents are the least ideological and issue-oriented of the seven identifier types.