The Polls: Presidential Greatness as Seen in the Mass Public: An Extension and Application of the Simonton Model

The Polls: Presidential Greatness as Seen in the Mass Public: An Extension and Application of the Simonton Model JEFFREY E. COHEN Fordham University I raise two questions in this article. In light of the scandals of the Clinton years, have the standards used to rate presidents changed or not? Second, do experts and informed citizens rate presidents similarly, and do they rely on the same criteria in their ratings? I use a C-SPAN poll administered in 2000 to experts, and through the Internet to the citizenry, as the data to address these questions. Results find great temporal stability in how presidents are rated. Furthermore, in applying a predictive model developed by Simonton, I find stability in the factors that predict presidential greatness ratings. In particular, experts and informed citizens rate presidents similarly and use similar criteria. Substantively, the most important and consistent predictor of presidential greatness is the number of years that the president served in office. This finding brings us full circle to a question that motivates much scholarship on the presidency: why do presidents get reelected for a second term? Introduction While presidential greatness rankings may not objectively inform us about who the best presidents are, they do tell us much about the factors that raters employ when comparing presidents. To date, most attention has focused on expert ratings of presidents, beginning with Arthur Schlesinger, Sr. s 1948 survey of historians. Several surveys of the mass public, conducted mainly by the Gallup organization, have found that only the more recent and historically more important presidents (e.g., Washington, Lincoln, and Theodore and Franklin Roosevelt) appear on mass public ratings. The public fails to rank most other presidents, presumably because of lack of knowledge. Thus, it is not easy to compare how experts and the general mass public view presidents. In 2000, C-SPAN conducted two polls, one of experts and one of viewers who responded to an Internet survey. Importantly, the two sets of raters responded Jeffrey E. Cohen is professor of political science at Fordham University. His most recent book, Presidential Responsiveness and Public Policy-Making, received the 1998 Richard E. Neustadt Award of the Presidency Research Group of the American Political Science Association. Presidential Studies Quarterly 33, no. 4 (December) 2003 Center for the Study of the Presidency 913

914 PRESIDENTIAL STUDIES QUARTERLY / December 2003 to the same questions, allowing us to compare how experts and the informed public view presidents. Are the opinions of experts echoed in the informed mass public or not? More importantly, do the two sets of raters employ the same factors in rating presidents? Answers to these questions will give us a better sense of the importance of certain factors in our assessments of presidential performance. If we find that informed citizens rate the presidents the same way as experts do and employ similar criteria in deciding which presidents are great, then the greatness ratings may tell us something about our political culture and the standards by which our political culture assesses presidents. In the sections to follow, I review past research on presidential greatness, focusing most heavily on the work of social psychologist Dean Keith Simonton, who has systematically studied presidential greatness. Simonton s (1987; 1991; 1992; 1993; 2001) research aims to try to explain why some presidents are perceived as great and others not. He has found that a handful of factors about the presidents and the times in which they have served account for a large component of the variance in presidential greatness rankings. In this article, I use Simonton s model as a framework, extending it to the C-SPAN surveys, and ask whether his framework helps us understand how informed citizens, such as the C-SPAN respondents, assess presidents. In the next section I review past research on presidential greatness. Then I describe the C-SPAN presidential greatness polls and assess their reliability and validity. I compare the C-SPAN poll results with previous rankings, as well as compare how the experts and informed public rate presidents. An analysis of the C-SPAN ratings using Simonton s model follows. In the conclusion, I place the findings into perspective. Past Research on Presidential Greatness In 1948, historian Arthur Schlesinger, Sr. began the systematic study of presidential greatness when he polled 55 experts on American history, asking them to rank the presidents (Schlesinger 1948). Since then, presidency experts have been periodically repolled a dozen times or more. 1 Some critics have charged that these expert polls are biased because the ideological and/or partisan preferences of the experts affect how they rate the presidents and that the panels of experts tend to be filled with liberals and/or Democrats (see, for instance, Piereson 1997; Lindgren and Calabresi 2000). Analysis of the ratings of experts, however, has not found such a bias. Maranell (1970; Maranell and Dodder 1970) unearthed no correlation between the ideological or partisan leanings of individual raters and their presidential rankings (see also Murray and Blessing 1994). Simonton (1987; 2001), in exhaustive studies, has found that the many 1. The many presidential rankings include Bailey 1966; Holmes and Elder 1989; Kynerd 1971; Lindgren and Calabresi 2000; Maranell 1970; Maranell and Dodder 1970; Murray and Blessing 1994; Neal 1982; Pederson and McLaurin 1987; Piereson 1997; Ridings and McIver 1997; Schlesinger, Jr. 1962; 1996; 1997; and Spangler and House 1991. The C-SPAN, 2000 ratings, which are used in this article, are found at the website http://www.americanpresidents.org/survey. See Bose 2001 for a review and comparison of many of these presidential rankings.

Cohen / PRESIDENTIAL GREATNESS AS SEEN IN THE MASS PUBLIC 915 expert polls of presidential greatness all produce similar rankings. This leads him to contend that a firm consensus [exists] about the best and worst presidents in U.S. history (Simonton 2001, 294). As a consequence, Simonton argues that research efforts should focus on trying to understand the components and predictors of presidential greatness, where a healthy debate exists (see Bose 2001; Deluga 1997; 1998; Greenstein 2000; Kenney and Rice 1988; Landy and Milkis 2000; McCann 1992; 1995; Nice 1984; Simonton 1987; 1991; 1992; 1993; 2001). The debate essentially boils down to whether greatness is a function of the president s personal attributes, such as personality and character, or whether it is merely a function of the times in which a president served. Simonton s analysis finds that the times seem to matter more than personal presidential characteristics. His model consists of the following six attributes: (1) the number of years that the president served, (2) whether the president served during wartime, (3) whether the president was assassinated, (4) whether a major scandal occurred while the president served, (5) whether the president was a war hero, and (6) whether the president could be described as intellectually brilliant. Simonton s (2001) analysis of the Ridings-McIver (1997) ratings finds that character (scandal) was less important that more situational factors in understanding overall presidential greatness. One wonders, however, if this is still the case in the wake of Bill Clinton s scandalplagued administration. The Clinton case raises several questions. First, in light of Clinton s behavior, have experts adjusted the standards by which they judge presidents? If character has become more important to the experts because of Clinton s behavior (or the high level of media attention to his behavior), then perhaps the factors that affect people s ratings of presidents may not be as stable as Simonton suggests. Second, do experts and informed citizens use the same criteria in rating presidents? During the Lewinsky scandal and congressional impeachment period, a seeming gap existed between the opinions of experts, such as Washington insiders and journalists, and the mass public. The latter seemed less concerned about Clinton s affair with Monica Lewinsky than political elites and the intellectual class. Does that gap exist in presidential ratings data too? The C-SPAN polls give us an opportunity to address these questions. The C-SPAN Polls Thus far, presidential ratings have relied almost exclusively on the assessments of experts, mainly historians and biographers. The Gallup organization conducted several mass polls, but found the mass public can only rate recent presidents and great historical figures such as Washington, Lincoln, and the two Roosevelts. In 2000, C-SPAN overcame this problem by conducting an Internet poll in which viewers could answer the same questions about presidents that were posed to a panel of experts. Although the C-SPAN Internet respondents are not representative of the mass public in general, presumably being more informed than the general public, the fact that they responded to the same questions as experts allows us to compare how an informed public and experts see the presidents. The C-SPAN public poll was open to viewers for a ten-day period

916 PRESIDENTIAL STUDIES QUARTERLY / December 2003 in late December 1999. In all, 1145 people participated in that survey in addition to the 58 experts. Table 1 presents the overall or summary C-SPAN presidential greatness ratings of experts and Internet respondents for presidents Washington through Clinton. Both the experts and Internet respondents were asked to rate presidents overall and on ten dimensions. The ten C-SPAN dimensions are: Public persuasion Relations with Congress International relations Economic management Administrative skills Moral authority Crisis leadership Vision and agenda setting Performance within the context of the times Pursued equal justice for all In contrast, the 1997 Ridings-McIver poll asked raters to judge presidents on five dimensions: leadership qualities, accomplishment, political skill, appointments, and character and integrity. Clearly, there is much item overlap between the C-SPAN and Ridings-McIver polls, but the C-SPAN survey goes into more detail and is less ambiguous than the more general Ridings-McIver categories. Unlike other polls, which only ask raters to rank the presidents, C-SPAN asked both sets of raters to assess whether the president possessed the specific trait. Cumulating across raters creates a percentage (which can also be converted into a rank order). Previous studies only ranked the presidents. Often these rankings are treated as numerical values when being analyzed (e.g., Simonton, 1987; 2001). The C-SPAN scores, however, are already scaled variables. Table 2 presents the correlations between each dimension and the overall presidential rating. For the C-SPAN viewers, the correlations range from.80 to.99 and all are highly statistically significant. As Simonton (2001) found, character and integrity (here morality) fit less well than the other items, but still was strongly associated with overall assessments. We find a similar story when inspecting the correlations for the experts. Here the justice score correlates the weakest, but still strongly at.72. Again, all correlations are statistically significant at the.001 level or better. Simonton s (2001) analysis of the five Ridings-McIver aspects of greatness found that they scale on one dimension. I replicate Simonton s approach with the ten items for both the C-SPAN experts and viewers. In each case, factor analysis produced only one factor. For the viewers, this factor accounted for 92 percent of the variance in the individual variables and 94 percent for the experts (Table 3). Thus, echoing Simonton, there seems to be one underlying greatness dimension, with several or more aspects entering into a president s greatness score.

Cohen / PRESIDENTIAL GREATNESS AS SEEN IN THE MASS PUBLIC 917 TABLE 1 C-SPAN Expert and Viewer Overall Rankings of Presidents President Experts Viewers Abraham Lincoln 1 1 Franklin Roosevelt 2 4 George Washington 3 2 Theodore Roosevelt 4 3 Harry S. Truman 5 7 Woodrow Wilson 6 13 Thomas Jefferson 7 5 John F. Kennedy 8 12 Dwight D. Eisenhower 9 8 Lyndon Baines Johnson 10 19 Ronald Reagan 11 6 James K. Polk 12 17 Andrew Jackson 13 14 James Monroe 14 9 William McKinley 15 18 John Adams 16 11 Grover Cleveland 17 21 James Madison 18 10 John Quincy Adams 19 15 George Bush 20 16 Bill Clinton 21 36 Jimmy Carter 22 27 Gerald Ford 23 23 William Howard Taft 24 24 Richard Nixon 25 20 Rutherford B. Hayes 26 26 Calvin Coolidge 27 22 Zachary Taylor 28 25 James Garfield 29 28 Martin Van Buren 30 30 Benjamin Harrison 31 31 Chester Arthur 32 34 Ulysses S. Grant 33 29 Herbert Hoover 34 33 Millard Fillmore 35 37 John Tyler 36 32 William Henry Harrison 37 35 Warren G. Harding 38 40 Franklin Pierce 39 39 Andrew Johnson 40 38 James Buchanan 41 41 Source: C-SPAN; available from http://www.americanpresidents.org/ survey.

918 PRESIDENTIAL STUDIES QUARTERLY / December 2003 TABLE 2 Correlations between Rating Dimensions and Overall Presidential Greatness Ratings by Experts and C-SPAN Viewers Dimension Experts C-SPAN Viewers Public persuasion.91.91 Moral authority.86.80 Relations with Congress.86.90 Crisis leadership.97.96 International relations.88.91 Vision and agenda setting.96.97 Economic management.91.93 Performance within the context of the times.99.99 Administrative skills.90.96 Pursued equal justice for all.72.83 Source: C-SPAN; available from http://www.americanpresidents.org/ survey. TABLE 3 Factor Loadings of Rating Dimensions and Overall Presidential Greatness Ratings by Experts and C-SPAN Viewers (Principal Components Factor Analysis) Dimension Experts C-SPAN Viewers Public persuasion.90.90 Moral authority.85.79 Relations with Congress.85.89 Crisis leadership.97.96 International relations.88.90 Vision and agenda setting.96.97 Economic management.91.93 Performance within the context of the times.99.99 Administrative skills.90.96 Pursued equal justice for all.68.80 Source: C-SPAN; available from http://www.americanpresidents.org/ survey. Despite some difference in the loadings of particular items across the two factor analyses, their similarity is striking. In both cases, performance within the context of the times was the highest loading factor, followed closely by vision and agenda setting. For the experts, the factor loadings ranged from.68 to.99; for the viewers, the range was.79 to.99. Six items loaded at.9 or better for the experts; seven did so for the viewers. And while pursuit of equal justice was the weakest loading item for the experts (.68), it was the second lowest item for the viewers (.80), only a shade stronger than moral authority, which loaded at.79. And for the experts, moral authority tied for next to last at.85 with relations with Congress. Thus, not only do the experts and

Cohen / PRESIDENTIAL GREATNESS AS SEEN IN THE MASS PUBLIC 919 informed citizens appear to rate presidents similarly on the same criteria, but these results strongly echo Simonton s earlier analyses (1987; 2001). Predicting Presidential Ratings One of Simonton s innovations was to ask what factors account for presidential performance and greatness. After exhaustive research over nearly two decades, Simonton settled on a predictive statistical model. His model is composed of the six variables mentioned above. Simonton found that longer service, assassination, wartime tenure, war hero status, and intellectual brilliance enhance a president s greatness score. Scandals lower the score. While some important challenges to Simonton s model exist (e.g., Deluga 1997; 1998; Kenney and Rice 1988; McCann 1992; 1995; Nice 1984), none of the rival formulations statistically outperform his model (Simonton 1987; 1991; 1992; 1993; 2001). In fact, most of these alternatives appear to be either components of his fuller model (e.g., Kenney and Rice 1988; Nice 1984) or refinements of points that he has made (e.g., Deluga 1997; 1998; McCann 1992; 1995). Given the performance of the Simonton approach, we ask how well it holds up with these new data, especially with the C-SPAN viewers. In other words, does the Simonton model only apply to the greatness rankings of experts or does it also generalize to informed citizens? Table 4 presents results that employ Simonton s model for the C-SPAN experts and viewers. In general, the results are quite similar to Simonton s earlier effort. This model accounts for a healthy amount of the variance in the overall presidential greatness scores, TABLE 4 Impact of Simonton Variables on Overall Presidential Greatness Scores, C-SPAN Experts and Viewers Experts Viewers Variable b SE p* b SE p* Years 3.70.68.000 2.87.61.000 Hero 10.46 4.18.009 9.65 3.73.007 War 9.15 3.84.01 6.18 3.43.04 Assassination 16.65 5.55.003 10.08 4.96.025 Scandal -9.06 4.61.029-7.53 4.12.039 Brilliance 4.65 1.62.004 4.10 1.45.004 Constant 30.54 3.72.000 41.03 3.33.000 R 2 /Adj. R 2.73/.68.67/.61 F 15.13 11.44 Source: C-SPAN; available from http://www.americanpresidents.org/ survey. *p = one-tailed tests.

920 PRESIDENTIAL STUDIES QUARTERLY / December 2003 with adjusted R 2 s of.68 for experts and.61 for viewers. We are slightly better able to account for the ratings of experts than viewers, a finding that makes sense, given that we are unclear as to the level of historical knowledge of the viewers and the potential sampling biases in a self-administered poll. Still, it is remarkable how closely the two sets of evaluators agree on the variables that account for higher presidential greatness scores. In both cases, all six variables are significant at.01. The two sets of evaluators apply similar weights to each of the six factors. For example, each year that a president served in office raises his score by 3.7 points for the experts and nearly 2.9 for the viewers. War hero status increases a president s score by 10.46 points for the experts and 9.65 for the viewers. Intellectual brilliance boosts a president s score by 4.65 points for the experts and 4.10 for the viewers. And scandal, one of the motivating factors for this article, was of broadly similar consequence to both panels: for the experts, a scandal lowered a president s score by 9.06 points, while for the viewers the drop was 7.53 points. One also observes a few differences in weighting across our two panels. Wartime tenure increases a president s score by 9.15 points for experts but is less weighty for viewers, at only 6.18. Perhaps the ambiguity of Vietnam has greater weight in the minds of viewers than experts, the latter perhaps possessing greater historical knowledge about other wars. And assassination weighs more heavily for experts than viewers (16.65 to 10.08 points). This may signal lower levels of information on assassinated presidents, like Garfield, who served abbreviated terms, but who some historians think of highly, perhaps because of his pre-presidential career. More important than these differences in the weights of individual variables, however, is the overall resemblance across the two estimations. Experts and informed citizens view presidents similarly. Moreover, these results correspond well with Simonton s analyses (1987; 2001). We should be cautious in comparing these results to Simonton s analysis of the Ridings- McIver poll because the Ridings-McIver poll only used rankings, and here we are analyzing percentage ratings. Still, the results are remarkably alike. Like our results, Simonton found that all six of his variables are statistically significant and correctly signed. Most important, not only does Simonton s model hold on these new data, but his predictive model applies to informed viewers, as well as experts. The Simonton Model and Specific Presidential Traits Both the C-SPAN and the Ridings-McIver surveys also ask respondents to rate the presidents on specific dimensions. Earlier, we reported that the ten C-SPAN traits load on one dimension, although some of the traits load more heavily than others. Simonton (2001) similarly found a uni-dimensional structure to the five Ridings-McIver traits (leadership, accomplishments, political skill, appointments, and character/integrity). Moreover, he found some consistency in the ability of his model to statistically predict the five traits. All save the character/integrity trait displayed R 2 s from.67 to.79. His model could account for only 46 percent of the variance in character/integrity. For none of the five traits, however, did all six of his variables reach statistical significance, indi-

Cohen / PRESIDENTIAL GREATNESS AS SEEN IN THE MASS PUBLIC 921 cating that respondents distinguish among the five traits. Yet, years in office emerged as a significant predictor each time. I repeat the Simonton exercise by regressing each of the ten C-SPAN traits on Simonton s predictive model. Table 5 presents the results for the C-SPAN viewers and Table 6 for the C-SPAN experts. I reduce the mass of available information by only presenting the regression coefficients (b s) for variables that reach the.05 level of statistical significance. We can compare the impact (b) of the same variable across equations because each dependent variable is measured in percentage terms. TABLE 5 Impact of Simonton Variables on Presidential Greatness Ratings Dimensions, C-SPAN Viewers (Regression Coefficients at p.05) Dimension Years Hero War Assassination Scandal Brilliance Constant R 2 /Adj. R 2 Persuasion 3.90 16.18 19.78 4.85 35.36.77/.73 Moral Auth. 2.22 10.39-22.00 50.08.42/.32 Congress 2.46 7.98-8.20 41.13.50/.41 Crisis 3.38 17.33 11.01 12.82 4.35 36.05.69/.63 Inter. Rel. 2.55 7.74 4.64 44.62.55/.47 Vision 3.50 11.16 7.29 14.64 6.15 36.02.68/.62 Economy 2.74 9.74 3.16 38.92.61/.55 Context times 3.59 11.71 6.46-10.63 4.55 37.32.66/.60 Admin. 2.42-11.10 4.37 45.11.69/.64 Equal justice 1.92 4.23 45.68.46/.36 All entries are regression coefficients that are significant at the.05 level or better. Empty cells indicate that the regressor is insignificant. Source: C-SPAN; available from http://www.americanpresidents.org/survey. TABLE 6 Impact of Simonton Variables on Presidential Greatness Ratings Dimensions, C-SPAN Experts (Regression Coefficients at p.05) Dimension Years Hero War Assassination Scandal Brilliance Constant R 2 /Adj. R 2 Persuasion 2.90 9.97 17.38 2.32 40.81.73/.68 Moral Auth. 4.07 14.47 18.13-26.72 4.31 33.66.63/.57 Congress 3.00 14.43 32.72.48/.39 Crisis 4.28 17.63 15.50 20.91-11.87 4.86 24.94.71/.66 Inter. Rel. 3.52 8.78 3.93 36.89.57/.50 Vision 4.51 13.57 11.95 19.16-10.65 7.43 25.63.70/.65 Economy 3.10 9.68 12.78 4.53 30.98.63/.56 Context 4.71 12.68 9.49 19.98-13.19 4.74 25.49.72/.67 Admin. 2.69 10.22-14.73 4.60 38.49.67/.62 Justice 2.08 15.83 5.25 33.00.39/.28 All entries are regression coefficients that are significant at the.05 level or better. Empty cells indicate that the regressor is insignificant. Source: C-SPAN; available from http://www.americanpresidents.org/survey.

922 PRESIDENTIAL STUDIES QUARTERLY / December 2003 From this wealth of detail, we can make several comparisons. First, we can compare the results in general with Simonton s. Resulting R 2 s are similar to Simonton s. Most fall in the.57 to.73 range for the experts and.50 to.77 range for the viewers. Just as Simonton found, I was less able to predict the experts ratings of character (here termed moral authority), which produced an R 2 of only.42, similar to Simonton s R 2 of.46 for the character/integrity trait. However, I was better able to predict the viewers moral authority ratings, producing an R 2 of.63. In contrast, the six variable models produced R 2 s of only.39 for equal justice for all and.48 for relations with Congress. Because the five Ridings-McIver traits do not correspond exactly to the ten C- SPAN traits, we cannot directly make comparisons of the impact of specific variables. Thus, we turn our attention to the comparisons between the C-SPAN experts and viewers, who were asked to rate the presidents on the same ten traits. One remarkable consistency emerges from inspecting the two tables. For both experts and viewers, years in office is a significant predictor for each trait. This echoes Simonton s finding that years in office was a significant predictor for each of the five traits that he looked at. The second most common significant predictor was intellectual brilliance, which was significant nine of ten times for the experts and eight of ten for the viewers. Intellectual brilliance seemed not to matter for experts and viewers ratings of skill with Congress and viewers ratings of moral authority. Being assassinated is more consequential for experts than viewers, reaching significance eight times for the former group but only four times for the latter. The other factors reach statistical significance about the same amount for experts and viewers. In summation, our results are comparable to Simonton s. Generally, his sixvariable models account for from one half to three quarters of the variance, with years in office the most consistent predictor. The longer a president served in office, the higher his overall ratings and the higher the rating for each specific trait, no matter who is doing the rating, experts or viewers. Conclusion In this article, I utilized C-SPAN s presidential greatness survey of experts and viewers to address two questions. One, given the scandals of the Clinton presidency, have character traits become more important in assessing presidents? Two, do experts and the informed citizenry rate presidents the same way, using the same criteria? As to the first question, character seems no more important in this most recent survey of experts than it did in previous ones. Thus, there seems to be great temporal stability in how presidents are rated. Similarly, experts and informed citizens generally seem to rate presidents the same way using the same criteria. The most important substantive result of applying the Simonton model was the finding that years in office consistently emerged as a statistically significant predictor, echoing Simonton s (2001) previously reported finding. In the remainder of this conclusion, I want to address the meaning of this finding and its implications for the study of presidential greatness.

Cohen / PRESIDENTIAL GREATNESS AS SEEN IN THE MASS PUBLIC 923 The years in office variable basically distinguishes between presidents who served for one term versus those who were reelected and served a second term, except for presidents who died in office or left office (e.g., Nixon) before their terms expired, and of course, FDR, who was elected four times. From this, I am led to ask, why do presidents get a second term? Two possibilities present themselves. One, the public rewards presidents who have done a good job in office with a second term, or, the public retains in office presidents who they feel possess great leadership qualities. One must question the plausibility of the second possibility. First, one can make the case that several presidents of mediocre talents won reelection (e.g., Grant), while others of more demonstrable leadership talents (perhaps Carter), were defeated in their reelection bids. It seems more likely that the public rewards presidents with a second term when the president presided over good times as opposed to bad times and when the president appeared successful in dealing with problems as opposed to being incapable or making problems worse. Thus, borrowing from the presidential approval and elections literatures, when the economy is in good shape and the nation is at peace or if at war, appears to be succeeding in the war effort presidents will tend to be reelected. We must then ask: do presidents possess the policy instruments to produce good times? And if presidents possess such policy instruments, how important are their greatness qualities to selecting policy courses that produce good times? In others words, the literature on presidential greatness and the analysis presented here leads us to the question of the impact of personal attributes on presidential decision making, a foundation question of presidential studies. Thus, presidential greatness studies are not mere parlor games, but are rooted in the same concerns as presidential studies in general. References Bailey, Thomas A. 1966. Presidential greatness: The image and the man from George Washington to the present. New York: Appelton-Century. Bose, Meena. 2001. Presidential ratings: Lessons and liabilities. Paper presented at the 2001 American Political Science Association Meeting, San Francisco, CA, August 29-September 2, 2001. Deluga, R.J. 1997. Relationship among American presidential charismatic leadership, narcissism, and related performance. Leadership Quarterly 8: 51-65.. 1998. American presidential proactivity, charismatic leadership, and rated performance. Leadership Quarterly 9: 265-91. Greenstein, Fred I. 2000. The presidential difference: Leadership style from FDR to Clinton. Princeton, NJ: Princeton University Press. Holmes, Jack E., and Robert E. Elder. 1989. Our best and worst presidents: Some possible reasons for perceived presidential performance. Presidential Studies Quarterly 19: 529-57. Kenney, Patrick J., and Tom W. Rice. 1988. The contextual determination of presidential greatness. Presidential Studies Quarterly 18: 161-69. Kynerd, T. 1971. An analysis of presidential greatness and president rating. Southern Quarterly 9: 309-29. Landy, Marc, and Sidney M. Milkis. 2000. Presidential greatness. Lawrence, KS: University Press of Kansas. Lindgren, James, and Steven G. Calabresi. 2000. Ranking the presidents. Wall Street Journal, November 16, 2000, p. A26.

924 PRESIDENTIAL STUDIES QUARTERLY / December 2003 Maranell, Gary M. 1970. The evaluation of presidents: An extension of the Schlesinger polls. Journal of American History 57: 104-13. Maranell, Gary M., and R. Dodder. 1970. Political orientations and the evaluation of presidential prestige: A study of American historians. Social Science Quarterly 51: 415-21. McCann, S.J.H. 1992. Alternative formulas to predict the greatness of U.S. presidents: Personological, situational, and zeitgeist factors. Journal of Personality and Social Psychology 62: 469-79.. 1995. Presidential candidate age and Schlesinger s cycles of American history (1789-1992): When younger is better. Political Psychology 16: 749-55. Murray, Robert K., and Tim H. Blessing. 1994. Greatness in the White House: Rating the presidents from George Washington through Ronald Reagan. 2d updated ed. University Park, PA: Pennsylvania University Press. Neal, Steve. 1982. Our best and worst presidents. Chicago Tribune Magazine, January 10, pp. 8-13, 15, and 18. Nice, David C. 1984. The influence of war and party system aging on the ranking of presidents. Western Political Quarterly 37: 443-55. Pederson, William D., and Ann M. McLaurin, eds. 1987. The rating game in American politics. New York: Irving Publishers. Piereson, James. 1997. Historians and the Reagan legacy. The Weekly Standard, September 29, pp. 22-24. Ridings, William J., Jr., and Stuart M. McIver. 1997. Rating the presidents: A ranking of U.S. leaders, from the great and honorable to the dishonest and incompetent. Secaucus, NJ: Citadel Press. Schlesinger, Arthur M, Sr. 1948. Historians rate U.S. presidents. Life, November 1, 1948, pp. 65-74. Schlesinger, Arthur M., Jr. 1962. Our presidents: A rating by 75 historians. New York Times Magazine, July 29, 1962, pp. 40-43.. 1996. The ultimate approval rating. New York Times Magazine, December 15, 1996, pp. 46-51.. 1997. Rating the presidents: Washington to Clinton. Political Science Quarterly 112: 179-90. Simonton, Dean Keith. 1987. Why presidents succeed: A political psychology of leadership. New Haven, CT: Yale University Press.. 1991. Predicting presidential greatness: An alternative to the Kenney and Rice contextual index. Presidential Studies Quarterly 21: 301-05.. 1992. Presidential greatness and personality: A response to McCann. Journal of Personality and Social Psychology 63: 676-79.. 1993. Putting the best leaders in the White House: Personality, policy, and performance. Political Psychology 14: 537-48.. 2001. Predicting presidential performance in the United States: Equation replication on recent survey results. Journal of Social Psychology 141: 293-307. Spangler, W.D., and R.J. House. 1991. Presidential effectiveness and the leadership motive profile. Journal of Personality and Social Psychology 60: 439-55.