Introduction

We so often hear of liberals and conservatives battling out the issues in the United States Congress that we may presume that these individuals oppose each other on most details. Given this presumption, our tendencies are to place liberals and conservatives at the poles of a one-dimensional continuum, as if liberalism and conservatism were diametric opposites. What if there were not a continuum and Congresspeople were classified both liberal and conservative? Would not this change the way in which we view Congresspeople and perhaps even ourselves? This possibility warrants further examination.

On grounds of reliability and validity we may critique the one-dimensional liberal/conservative continuum. A continuum is very difficult to apply to the U. S. Congress, because placing Congresspeople on the endpoints of a continuum is unreliable and invalid. Simply stated, no liberal/conservative continuum describes the United States Congress well.

My claim that no liberal/conservative continuum can describe Congress may arouse skepticism. However, even if I persuade the reader that there is more than one-dimension to a Congressperson’s ideologies and issue structure(s) that will justify this effort.

Because there is no standard definition for liberalism or conservatism we can ask the following question. Who are liberals and what is conservatism? This problem complicates the use of a liberal/conservative continuum, as does the invalidity of this one-dimensional measurement. It is more reliable, thus aiding the validity, to use a multi-dimensional structure to measure liberalism and conservatism. Finally, the use of group scores and Conservative Coalition votes to establish a continuum may lead to invalid measurements. The overall, contention is that no single-dimensional liberal conservative continuum can be used to classify Congresspeople.

The foremost challenge that confronts the liberal/conservative continuum is definitional. Webster’s Collegiate Dictionary defines liberalism as "a political philosophy based on belief in progress, the essential goodness of the human race, and on the autonomy of the individual and standing for the protection of political and civil liberties." The same dictionary also defines conservatism as "a political philosophy based on tradition and social stability, stressing established institutions, and preferring gradual development to abrupt change." Essentially this dictionary suggests that liberals and conservatives differ on the grounds of government intervention, conduct of the individual, and rate of social change. This seems to be a fair and relatively unbiased interpretation, but this is not how the average person or Congressperson might define the same terms.

In 1981, U.S. News & World Report interviewed ten Congressmen, five were to define what a liberal was and five were to discuss what a conservative was. Their answers did not reflect what had been defined in the dictionary. Senator Paul Laxalt (Republican-Nevada) stated that,

Senator John Tower (Republican-Texas) also contributed that; "Conservatives have more faith in people than in government institutions, and would keep government out of issues that can be handled in the private sector." The contradiction here is significant. According to the dictionary definition, conservatives favor institutions and stable social conditions. Yet, Senator Tower contends that conservatives have little faith in government institutions.

The liberal contribution to this example came from Representative Ted Weiss (Democrat – New York).

The representative suggests that the government should provide substantial assistance to citizens, however this does not express economic freedom for the individual. These two definitions (Webster’s and Weiss) of liberalism clash on the grounds of self-reliance. Representative Weiss claims that self-reliance is not necessary because governmental aid should provide for many. Webster’s Collegiate Dictionary implies that liberalism is entirely based on the autonomy of the individual and the ability to be economically independent from government.

The point in using these examples is not to state that any source provides the correct and only definition of a term. Consider that "All definitions are to some extent arbitrary, subject to the judgement of the definition writer." The overall point is to help establish the idea that definitions of liberal and conservative vary widely. These two terms are associated with the endpoints of the liberal/conservative continuum, and are so varied and contradictory that defining the liberal/conservative continuum precisely is impossible. Thus it becomes implausible to even apply the liberal/conservative continuum to anyone including Congress.

The liberal/conservative continuum fails us because it measures issues in only one dimension. A single dimension measures at most one set of values. For example, an individual’s or in this case a Congressperson’s position on several social and economic issues is placed on a line indicating a degree of liberalism or conservatism. This would be successful if economic and social issues varied along a single dimension. Research has shown (Shafer and Claggett, Conover and Feldman, and Kerlinger) that liberals and conservatives are not polar opposites and that issue structures for liberals and conservatives must be measured in multiple dimensions. Suggesting more than one dimension undermines the validity of a one-dimensional continuum because measurements being taken are not measuring all that they purport to measure.

While many try to place liberals and conservatives on the same line, liberals and conservatives do not view issues from the same backgrounds or issue structures. For example, a person might feel that he or she is very liberal concerning social issues, like racial equality. However, the same person might be very conservative on policies affecting the economy. The individual cannot be labeled as a devout liberal or conservative, because the individual cannot place himself/herself on a single-dimension. Congresspeople also cannot be placed on a continuum because they don’t fall into the same category over every issue. As Conover and Feldman so eloquently stated, "In essence then, it is posited that liberals and conservatives view the political world not from different sides of the same coin, but rather, if you will, from the perspective of entirely different currencies."

In light of this research, it is easier to understand why a one-dimensional continuum, like Hill et al., is invalid. Kim Quaile Hill, Stephen Hanna and Sahar Shafqat recently published an article in which they reported their research findings. Hill et al., studied campaign issues and used this data to determine liberalism and conservatism for the 101^st Congress. This form (one-dimensional continuum) of assessing liberalism and conservatism does not measure all facets of a Congressperson’s ideology. It has been established that an individual can have both liberal and conservative ideologies, or at least liberal and conservative values. As a result a single dimension which only measures one ideological dimension per individual is invalid. It does not reflect what it purports to measure, which is an overall ideological assessment.

Conover and Feldman’s research shows that there tend to be relatively few points where liberals and conservatives can be evaluated together (Appendix A). In Appendix A, it can be observed that, while there are seven categories, only one provides a directly polar opposition between both liberalism and conservatism. A similar challenge leads to Shafer and Claggett to devise a two-dimensional measurement (Appendix B). What this figure shows is that, given multiple dimensions, people can more accurately define their liberalism and conservatism by not having to choose one over the other. This type of measurement allows for individuals to agree with both groups on various subjects.

Congress makes policy decisions that are not based on only one value structure. Congresspeople vote towards the policies that best represent their ideologies and their constituents. The crucial note here is that a Congressperson might only agree with part of a bill or resolution, but favors that over the alternative vote. Sometimes it is hard to understand why a Congressperson might vote for something that is not typical of his/her position on the one-dimensional liberal/conservative continuum on which others have placed them.

In short, multiple dimensions provide a more valid measurement of an individual or Congressperson’s ideology. This method will provide more accurate results, because it allows for people to have more than one inclination unlike a one-dimensional continuum. The research on multiple dimensions should make people consider that while the one-dimensional liberal/conservative continuum might have worked well in post-World War Two society, it is no longer accurate. However, times have changed. Congressional voting behavior is more complex today, because of influence from Political Action Committee’s, the White House, lobbyists, etc. As a result a multi-dimensional model of liberalism and conservatism is greatly needed. Just like a phone line that is disconnected after it is no longer needed, so too should the liberal/conservative continuum.

"A wealth of past scholarship has demonstrated that legislators’ personal ideology is highly associated with the ideological liberalism of the policy choices they make in public roll call votes in Congress." While this might be true, there are two problems that come up when ideologies are directly associated with roll call votes. The first problem that occurs involves the validity of roll call votes when used to measure ideologies. The second issue arrives when interest groups use the roll call votes to determine how many times a Congressperson voted for their position.

When roll call votes are used so often to measure ideologies it seems hard to believe that they are invalid, that is they don’t accurately measure what is supposedly being measured. Keith Poole and Howard Rosenthal have produced two main research articles on the patters of roll call voting. In 1991, Poole and Rosenthal establish that there are two main factors that directly affect congressional voting behaviors. "The extent of preference is expressed by a utility function. The closer an alternative is to the legislator’s ideal point, the greater the preference for alternative, and the higher the utility." This suggests that Congresspeople vote according to what fits their position best. We are not always positive that this is a direct correlation to ideology, and it is hard to separate the two. Preferences of Congresspeople might also vary with ideology. Another element that affects voting significantly is lobbying. "Thus the lobbying process-involving interest groups and the White House – can be seen as a set of efforts to alter the location of the cutting line on an issues." Here it can be seen that roll call voting is tainted. Lobbying affects the results of votes, making them unreliable. Therefore, roll call votes may not be good measurements of liberalism or conservatism for two reasons. They need not be direct reflections of ideologies and are subject to influence, like lobbying. The second reason, as Poole and Rosenthal suggest, is that a vote may be resultant of a utility function and that function may reflect party, district, re-election themes, and more.

Group ratings are invalid because they are not measuring what they purport to measure. A vote in favor of an interest group is not necessarily connected to support of that group or ideological alignment with that group(s). A group rating is achieved using roll call votes and for the before stated reasons there is a problem with that method.

A specific example is from Senator Patty Murray (Democrat – Washington). Murray’s group rating from the League of Conservation Voters, a liberal interest group, was 85. Her rating from the Chamber of Commerce of the United States, a conservative interest group, was 17 in 1996. Other ratings included a 90 from Americans for Democratic Action, a 0 from the American Conservative Union; the ACLU rated her with a 59 and the National Federation of Independent Business rate her 36. Combined, these ratings give no real indication of how liberal or conservative Murray is. The groups themselves range greatly in their degrees of liberalism and conservatism. Thus, it would be invalid to take the aggregation of these ratings and assign Murray to a point on the liberal/conservative continuum. Group ratings are not good tools to measure liberalism or conservatism when linked to the liberal/conservative continuum because they are invalid.

Group ratings are derived at from roll call votes. As previously noted this is an unreliable and invalid measure. The unreliability of roll call votes can be shown easily. Given the same bill or resolution to vote on repeatedly, two things can happen. Lobbying tactics can alter voting in each repeated test. Also, Congresspeople may go back and forth over which vote (yea or nay) best represents the ideal stance. As a result of changing opinions, roll call votes are unreliable gauges of ideology because of their inability to produce consistent results repeatedly. Given the validity and reliability problems with roll call votes, a better understanding of why Group Ratings can’t accurately be arrived at. Group ratings can measure what they set out to measure (see footnote #16), but should not be applied to the liberal/conservative continuum. Group ratings do not measure ideology so they are not useful gauges of liberal/conservative ideology.

Overall, the use of roll call voting and group ratings are unreliable and invalid. When used to establish a continuum roll call votes become invalid because they do not measure what they purport to measure. Roll call votes do not determine or measure liberal and/or conservative ideologies. Also, roll call votes are unreliable because they can change over repeated tests and thus are inconsistent. Group ratings may be considered invalid as well, because of the varied data. While this same data is reliable it still does not measure congressional ideology. However, every group is different and thus will return contrary measures of ideology. When joined to determine composite ideologies, it becomes highly invalid. They do not measure congressional liberalism or conservatism; rather they measure how dissimilar each group is.

Conservative Coalition votes are yet another method used to determine the number of liberals and conservatives in Congress. This measurement too is reliable producing consistent results, but is not a valid measure of liberalism or conservatism. If you look at appendix C the trend of Conservative Coalition suggests that there are relatively few liberals in Congress and that most people are conservative. For example, in 1988 there were 10 votes on which the Conservative Coalition appeared in the Senate and they won with 97%. This might be mistaken for an indication that most Senators and even Representatives are conservative. Using Conservative Coalition Votes to measure liberalism and conservatism, may tend to eliminate moderates, which turns the continuum into two poles. This measurement moves the center of the continuum so that a Congressperson is extremely liberal or conservative. This displacement of the center does not allow for mixed values. Acknowledging this proves that Conservative Coalition Votes do not measure congressional liberalism or conservatism well. Thus, if used to establish a continuum, it is invalid. Conservative Coalition Votes is a flawed measurement and cannot depict the accurate liberal and conservative ideologies of Congresspeople.

While Conservative Coalition voting is an interesting phenomenon, it should not be used as the sole indicator of ideology. When this measurement is applied to the liberal/conservative continuum, it greatly skews an already flawed method. The usage of Conservative Coalition votes is highly invalid and should not be used to show how liberal or conservative a Congressperson is. This example is to show yet another method that is taken advantage of to create a single dimension continuum that is truly inapplicable to Congresspeople.

If the liberal/conservative continuum is invalid and unreliable, why do people continue to use it? The one-dimensional continuum is used to quickly sum up a Congressman/woman. Rather than take the time necessary to measure the entire ideology of a Congressperson, most, like the mass media opt for a quick solution. Yet another reason that the continuum is still in use is because those who use it resist change. Most have used the one-dimensional continuum for so long that they refuse to accept any other possibility.

Individuals might contest some to the assertions regarding group ratings and roll call votes that were mentioned previously. The two are invalid and unreliable determinants of liberalism and conservatism. Everyone is entitled to his or her own interpretations, but the contention of this paper still stands. A single dimension in and of itself is not a good measure of congressional ideology. Thus, group ratings and roll call votes, while consistent, are not the proper data to be used when making assumptions about congressional ideology.

Through various points this paper has shown that the liberal/conservative continuum cannot be applied to Congress. This is because of several validity and reliability issues. The issues stem from methods of measuring liberalism and conservatism. There are also problems with definitions, which directly affect measurement and above all liberals, and conservative’s ideologies cannot be measured in a single dimension. The value structures of Congressmen indicate that there are multiple dimensions. Hence this paper arrives at the conclusion that the liberal/conservative continuum cannot be applied because it is not a valid or reliable measurement of congressional ideologies.

Appendix A

Appendix B

Appendix C

Barone, Michael and Grant Ujifusa. The Almanac of American Politics: 1998.

Washington, D.C.: National Journal, 1997.

Carmines, Edward G. and Richard A. Zellar. Reliability and Validity Assessment.

Beverly Hills, CA: Sage Publications, 1979, 11-12.

Congressional Quarterly. The 53^rd Annual CQ Almanac: 1997, 105^th Congress, 1^st

Session. Washington, D.C.: Congressional Quarterly Inc., 1998.

Conover, Pamela Johnston and Stanley Feldman. "The Origins and Meaning of

Duncan, Philip D. and Christine C. Lawrence. Congressional Quarterly’s Politics in

America: 1996, the 105^th Congress. Washington, D.C.: CQ Press, 1995.

Duncan, Philip D. and Christine C. Lawrence. Congressional Quarterly’s Politics in

America: 1998, the 105^th Congress. Washington, D.C.: CQ Press, 1997.

Hill, Kim Quaile, Stephen Hanna, and Sahar Shafqat. "The Liberal-Conservative

Kerlinger, Fred N. Liberalism and Conservatism. Hillsdale, N.J.: Lawrence Erlbaum,

1984.

Merriam-Webster’s Collegiate Dictionary, 10^th ed. Springfield, Mass.: Merriam-

Webster, Inc., 1996.

Poole, Keith T. and Howard Rosenthal. "A Spatial Model for Legislative Roll Call

Analysis." American Journal of Political Science February 1985.

Poole, Keith T. and Howard Rosenthal. "Patterns of Congressional Voting." American

Journal of Political Science February 1991.

Shafer, Byron E. and William J. M. Claggett. The Two Majorities: The Issue Context of

Modern Politics. Baltimore: The John Hopkins University Press, 1995.

Stanley, Harold W. and Richard G. Niemi. Vital Statistics of American Politics: 2^nd

Edition. Washington, D.C.: CQ Press, 1990.

Stone, Marvin. "‘Liberal’ and ‘Conservative’." U.S. News and World Report 2

November 1981, 96.