Framing the Message of Social Security in the U.S. Senate and the Media

I. Introduction

Do the size and electoral power of a state, the interest group rating, or the date of re-election influence the rhetoric of a U.S. Senator? This paper will seek to discover factors that indicate how often a senator falls in line with their specified party’s position. The party position involves the code words or frames used to sell or ‘spin’ a particular position. Both the Republicans and the Democrats have indirectly established sets of words used to structure their particular position on the issue of Social Security reform. These words act as frames to shape or a particular position. George Lakoff’s theory on framing acts as the bases for this Social Security study. His theory about the creation and repetition of frames by political parties is a major factor in the overall model. The repeated use of a party frames establishes a senator’s policy position to his fellow politicians as well as the public, the use of the opposition frames however supports the opposition (Lakoff, 23). Press releases and floor statements are one way in which a senator can communicate their position to the public and their constituents, as well as a means by which they can circumvent the warping lens or pen of the media. The rhetoric of Social Security reform reveals several challenges to party and message cohesiveness. The push for reform of Social Security comes down from The White House domestic agenda, which can lead to inter-party conflict, and second, the media’s holds the power to manipulate the tone and attitude of the debate.

This study of senatorial rhetoric and Social Security reform will progress through a background look into public conceptions associated with Social Security to the look at the role the and power struggle between the media and politicians. Ultimately, I will make projections about the relationships of a series of independent variables on U.S. Senators’ use of an established set of party frames. The second part of this paper looks at the relationship of a senator’s public statements and the frequency of those same frames in the print media the following day. Similar to a study by Pat Sellers on congressional media coverage, The Washington Post and Washington Times will serve as the liberal and conservative newspapers. Finally I will present the predicted values and regression analysis.

II. Literature Review

 The formation of a cohesive massage concerning the Social Security reform movement currently facing congress depends upon several factors. The willingness of a politician to vocalize his or her party’s message may or may not represent their opinions or even the opinions of their constituents. Social Security encompasses an uncommon host of complications due its nature as a comprehensive government program. This study looks into the variables driving senators’ public rhetoric; however, several factors related to the debate over Social Security prove prevalent to a discussion on Social Security reform. When asked about Social Security as a program and its possible reform, several studies show that the immediate perception of the program varies widely. Some of those polled relate the program to race, partly affiliation, the elderly, and others to the poor. These impressions prove prevalent when discussing Social Security because those who shape policy and try to implement reforms must consider these varying views so they may successfully sell the reform to the public.

Issue framing represents an important aspect of how the president, a party, or a senator expresses their position. It can also relate to an individuals perception of a politicians opinion. For example, “framing effects occur when different presentations of an issue generate different reactions among those who are exposed to that issue” (Jacoby, 2000). Jacoby’s study opperationalizes the manner in which a person reacts to the problem of government spending. Government spending proves gravely important when discussing Social Security because it represents a huge amount of government funds and it encompasses a wide variety of personal opinions. The relation of race to Social Security may not be an immediate association for everyone, but the works of Jacoby, Clawson, and Winter demonstrate its importance, especially when related to the issue of government spending.

Opinions on Social Security vary according to perceived beneficiaries, “racist feelings decrease support for government spending, while positive evaluations of welfare recipients and the poor both tend to increase it” (Jacoby, 2000). Conservative and white are the two most common associations for Social Security (Winter, 2003). How does a government program like Social Security gain not only a white face, but a party affiliation? Winter’s group theory explains the conditions under which an issue can become associated with race despite the lack of apparent racial connection as well as the connection to a party. The importance of the group theory extends beyond simply an association with race but to a framer’s ability to link policy to seemingly unrelated topics. In separate analysis, Clawson’s findings about the media back up Winter’s theory. First, media images of whites tend to be much more positive then those of blacks ( Clawson, 2003). Second, the media tend to focus more of their time and attention to the negative rather than the positive ( Clawson, 2003). The media’s ability to put a face on policy comes from their manipulation of the image. They provide the lens through which much of the public get exposed to public policy. Coverage of government programs tend to focused more on the problems associated with them rather than the successes. Winter’s group theory comes into play here. Clawson sites speeches by several conservatives referring to Social Security as a positive program, rewarding those who have worked hard all their lives. In addition, “Republicans frame Social Security in terms of equality to a much greater extent than do independents or Democrats” (Winter, 2003). Neither of these positive frames of equality and individuality typically applies to negatively framed blacks. Current republican frames for Social Security put an emphasis on stability and a safeguard for future generations along with the creation of a “nest egg” to support and reward the hard working individual. These frames according to the group theory, reveal the inherent whiteness of Social Security.

The Clawson experiment involves a mixture of several hundred participants, predominately college students. Each participant answers questions about government spending after receiving information on Social Security. The paper containing the facts about Social Security includes an image of either a black couple or a white couple. The results show, conservative participants favoring Social Security reform after viewing the white couple and less positive opinions about reform when presented with the black couple ( Clawson, 2003). The reason for the racial difference goes back to assertion that the motivation for public opinion reaction directly relates to the perceived beneficiaries (Jacoby, 2000). The perceived distinction of Social Security from other welfare programs exposes on the important link between the media and political framing. The ever changing and unpredictable lens of the media forces politicians to craft their own message in anticipation.

Politicians success and effectiveness at implementing their policy initiatives depends largely on their ability to build and maintain a cohesive coalition of support. Despite preconceived opinions about Social Security, what kinds of obstacles do politicians face when communicating their message and how do political scientists propose they deal with them? Coordination of media activity is vital for the senate given the decreased incentive for collective action (Sellers, 2002).

The feasibility of implementing wide sweeping reforms represents a battle for the President, Congress, the media, the evolving characteristics of the program, and the logistics of the policy campaign. To begin, the Republican National Committee (RNC) work as one when it comes to policy initiatives. The roles of the DNC and RNC have evolved from simply getting the president re-elected to helping him establish support for his initiative and build a cohesive foundation in congress. A study of media party image and the evening news in 2004 surprisingly reveals, “that it is extremely difficult for the presidential party in Congress to present a “unified front” with the president in the news” (Groeling, 2004). The president’s party in congress appears more likely to criticize him in the news, thus damaging his policy efforts more than opposition complaint. The president is currently spending a large amount of political capital on the Social Security reform initiative. Despite the likelihood for public critique, “the president’s partisans in Congress actually support their leader’s position about 75 percent of the time in their voting but only in about 40% of their evaluations in the national news” (Groeling, 2004). The battle waged across media outlets by politicians represents the ongoing campaign that now exists in modern politics. Despite close proximity in Washington, many politicians have begun to use the media to communicate their policy position.

Re-election in 2006 represents one important variable in determining a senator’s public rhetoric and adherence to party frames. The importance of the re-election campaign adds another complicating dynamic to the president and the party’s ability to sustain his message. The tools of a campaign have blended themselves into the governing and negotiating tactics of policymaking. Polls, town hall meetings, strategic public statements, and party driven coalitions, shot to make headlines and the top of the fold, just as they were for the campaign of 2004 (Ornstein and Mann, 2000). The media’s ability to guide any form of campaign symbolize their growing involvement in daily governing. As a result, politicians adjust to the requirements of the media and the need to appear “in-touch” with the public.

The evolution of the role of the instant news media forces politicians to focus an increasingly large amount of time on crafting a broad public opinion for fear of spending too much of their political capital on one initiative. The 24-hour news networks ensure that the average American gets their news from a media source. “By appealing to the media, one can attempt to indicates one’s preferences, respond to ongoing events, and attempt to persuade en masse en entire and desperate set of political actors across branches and levels to the correctness of one’s stance” (Cook, 124). The media, however, determines the events that receive attention and the lens through which those events reach the voting public. Politicians now seek to craft the image of their message to appeal either to a target audience or the masses. This new role of creating “crafted talk” serves to simplify a complicated policy issue for its use in the media (Jacobs and Shapiro, 27). The need for crafted talk does not necessarily symbolize pandering on the part of politicians, Jacobs and Shapiro suggests that politicians do not pander; instead, they manipulate their actions and message to appease public opinion. Clinton created focus groups to discuss health care reform when in fact their purpose was simply to give the impression of collective decision making (Jacobs and Shapiro, 90).

The management and anticipation of public opinion on an issue like Social Security has already proven to evoke a myriad of different emotions depending on the manner in which the media or politicians present their story. This study will not investigate racial implications associated with Social Security reform; however, any association the public adds to an issue be there a direct link or not, drives the need for politicians to anticipate a response. They must walk a fine line of responsiveness so as not to give the impression of pandering to the public. In their work, Politicians Don’t Pander, Jacobs and Shapiro portray Clinton’s use of focus groups to present the image of appealing to the public while in reality following his own policy goals. Social Security reform, like HealthCare, represents a huge barrier naturally generated by well-established programs, for which many depend upon and feel entitled to (Beland, 2004). Politicians, party’s, and interest groups all have something to say about the reforms, they all want to make a headline, but in the end they are all at the mercy of the reporters pen. Politicians do not pander to the whim of the public, but they do engage in year round campaigns to sell themselves and their agenda. The rhetoric they choose and the vehicle through which they convey their message represents a calculated move. The debate of social security supports Ornstein and Mann’s theory of the never-ending campaign as well as Jacobs and Shapiro’s discussion of crafted talk. The factors motivating public rhetoric on a controversial issue like Social Security and the willingness of a politician to fall in line with their party message depends on a host of variables. U.S. Senators are motivated by the electoral size of their state, the date of their next election, and interest group ratings. As for citizens, their “opinions will be shaped by politicians’ rhetoric, by the actions of interest groups and the champions of various causes, and by the way the mass media cover the unfolding story. Also relevant is how legislators anticipate that public opinion might evolve after a plan is approved and implemented” ( Arnold, 1998). The best guides for analyzing senatorial public rhetoric are the most traditional, floor speeches and press releases.

III. Hypothesis

The purpose of this study is to determine how a Senator’s re-election date, their previous voting record on Social Security along party lines as indicated by interest group ratings, and the electoral size of their state influence their rhetoric choices. A politician’s willingness to take a position on the issue of Social Security depends heavily on party momentum and message decisiveness. When a politician chooses to take a side, they likely take up the same rhetoric as their partisans. The Senate, unlike the House of Representatives, is a small governing body of experienced politicians. Their six-year terms provide time for members to take decisive positions on policy before moving towards the center in anticipation of an election. The rise of partisan politics corresponds to the decline in responsiveness by politicians, “the change in the structure of American politics…created incentives for politicians to pursue relatively extreme policy goals, discount centrist opinion, and compete with political rivals and journalists to direct public perception in order to minimize the risk of electoral retribution” (Jacobs and Shapiro, 151 ). As a result, senatorial rhetoric should likely reflect party frames. As discussed previously, the ability of a party to mount a successful campaign comes down to unity. Success requires message repetition (Lakoff, 23). The use of frames should reveal aspects of seemingly unrelated opinions. Strategically chosen frames manipulate the public’s association of Social Security to a positive or a negative attitude, as was seen with the link to race. The modern evolution of crafted language combined with the permanent campaign should result in a representative whose rhetoric depends on an intricate relationship of power, reelection, and policy. The set of established democratic and republican frames represent the dependant variable and while the independent variables fall under the afore mentioned categories. If a senator faces re-election in 2006, their use of party frames should be less than the mean of 5.1 for republicans and 11.4 for democrats. Secondly, the number of frames mentioned should rise in relation to the number of electoral votes per state. The third independent variable consists of interest group ratings about a senator’s stance on Social Security. The higher the rating given by a specifically chosen conservative interest group the greater the likelihood that senator will use a higher than average number of frames. The same notion of support scores follows for democrats and an interest group representing a liberal leaning constituency. Given the heated nature of the political debate on Social Security, its position as a massive, long-standing government program, and the argument discrepancies over its financial future, the unified stance of a party has never been more vital, or public.

The natural diffusion of power in a governing body like the Senate poses an ideal forum in which to study rhetoric. Steeped in the tradition of prolonged floor speeches, the senate’s sources for rhetorical data are voluminous. Senators have ability to garner a greater amount of publicity because media outlets typically see senators as greater authoritative figures than the majority of members in the house. A motivating factor behind the senate choice for the study combines the simplicity of one hundred units of analysis versus 450, and the greater likelihood that senator’s take more resolute positions on Social Security reform. The senate facilitates another interesting reason for investigation; republicans do not hold sixty percent of the chamber. Close margins in the chamber make the jockeying for power and position cohesion increasingly significant. Despite the Groeling theory, I expect Republicans to have a more cohesive and vocal message because its majority position in both chambers of the Congress and The White House.

The final conjecture of the study involves the portrayal of the republican and democratic frames in the news media. As noted above, the media prints what it pleases while politicians from the White House to Capital Hill jockey to have the massage they want above the fold. Politicians on both sides of the isle disagree about bias in the press. To analyze this complaint, a complied list of republican frames and democratic frames relating to social security will be tested against articles in the Washington Post and Washington Times from January through February of 2005. Two widely calculated papers in Washington with reported biases to the left and right. The heated and partisan debate over Social Security reform should be evident in the media bias of the storied in the Times and Post.

IV. Data

The collection of data involves a wide range of congressional and media resources. A senator present a single unit of analysis. Unless specifically designated, the data does not automatically bias the senators by their party affiliation. The set of ten democratic frames and ten republican frames act as the two dependent variables. The word choice for the frame sets came from articles and position statements on Social Security from the web sites of the DNC and the RNC as well as, articles from cnn.com, the New York Times, the Associated Press, and the White House website. Publications supporting the two positions all provide useful sources for codeword collection. Headlining interest groups like the AARP who speak decisively about Social Security reform are useful in confirming a party’s frames. The democratic frames consist of; “gambling,” “unnecessary,” “risk,” “scheme,” “privatization,” “debt,” “deficit,” “cutting benefits,” “dismantle,” and “irresponsible.” For example, press releases from the website attack the credibility of the president’s statistical data; they question his motivation, and the legitimacy of the “crises.” The press releases in February pay particular attention to the question of cutting benefits while news articles from CNN and the New York Times that lay out both sides of the story quote politicians and pundits paint images of risk and growing debt. The choice of spending attacks represents and effort by the democrats to link Social Security reform with the growing costs of the conflict in Iraq, tax cuts encouraged and implemented by the president, as well as the growing trade and budget deficits. Republicans use language connecting the Social Security reform to tax cuts and on going military conflict but they put a positive spin on these same issues.

The frames for the president and his partisans consist of, "crises," "save," secure," voluntary personal accounts," "modernize," "preserve," "problem with Social Security," "strengthen," "future/next generations," and "nest egg." The president links his Social Security agenda to the feelings of vulnerability after September 11 th through words like “secure,” “save,” and “crises.” Words like “preserve” and “nest egg” link to tax cuts by drawing attention to the individual. Winter’s assertion that the conservative party focuses attention to the individual appears valid in this list of frames. Like all framing efforts, these republican code words and those of the democrats try to evoke support or doubt for the Social Security reforms by drawing associations to seemingly unrelated issues. To ensure accuracy in data collection, each floor speech and press release are read and coded based on the above words. The manual approach to coding ensures the appropriate context of the frames and an accurate depiction of the Senator’s rhetorical intentions. Coding the press releases and floor statements involves the use of both republican and democratic frames in order to determine the frequency with which senator’s use opposition language. This creates four more independent variables for the third test in the media outlets.

The independent variables should reveal motivations and considerations made by senator’s when determining public rhetoric. The first independent variable involves whether or not a senator faces re-election in 2006. The choice of this variable seems especially relevant for the senate because of the likelihood that their public statements will move more toward the center as re-election approaches. The Senate website, www.seante.gov, provides the dates for senatorial re-election as well as access to the Congressional Record, which comprehensively records all remarks and actions on the floor of the Senate.

The second independent variable questions the importance of a senator’s electoral power. The electoral variable indicates the number of electoral votes for each state. The number of electoral votes mirrors the population of the state; therefore the variable symbolizes both the number of electoral votes and the population strength of the state. Like the question of election in 2006, the number of electoral votes represents a unique position in the Senate because the fundamental premise behind the governing body of the senate is that it tries to safeguard all states sovereign governing power. An interesting point of consideration in the final analysis involves the number of party frames from a large state like New York with Senator Clinton and 31 electoral votes versus Senator Thune from South Dakota with only 4 electoral votes. Both are prominent members who hold a great amount of name recognition. The third group of variables represents an outsider link to governing.

The Project Vote Smart website provides an extensive list of interest group ratings on a broad range of issues. The site sorts the interest groups by issue affiliation. The ratings indicate how the representative’s voting history matches up with the agenda supported by the group. For the purpose of impartiality, the National Committee to Preserve Social Security and Medicare serves as the democratic interest group rating. The group’s agenda supports the agenda set forward by the Democratic Party This organization represents an estimated 4million senior citizens and is strongly opposed to the president’s privatization plan. The Republican interest group, Retire Safe, supports the republican agenda. Many of the seniors groups, most notably the AARP oppose the president’s plan for reform. Retire Safe has supported a republican reform agenda. The organization questions the president’s plan but it does not criticize it. For the sake of comparison, Retire Safe represents 300,000 seniors while NCPSSM’s claims 4million. The internet facilitates many of the resources. The websites of senators provides their press releases and some floor statements while the Congressional Record logs all proceedings from the floor of the senate chamber.

The websites of many senators and the frequency with which they are updated varies greatly. Newly elected senators do not typically issue many press releases nor are their websites very extensive, some long time members also do not up date their sites for what appears to be many months. For example, Jim DeMint from South Carolina and Tom Coburn of Oklahoma fail to update their sites for 2005, even by the end of February. The logistics of accessing a representative’s website and the location of their releases depends mostly on their tenure. Prominent experienced members tend to have well-organized sites, those who are new and or less prominent present more of a challenge when attaining their remarks.

Due to these technical limitations, the statistical data on some senators may be un accounted for. The period for observation ranges from January 1, 2005 through the end of February 2005. This time marks the initial push made by the administration for Social Security reform. Conflicts related to the time frame involves a diminished volume of discussion and press releases due to the presidential inauguration and swearing in of new representatives. Much of the month of January was spent swearing in new representatives at the start of the month and the president towards the end of the month. Discussion over Social Security reform was also overshadowed by battles to confirm new presidential appointees, most notably, Condoleezza Rice as Secretary of State. The volume of releases and the need to take a decisive stand in the media diminishes significantly when the focus of daily activities revolves around confirmations and filibusters. The floor of the senate is a public stage for many especially around the time of the inauguration. It represents the best way for a senator to communicate directly with the public and the president.

The final part of the data collection requires the use of Lexis Nexis. This service catalogs all articles from newspapers all over the world. It files them according to date, region, and outlet. The ten republican and democratic frames can be plugged into the search engine and then the service will search all articles from each paper during the specified two-month period. Lexis Nexis makes the search in the media outlets significantly less burdensome than that of the senatorial press releases. The test aims to investigate the legitimacy of the claim that TheWashington Post prints on a liberal bias while the Washington Times prints more conservatively biased articles. Sellers assumes this same theory in his test of massage and its manipulation by senate leadership and the press (Sellers, 2002). The compilation of data for the independent variables, dependent variables, media outlets, and senatorial releases around the notion of power, policy, and reelection in the media should yield significant results.

V. Regression Models and Statistical Results

The use of the regression model is the best way to detect a relationship. Regression tests show the significance of a relationship. Specifically a regression test shows, “the relationship between the mean value of a random variable and the corresponding values of one or more independent variables” (Dicionary.com). The two dependent variables, democrat and republican frames are run in two different regression tests. The first regression tests the relationship between the democratic frames and the three independent variables. The second test is the exact same, save the use of republican frames at the dependent variable. The third, forth, fifth, and sixth regression tests involves the media outlets of the Washington Post and Washington Times as the dependent variables and the number of republican and democrat frames as referenced by both democrats and republican senators as the four independent variables.

The regression test produce highly correlated interest group ratings. The highly correlated variables make it necessary to use only one of the ratings. Due to the relationship, the interest group rating for republicans can be used without worry of compromising the test. The first test of democratic frames produced significant results for all variables.

Insert Table 1 Here

The democratic frames variable is not normally distributed. Its mean of 11.41 has a standard deviation of over 41. As a result, the predicted values for two of the variables within one standard deviation above and below the mean produce a negative result. This however, can be compensated for without compromising the results. Despite the change in the variable choice most significant variable influencing the use of democratic frames proved to be the interest group rating. Ratings by interest groups ranged from zero to 100, with a mean of 54.57%. According to the predicted value, as the percentage of approval increases from 13.12 %( one standard deviation below the mean) to 96.02% (one standard deviation above the mean) the number of frames falls from 25.59 to zero. The negative coefficient of -.316 demonstrates a negative relationship that corresponds with the predicted theory. Senators who have low ratings from conservative groups apply a greater volume of liberal language. The P> t value of .001 proves the variable to be very significant. The second significant variable, the electoral size of a state also produces a significantly positive relationship.

The second significant variable of electoral size demonstrates a relationship between the electoral power of a state and the number of frames used by a senator. The positive coefficient value of 1.34 indicates a positive relationship. The significance of the relationship, P> t .002 confirms its strength. The values for electoral size range from 3 for a state like North Dakota to 55 for California. When adjusted by one standard deviation in both the positive and negative directions the predicted values exhibit the same values for the interest group rating. As the size of a senator’s electoral power rises from 3 to 20 the number of democratic frames increases from zero to 25.28. Therefore, the larger the state the more likely a senator will use democratic frames. Unlike the interest group variable, electoral size does not relate to partisanship. Its relationship demonstrates an expected value but one that might anger the founding fathers. The third and final significant variable for democratic frames involves whether of not a senator faces re-election in 2006.

Part of the strength of the senate comes from the job security of a six-year term. Politicians have time and enough political capital to take decisive stands on issues without fear of devastating their political careers. The election year variable simply tests if the senator faces re-election during the next midterm elections in 2006. Despite a weak significance of P>t .046, the predicted values show that if a senator faces reelection in the near future, the number of frames they use falls from 20.38 to 4.59. The negative relationship, as indicated by the -16.8 coefficient value and the results match with the prediction that decisive rhetoric declines as elections approach. Unlike the previous variables, the range does not include zero. The most significant relationships include a range of 25 democratic frames.

The overall model produces an R-squared of .214. This value explains the percent variation in the dependent variable as explained by the overall model. 21.4% demonstrates a weak but legitimate relationship. The R-squared value reflects the relative strength of the overall model. The significance of the all three variables and the relative strength of the R-Squared, this democratic frame regression model is a successful model. The best understanding of the possible implications of the results comes from comparison with the second republican frames model.

The second regression model uses republican frames and the same three independent variables, interest group ratings, electoral size, and re-election in 2006. The predicted relationships for this dependent variable mirror that of the previous. The use of the republican interest group ratings should give a positive significant relationship and correspond to a rise in the number of republican frames. Like the prior test, the increasing electoral size should be consistent with a larger number of frames. Thirdly, the approaching midterm election should match up with a decline in decisive rhetoric.

Insert Table 2 Here

The outcome of the regression test reveals many unexpected results. The normal distribution of the republican frames dependent variable does not require a correction to the predicted values to account for negative results as was seen in the democratic frames model. Also unlike the previous model, only two independent variables demonstrate significance, the interest group rating and the electoral size. The p value of .065 for re-election discounts it from significance.

As predicted, the republican interest group rating displays significance, however both the P> t value of .023 and the coefficient value of -.053 do not match expectations. The weak p value indicates that the relationship between the level of support and the number of frames does not reach expectations. In addition, the negative coefficient value shows the relationship to be negative. As the rating for a senator climbs from 13.12% to 96.02% the predicted number of republican frames falls by exactly half, from7.62 to 3.31. Little evidence explains this negative relationship, save the Groeling theory of inverse partisan support. Groeling argues that partisans do not tend to use the president’s rhetoric nor do they statistically support him publicly on decisive issues, but ultimately they support his policy initiatives when they vote (Groeling, 2004). Unlike this surprisingly weak and negative relationship, the second variable of electoral size proves to be extremely significant and positive.

Increasing electoral size corresponds very strongly with a growing use of republican frames. The p> t value of 0.000 combined with the positive coefficient of .445 prove the prevalence of power in the republican rhetoric. The predicted number of frames range from 1.22 to 9.72 as adjusted along the standard deviation, one above and one below the mean. The number of republican frames exhibits smaller usage than those of the democrats who ranged from zero to 25 when linked to electoral size. The lack significance in relation to the elections in 2006 presents a surprising result, but its insignificance does not appear to harm the overall strength of the test.

The R-squared value of .243 shows a rise from the democrat model. While still in the mid twenties, the relationship of the two significant independent variables explains about a quarter of the variation in the dependent variable of republican frames. The strength of the significance for the republican regression model and the democrat model express a strong relationship, unlike the results from the four media regression tests, which yield only one significant and unpredicted result.

Prior to discussing each regression model, the independent variables require further explanation. As stated in the explanation of the data, each senator’s press releases and floor statements undergo manual coding for both sets of Social Security frames. A log of the number of both party’s frames shows the volume of democratic frames referenced by both republican senators in their releases and by democratic senators in their releases. The same follows for republican frames. This produces four different sets of independent variables with which to test the bias of the media outlets.

The first test as seen in tables three and four, over the two-month period there no relationships develop that indicate the bias in the Post. LexisNexis sorts all the articles according to the republican and democratic frames with in twenty-five words of Social Security. The articles that match with the republican frames represent one dependent variable. The set of “republican” articles are tested according to the frames mentioned by senators. The regression model for all the media tests lag one day in an effort to link the frames mentioned by senators and the use of them in the newspaper the next day.

Insert Tables 2 and 3 here

The first set of regression models run on the republican and democratic framed articles from the Washington Post provide no significant relationships. All the variables have extremely high p values. The R-squared value .008 also supports the fact that no relationship exists between the Washington Post and the amount mentions by senators. The same is true for the results from the Washington Times, save the link between the repetitions of democratic frames by democrats. This specific piece of data does not match up with expectations.

Insert Tables 5 and 6 here

The Times represents a conservative leaning media source. The significance of the democratic representatives ability to consistently get their frames provokes a host of possible explanations as well was questions. Democratic senators released a considerably larger amount of releases than republicans. For example, on February 21 and 22 Senator Schumer from New York issued several press releases each very similar to the other and all containing numerous mentions of democratic frames. Senator Durban from Illinois also made several floor statements throughout the two-month period regarding Social Security reform. The data finds Schumer referencing the established democratic frames some 344 times and 172 times for Durbin during the short January and February period. Their number of mentions far exceeds that of the republicans and most other democrats. Therefore, it appears that a few of outspoken senators have the ability to impact the message and word choice in the media, most surprisingly in a reportedly conservative news outlet.

The R-squared value for the Washington Times democrat regression of only 10% supports the general lack of significance found in all of the regressions. The other regression tests run for republican framed articles in the Washington Times yields no relationships and an R-squared value of less than ten percent. The lack of significance in all of the media outlets presents several theories regarding the position of the media and the climate for Social Security reform during January and February, especially given the strength of the regression models for the senators.

 VI. Critical Analysis

Results from the regression models both support the initial hypothesis and pose questions about the absence of expected relationships. Overall, the proliferation of democratic frames in the media and by democratic senators depends more or independent variables than the republican message. The republican message evokes a sense of immediacy while the democratic message tries to diminish the sense of urgency and provoke doubt about the President’s agenda. The initial hypothesis that the political power of the President in politics and the media makes for the notion that he and his partisans in the Senate should experience greater success in the media and in the spread of their message. The results of the models reveal that the variables influencing a senator and their choice of rhetoric do influence their number of party frames. The ability of a Senator to assert his/her party’s frame on Social Security in the media appears weakened thanks to the influence of the afore mentioned independent variables on the quantity of party frames by each Senator.

The lack of a relationship with the frame of Social Security in news media and senatorial rhetoric can be explained by taking into account the many factors influencing each individual senator. For example, the size of a state and the date of re-election differ for each senator. This competition between electoral power and the factors associated with campaigns produce a battle over rhetoric and therefore and increased attention to choice of frames by senators. Analysis behind the first regression tests reveal straightforward, significant results that can be summed up with the notion that politicians weigh several factors when deciding how exactly they communicate with the public. The ability of the press to shape a political message appears to force senators into greater consideration about policy specifics and the party affiliation of their message. Most of all, the regression results from tables one and two meet expectations but the lack of relationship in all Washington Post tests and all but one Washington Times tests present unexplained and seemingly irrational results. One possible explanation behind the absence of relationships relates to the size of the Post in comparison to the Times. The number of articles produced by the Times for the two-month period containing democratic frames numbers 92 versus the Post who numbers 154. The volume disparity matches again with republican framed articles reaching 59 for the Times and 96 for the Post. The volume inequality indicates the possibility for the Post to have a broader range of articles. The volume also signifies that the number of staff working on Social Security articles at the Times might be smaller and as a result, those who write about the reform effort might tend to site the same sources, thus resulting in a tendency toward bias. Other possible explanations include the fact that editors at the Times simply do not support the conservative agenda on Social Security. One more notion involves the idea that the democrats have done a better job at getting their desired message into the media. Ultimately, the weak significance and low R-squared value of ten percent calls into question the overall impact of the variable and raises the likelihood that Schumer’s large quantity of press releases skewed the overall data.

Reasons for the lack of significance in other variables include the short time frame, the fact that the public focus on politics in Washington centered mostly around the swearing in of new Senators and the president. The President’s State of the Union address also falls in the middle of the time rage. Most republican senators who comment on Social Security reform fail to use more than four or five republican frames and opt for ambiguous language in an apparent effort to avoid committing to a decisive stance early in the debate. This can be seen most notably in the releases responding to the State of the Union. Unfortunately, the public debate really sparked in March and April. Both the republicans and democrats markedly increased their efforts to strengthen their support base. The significance of the media coverage and the independent variables might have developed had the period been expanded. Ultimately, the use of party frames represents the ever-morphing power policy struggle in the senate.

VII. Conclusion

The battle over power, policy, and its relationship to the campaigns, reelection, and the media present a long running question about the importance of rhetoric for politicians. George Lakoff decrees the importance of frames and their ability to drive policy and public opinion. Overall Republican appear cautious about their use of decisive party frames for fear of unforeseen criticism. The frames established by both party’s attempt to link Social Security reform with discussion of poverty, race, Iraqi conflict. The results of Social Security models reveal specific influences on a senator’s public rhetoric. Republicans in the Senate released far fewer party aligned releases than Democrats, which disagrees with expectations of a party who hold the majority in both houses of congress and the White House. The results serve support a theory explored in the literature review by Groeling and Lichter, they observe that partisans of the president do not always support his policies in their public rhetoric but when it comes to voting they overwhelmingly support his initiatives (Groeling, Lichter, 2004). In the grand scheme of public opinion and political rhetoric the senate

Party Frames by Senators

Table 1. The Use of Democratic Frames on Social Security by Senators from January through February 2005

(standard errors)

Table 2. The Use of Republican Frames on Social Security by Senators from January though February 2005

(standard errors)

Washington Post

Table 3. Democratic frames in the Washington Post for January and February 2005

Lagged by One Day

(standard errors)

Table 4. Republican frames in the Washington Post for January and February 2005

Lagged by One Day

(standard errors)

Washington Times

Table 5. Democratic frames in the Washington Times for January and February 2005

Lagged by One Day

(standard errors)

Table 6. Republican frames in the Washington Post for January and February 2005

Lagged by One Day

(standard errors)

Predicted Values

Table 7. Predicted Number of Republican Frames used by Senators

Table 8. Predicted Number of Democratic Frames used by Senators

Table 9. Predicted Number of Democratic frames in the Washington Times

Created by Katie Parsley

Davidson College

Created on May 1, 2005