John W. Williams, Political Science,
Principia College, Elsah, IL 62028

This paper was originally presented at the 1993 annual meeting of the Illinois Political Science Association.

Campaigning for local political office, as in Jersey County, Illinois, appears to be conducted by instinct. The question of how to campaign most effectively resurfaces with every primary and general election. This is especially true in communities which draw little attention or guidance from political scientists, political communications experts and political consultants. Little research has been conducted on the role and impact of media in elections for public office other than at the national level. Among the limited research, McCleneghan's work on mayoral elections in Texas and New Mexico indicates that such elections can be more affected by the media than by personal attributes of the candidates, including the power of incumbency. Luttbeg found that local newspapers apparently raise public consciousness about city council and school board elections and, thereby, encourage voting participation. The purpose of this study is to shed light on the nature of media use for political information in local elections.

People get their information about state presidential primaries, governors, state legislatures and state supreme courts from a variety of sources, rather than from one medium, such as television or a newspaper. Reagan and Ducey found 75% of their respondents used multiple sources. Yum and Kendall found over 80% of their respondents relied on at least two media (television and newspapers), and 39% relied on three or more media. Berkowitz and Pritchard expanded their list of political media to "communication resources": "the news media (newspapers, television, magazines, radio), interpersonal communication (conversations with family, friends, colleagues), and information that emanates directly from a political campaign (ads, campaign literature, conversations with a candidate)."

In Jersey County, present and past candidates, both successful and unsuccessful, and political party officials, who were interviewed in preparation for this study, admitted that they do not know which campaign resources are truly effective and which are a waste of time and money. Decisions on which media to use appear to be made on the basis of personal preference, "gut instinct," tradition, or reaction to what other candidates are doing. For example, according to one party official, cable television (found only in the county seat of Jerseyville) has not been used because it would force everyone into the medium, thereby raising the cost of campaigning without any guarantee of effectiveness.

Nevertheless, two themes kept reappearing in the discussions with the politicians and political leaders of Jersey County. First, almost every person believed that the most widely consumed medium in Jersey County, thereby the best medium for transmitting political information, is the Jersey County Shopper, a weekly "throw-away" tabloid consisting of classified and display advertising. Second, all believed that personal contact between the candidate and the voter is the most effective medium for influencing voters, therefore the traditionally heavy emphasis on door-to-door campaigning and personal appearances at social gatherings. The two research questions were (1) which medium do Jersey County voters claim to get most of their "political information," -- information that assists their decision on how to vote; and (2) which medium do Jersey County voters claim most influences in their vote?

The goal of this study was to discover the "political" media consumption habits of the voters of Jersey County. It is a relatively small rural Illinois county (population just over 20,000) located at the confluence of the Mississippi and Illinois rivers about fifty miles north of St. Louis, Missouri.

The respondents were not randomly selected. Instead, with assistance of students from two local colleges, we attempted was to survey every voter at four specifically chosen polling places. According to past registration and voting patterns, Jersey County voters are roughly divided between Jerseyville, the county seat and largest town, and the outlying villages and rural precincts. Therefore, two Jerseyville precincts and two non-Jerseyville precincts were selected. Original exit polling was conducted at two polling places -- one rural and one city -- during the March 1990 Illinois primary. Subsequent exit polling was conducted at two more polling places -- one rural and one city -- during the November 1990 Illinois general election.

Jersey is a swing or split county. For example, the state representative, sheriff and county clerk are Republicans while the circuit judge, circuit clerk, treasurer and coroner are Democrats. The four precincts were selected because they have voted most consistently with the majority of the county. In the last 22 contested races in the past three general elections (1988, 1986 and 1984), the voters of Elsah2 voted identical to the county as a whole. Of the 22 winners, 13 were Democrats and 9 were Republicans. If the county voted for state senator Vince DeMuzio (Democrat), so did Elsah2. If the county voted for President George Bush (Republican), so did Elsah2. The results from Elsah2 were 100% consistent, regardless of the party of the winner, in the contested 22 votes. No other precinct could do better than 21 out of 22, including Jersey2, Jersey4 and Mississippi1. While six of 12 Jerseyville precincts voted with the county 21 out of 22 votes, only three (including Jersey2 and Jersey4) voted 21 out of the most recent 21 races. After discussion with the County Election Authority, Jersey2 and Jersey4 were selected because they were the most established, hence most likely to replicate past voting behavior, and the largest, hence most likely to produce a larger survey population. Mississippi1, the other rural precinct, was chosen for the same reasons. Based on those results, Jersey2, Jersey4 and Mississippi1 voted consistent with the county as a whole in 95.4% of the contested races.

[See Table I -- Precinct Voting Record]

The choice of the four precincts was further supported by the results of their voting in the November 1990 general election. All four of the precincts voted consistent with the county as a whole in the federal (U.S. Senator, U.S. House of Representatives), leading state (Governor/Lt. Governor, Attorney General) and county-wide races, even though their choice for Governor was defeated statewide.

[See Table II -- November 1990 Precinct Voting Record]

Based on interviews with local political leaders and candidates, a list was developed of the eleven media normally considered and used by Jersey County candidates during political campaigns. All of the media are local, including the weekly advertising-only Jersey County Shopper, which is delivered by mail to nearly every household in the county, and The Telegraph newspaper. With the closing of the Democrat-News (a free weekly newspaper) several years ago, the county was left without a local newspaper. The Telegraph, the daily regional newspaper based in Alton, integrated the staff and functions of the Democrat-News into a weekly supplement called the Tri-County Edition (for the three rural counties of Calhoun, Greene and Jersey). Local candidates also advertise on WJBM, the county's daytime AM radio station, which reaches Jerseyville and most of the northern and central portions of the county. Campaign season in Jersey County is announced with the proliferation of store, yard and pole signs, bumper stickers, handouts and flyers, and mailed literature. As is true in rural and small communities, much of the campaigning is interpersonal in nature -- the candidates, campaign workers and political officials go door-to-door, speak at all sorts of gatherings, march in parades and stand on street corners. As Lazarsfeld and others found with the two-step flow of influence and information model, political information flows among family, relatives, friends and neighbors, as well as co-workers and people crossing paths while shopping, in diners and at church.

Certain media, such as major St. Louis television and radio stations and the St. Louis Post-Dispatch, are missing from the list because local candidates cannot afford to use them. One local medium, a cable system reaching 1500 homes within the city of Jerseyville, was not included because it was never suggested in the preliminary interviews. Until specifically prompted, the candidates and political leaders appeared to have forgotten its existence. Likewise, the initial interviews did not alert me to a weekly newspaper based outside the county which is distributed in one precinct of the county not covered by the Shopper. However, that precinct, at the eastern edge of the county, was not surveyed.

Upon exiting the polling places, each voter was asked to complete a three page questionnaire that listed the eleven media and, for each media, to answer the following questions: "About how much information about this election did you get from...?" and "How much did these sources influence how you voted?" The respondents answered by circling "a lot," "some," or "none," on a three element Likert-scale. The respondents also answered a number of questions aimed at identifying political behavior.

There were several problems with this study. Since voters were surveyed at only four polling places, the results of this survey cannot be statistically inferred to the entire county. However, given that the four precincts appear to be "bellweather" precincts, the results may reflect county-wide behavior to some degree. Furthermore, the results reflect only those who vote in primary or general elections. Assuming that those who vote are representative of committed and involved voters, and that the precincts may reflect the county generally, the following information may be of some use.

FINDINGS -- GENERAL INFORMATION

A total of 542 of 980 voters (55.3%) answered the questionnaire. [See Table III for breakdown of response rate by precinct.] The number of valid cases for the first set of questions concerning the eleven media as sources of information averaged 494. The number of valid cases for the second set of questions concerning the eleven media as sources of influence averaged 478. The drop can be explained, in part, by the natural loss of interest in completing the questionnaire as the respondents worked their way through it.

There were 240 ballots (including absentees) cast in the two original precincts during the March primary, of which 130 were Democratic and 110 were Republican (Illinois is an open primary state in which party preference is determined by which ballot one asks for at each primary), for a 54.2% to 45.8% Democrat to Republican split. Of the 73.6% of the respondents who answered the specific question about political party affiliation, 54.2% said they were Republicans and 45.8% said they were Democrats.

The difference between claimed affiliation and ballot selection could be explained by the Illinois tradition of "cross-over" voting. The only contested county-wide race, for Democratic candidate for circuit court judge, may have attracted cross-over voting by Republicans. The only contested sub-county race was for the Republican candidate to the county board to represent the portion of the county which includes the Elsah precincts. In a county where local roots are important, Elsah Democrats may have crossed over to support the Elsah Republican against a non-Elsah Republican (the winner would face a non-Elsah Democrat). This assumption is supported by fact that 39 Republican versus 38 Democratic ballots were taken in Elsah2, the home precinct of the county board candidate.

Of the total respondents, 161 (29.7%) identified themselves as Democrats, 160 (29.5%) identified themselves as Republicans, 5 (.9%) claimed some other affiliation, while a plurality (216 or 39.9%) did not claim party affiliation. There were 53.4% women and 46.6% men. Of those indicating their age, the respondents ranged from 18 to 88, with an average age just under 48. The subjects appeared to be active voters, with 64% claiming to have voted in the last primary election and nearly 93% claiming to have voted in the last general election. While these people certainly do not reflect the American mainstream, they are the citizens who matter most to the candidates -- they vote.

FINDINGS -- MEDIA USE AND MEDIA INFLUENCE

Based on the mean or average score for each medium in each question, the following are the rankings of the media in order of source of most information and source of most influence:

[See Tables IV and V]

Clearly, those directly involved in the local political process predicted accurate. The voting public claims to obtain most of its political information from the Shopper, the throw-away paper, and is most influence by face-to-face contact with the candidates.

FINDINGS -- STRENGTH OF POLITICAL AFFILIATION

Recognizing the Illinois tradition of cross-over voting and Jersey County's inclination to split its ticket (in the last 22 contested county-wide elections, 13 Democrats and 9 Republicans won), questions were asked to explore the strength of political party affiliation.

Those who vote in primaries are generally committed voters. Of those responding, 96% voted in the last general election and 86% voted in the last primary election. Even though in Illinois one "declares" for a party in selecting a primary ballot, only 77.7% considered themselves aligned to a political party (54% Republican, 46% Democrat). Nearly two-thirds of those claiming party affiliation (62.8%; but only 46.6% of the total primary respondents) claimed that their affiliation was strong or very strong. However, only a third (33%) of the total respondents said they voted a straight ticket. Approximately a third said they had worked for a political candidate or party (31%), or contributed money to a candidate or party (38%) in the last election. Just under half said that they have put up a yard signs or bumper sticker (49%).

Of the combined primary and general election responses, two-thirds (63.7%) claimed party alignment (49.4% Democrat, 49.1% Republican, .5% other). Overall, 93.1% of those responding voted in the last general election and 76.1% voted in the last primary election. On the whole, they were less committed to a candidate or a party. Only 21.6% vote straight ticket, 20.0% worked for a candidate, 29.8% contributed money to a candidate or party, and 39.2% put up a yard signs or bumper sticker. [Calculations and tables, available from author, are based on frequencies and Chi-square/cross tabulations.]

DISCUSSION -- MEDIA USE AND MEDIA INFLUENCE

True to the common wisdom, more people claimed they got "a lot" of information from the Shopper than from any other source (24.9%). And, fewer people said they got no information from the Shopper than from any other source (24.1%). [See Table VI -- Source of Most Information] This supports the commonly held belief that Jersey County residents read the Shopper regularly and thoroughly. More campaign money is probably spent on ads in the Shopper than on any other mass media (ads in Telegraph and on WJBM). The ads include extensive copy, including lists of accomplishments and campaign promises, and formal and informal photographs of the candidates and their families. What the Shopper provided that no other media could, save mailed literature, was delivery of information to every voting household.

Mailed literature has the added advantage of being addressed to every voter (since mailing labels of voters are readily and cheaply purchased from the county election authority). This may explain why literature in the mail was rated as the second highest sources of political information. With the exception of the Shopper, more people claimed the got "a lot" of information from the mail (20.9%) and fewer claimed that got "none" (25.6%). Like the Shopper, direct mail can guarantee at least reaching the household and can include extensive copy about and photographs of the candidates and their families.

The next four sources -- the candidates, family and relatives, yard signs and bumper stickers, and political party officials -- indicate similarity in their impact. The ability of the candidates to transmit information may be related to the practice of campaigning door-to-door and at community functions (Jersey County's "chicken dinner" circuit ranges from biscuit and gravy breakfasts to estate auctions to the county fair parade). Furthermore, the candidates have been identified as the most influential media. The relative high ranking of family and relatives and political party officials may reflect Lazarsfeld's two-step flow in operation. In some instances, the officials served as surrogates for the candidates. An anomaly, however, is the high ranking of yard signs and bumper stickers as sources of information. Most have very little copy, usually identifying only the candidate's name, position sought and party. The ranking may be explained by the use of signs (including those in store windows and on billboards) to establish and increase name recognition.

Ads in the Telegraph ranked lower than yard signs and bumper stickers. This may reflect the limited circulation of the paper in the county. The Telegraph, a regional paid-circulation daily, focuses most of its attention on news of the communities down river from Jersey County. The circulation in the county is much smaller than that of the Shopper. However, the Telegraph publishes a special zoned supplement aimed at Jersey (and two neighboring counties) once a week.

Also true to common wisdom, more people claimed they were influenced "a lot" by the candidates directly than by any other source (30.8%). And, fewer people said that the candidates, rather than any other source, had no influence (27.9%). [See Table V -- Source of Most Influence] The influence of candidates was ranked substantially higher than that of any other media. Local politics in rural communities and small towns is still based on personal exposure. Candidates have long recognized the importance of meeting the voters face-to-face. One candidate told of how one potential voter responded: "I like the way you look me in the eye."

Five media -- political party officials, family and relatives, ads in the Shopper, literature in the mail, and handouts/leaflets/flyers -- were ranked as very similar. As with the candidates, party officials, family and friends form the personal contact network that creates the political community, consistent with the two-step flow model. Ads in the Shopper and direct mail were identified as the two most important sources of information. Along with the leaflets/flyers, these forms of media permit candidates to distribute substantial amount of information and persuasive material, including pictures of themselves with their families, in church or playing on a local softball team.

At the opposite end of the spectrum, over half of the respondents said they got no information from the ads on WJBM (58.1%) and over two-thirds said they were not influenced by the radio ads (70.0%). Less than 5% of the respondents said that WJBM had "a lot" of influence on their vote, the lowest of any source. Also, over half of the respondents said they got no information from coworkers (52.9%) and over two-thirds said coworkers had no influence on their vote (68.7%).

There may be a number of explanations for both rankings. Although it claims a 40% share of the local market (in part because there is no other local competition), the station's programming is aimed at the farm and retired sectors of the community. WJBM is a local AM radio station that does not reach all of the county (especially the southern part of the county and portions of the Elsah precincts). Due to the nature of AM radio, the station usually operates only during the daylight hours. Furthermore, due to the nature of its signal, the station is heard from Jerseyville northward. However, there was no significant difference between the rankings by the rural and city voters, so the station signal does explain the low ranking. However, to be effective, like any broadcast medium, radio requires repetition of message. Programming is designed with the recognition that listeners come and go, and that radio is usually a supplemental or secondary medium, often used as a companion or as background. Thus, the audience for radio advertising may not be there or may be transient to receive the necessary repetitive message.

Given that many Jersey County voters work outside the county, it is not unexpected for them to rank coworkers as a low source of local political information or influence. Furthermore, given the close nature of the community, local coworkers are as likely to be friends, neighbors or even relatives, and, as a result, be rated in other categories.

The circumstances of prior elections highlight the situation. The incumbent sheriff was challenged by a retired state trooper. The sheriff's daughter is married to the state trooper's son. The county treasurer and the state's attorney for the neighboring county are brothers of the deputy chairman of the county board and former chairman of the county Democratic party. Their sister is married to another member of the county board. The circuit clerk's brother is also on the county board, while the brother-in-law of the county clerk regularly challenges the deputy chairman of the board for his seat. It appears that everyone is on amiable terms and each relative has to run on his or her own merits. For example, the member of a large and well-known farming family (with members on the school board and board of trustees for the community college) failed to unseat the county treasurer.

The bottom line is that the political officials and past candidates were accurate in their instincts about the Shopper and face-to-face campaigning. However, they appear reluctant to avoid advertising in the other, lower ranked media.

DISCUSSION -- STRENGTH OF POLITICAL AFFILIATION

There is no statistical evidence of any difference between the two political parties in Jersey County and how voters perceive the strength of their commitment to their party, their practice of voting straight ticket, or their willingness to work for a candidate or the party or put up a yard signs or talk publicly about politics. On these matters, Jersey County Republicans are just like Jersey County Democrats. Voters who align themselves with either party are more likely to vote straight ticket (p<.001), work for a candidate or party (p<.001), contribute money (p<.001), and put up a yard signs or bumper sticker (p<.01). The voters who are strongly or very strongly committed their party are more likely to vote straight ticket (p<.001), work for a candidate or the party (p<.001), or give money (p<.01).

There is no statistical difference between the city and rural voters in their party alignments, how they perceive their commitment to their party, or in the frequency they voted in the past general or primary elections, vote straight ticket, or discuss politics. The rural and urban voters gave the same rankings of media as sources of information and influence (results were subjected to Chi-square/cross tabulation), although with several differences in intensity.

The differences appear in their use of the candidates (chi-square 9.0, df 2, p<.05) and party officials (chi-square 12.9, df 2, p<.01) as sources of information. This may be the result of the campaign behaviors of the candidates and the party officials. It is easier for candidates to campaign in more densely inhabited areas. All of the candidates went door-to-door and stood on street corners in Jerseyville, while only a few attempted to reach every household in the county. Also, as county seat and largest community (approximately one-third of the population of the county), Jerseyville is also center of county political activity. Party officials only appeared at political events, such as parades (in the two largest "towns") and rallies. Likewise, the two groups reported differences in their use of the Shopper (chi-square 10, df 2, p<.01), for which I am at a loss to explain. One the other hand, direct mail revealed no difference between the two groups of voters. This should have been expected since every voter was mailed literature and his study polled only voters.

A relatively low number of voters (especially in a primary system that requires one to publicly declare for a party to get a ballot) expressed strong or very strong allegiance to one of the parties (46.6%) or commitment to vote a straight ticket (33.1%). In other words, committed participants in the democratic process are open to being persuaded to cross party lines. The old wisdom that candidates should focus only on the uncommitted voters is not necessarily true. Candidates need to appeal to their own party members, because opposition candidates can successfully appeal to them. This tends to confirm that Jersey County voters, as a whole, swing their votes for reasons other than party. Are they voting based on qualifications, family or personal relationships, name recognition, or some other criteria? Further research should explore the reasons why Jersey County voters cast their votes for certain candidates.

However, we now have a fairly clear picture of the political media consumption patterns of Jersey County voters. The common political wisdom, at least in this instance, proved to reliable.

TABLE I

PRECINCT VOTING RECORDS

Question -- How consistent has each of Jersey County's 29 voting precincts been in voting with the majority of the county? Out of 22 contested votes in the last three general elections (1988, 1986, 1984) (includes President, Governor, Secretary of State, Attorney General, U.S. Senator, U.S. House of Representative, State Senator, State Representative, County Clerk, Circuit Clerk, Sheriff, State's Attorney, Coroner, etc.), 13 Democrats and 9 Republicans carried Jersey County. Lesser state constitutional positions, such as Secretary of State, Comptroller and Treasurer are usually contested but were not included because of minimal level of exposure accorded these races. The attorney general was included because for the last decade the position has been the highest state constitutional position held by the Democrats.

PRECINCT NUMBER OF CONSISTENT VOTES PERCENTAGE

Elsah 2 22 100%

Jersey 2 21 95.45%

Jersey 4 21

Jersey 8 21

Jersey 9 21

Jersey 11 21

Jersey 12 21

Mississippi 1 21

Otter Creek 21

Piasa 1 21

Elsah 1 20 90.9%

Fidelity 20

Jersey 7 20

Jersey 10 20

Jersey 1 19 86.36%

Richwood 19

Ruyle 19

Elsah 4 18 81.82%

English 18

Jersey 3 18

Mississippi 2 18

Piasa 2 18

Piasa 3 18

Jersey 5 15 68.18%

Jersey 6 14 66.74%

Quarry 1 14

Quarry 2 13 59.09%

Rosedale 13

Elsah 3 11 50%

Quarry 1 and 2 (the town of Grafton) and Rosedale are staunch Democratic precincts. With one exception (Quarry 1 voted for the Republican state representative in 1986), these three precincts voted straight Democrat in the 22 contested races. Elsah 3 (Principia College) is considered staunch Republican, although it voted for Michael Dukakis and Rep. Dick Durbin, a Democrat, in 1988, and Durbin and Sen. Alan Dixon, a Democrat, in 1986. In fact, the precinct helped a Democratic faculty member (now the college's president) win election to the county board.

TABLE II

PRECINCT VOTING -- NOVEMBER 1990

There were seven contested county-wide races on the 1990 ballot. There were contested races for three lower state constitutional positions (Secretary of State, Comptroller and Treasurer) which were not included in order to be consistent with Table I. The county voted for six Democrats and one Republicans. The four subject precincts voted with the county 100%. This raises the precincts' voting consistency since 1984 to at least 96.55%.

Actual County Precinct:

Position: Winner: Winner: E2 J2 J4 M1

------------------------------------------------------------

U.S. Senator D D D D D D

Governor/Lt. Gov. R D D D D D

Attorney General D D D D D D

U.S. Congressman D D D D D D

County Treasurer D D D D D D

County Sheriff R R R R R R

County Judge D D D D D D

CONSISTENCY IN VOTING WITH COUNTY: 100% 100% 100% 100%

TABLE III

RESPONSE RATES

ACTUAL VOTE NUMBER OF RESPONSE

(minus absentees) RESPONSES RATE

------------------------------------------------------------

ELSAH 2 73 53 72.6%

JERSEY 2 153 110 71.9%

JERSEY 4 359 199 55.4%

MISSISSIPPI 1 395 179 45.3%

------------------------------------------------------------

TOTAL: 980 541 55.2%

TABLE IV

SOURCE OF MOST INFORMATION

TABLE V

SOURCE OF MOST INFLUENCE