A Community Deliberative Polling Event: The Economic Impact of Walmart

Christopher Latimer, Karen Hempson, and J. Richard Kendrick, Jr.



This article presents the results of a deliberative poll in which members from the local community and college students from SUNY Cortland discussed the economic impact of Walmart on a small town. We review the literature concerning deliberative polling and describe the process of the deliberative polling event. Our examination of the data focuses on net changes in the participants’ opinions and gross changes in the participants’ opinions. We discuss the trends and implications of the opinion shifts and outline future research. The results illustrate that the process of deliberation affects changes in attitude items at both the individual and group level.

Social and political apathy affects both students and the broader society. Putnam (1993) contends that this phenomenon is connected to a decline in social capital, which he defines as “the features of social organization such as networks, norms, and trust that facilitate coordination and cooperation for mutual benefit”(p. 36). Putnam thinks that citizenship is affected by social capital. Moreover, involvement in civic life has been connected to lower absenteeism and reduced apathy. Social capital theory assumes that engagement on any level will enhance social trust and efficacy in citizenship, thereby strengthening democracy.

Since the 1980s, a number of methods for involving citizens and making their voices heard have been advanced. These include focus groups (Kreuger & Casey, 2000; Morris, 1999; Morrison, 2003); citizen juries (Coote & Lenhaglan, 1997; Smith & Wales, 2000; Niemeyer & Blamey, 2003); planning cells (Renn et al., 1984, 1993); citizen panels (Kathlene & Martin, 1991; Bowie et al., 1995), and consensus conferences (Einseidel, 2002; Andersen & Jaeger, 1999).

Another approach used for involving citizens by providing them a public forum for discussion is the deliberative poll. The deliberative poll is different from other methods because it allows for estimating informed opinion while retaining the possibility of generalization to the overall population through random sampling (Fishkin, 1991, 1995, 1997; Ackerman & Fishkin, 2002; Hough & Park, 2002; Hansen & Anderson, 2004; Luskin, Fishkin, & Jowell, 2002). This article presents the findings from a deliberative polling event at SUNY Cortland.

The deliberative polling method attempts to combine the depth of a qualitative analysis with the generalizability of a representative poll. Where decisions are being made without proper information, we might reasonably expect that these decisions may not entirely reflect an individual’s “true” interests. Some argue that the problem of an uninformed public and disengagement could be remedied through deliberation. As James Fishkin et al. (2000) contend:

While there is disagreement about how much lack of information and interest affects people’s views [and engagement], it is possible that preferences would be noticeably different if everyone was more knowledgeable about, attentive to, and reflexive about the issues involved. In deliberating, it is hoped that citizens will develop informed, or more reflective, preferences than would otherwise be the case (p. 657).

Those participating in a deliberative poll may shift from a position of ignorance and/or disengagement to a position of measured opinion and/or civic engagement. Fishkin argues that contemporary democracies fail to provide ordinary citizens with a means to have their voices heard. As a result, individuals believe that being informed and engaged have no utility for them. This is known as rational ignorance (Converse, Miller, & Stokes, 1960). Participation in a deliberative poll may reduce the impact of an individual’s rational ignorance by providing them with the incentive and opportunity to think about important policy issues and engage in a process that values their opinions.

The basic design of a deliberative poll includes contacting, inviting, and polling a representative sample to attend a one- or two-day deliberative polling event at a common location. Participants are then provided with carefully balanced briefing materials laying out the major arguments for and against a given set of policy proposals, policy related issues, or electoral alternatives. The participants engage in dialogue with experts and decision makers based on questions they develop in small groups with trained unbiased moderators. After the deliberations, the sample is once more provided the original questions. The consequential changes in opinion signify the conclusions the general public would reach if they had an opportunity to discuss an issue, engage with alternative points of view, and become more informed.

The goal of this project was to determine if deliberation would have any impact on participants’ opinions, regardless of whether they started from a positive or negative view of the selected issue. Increased deliberation affords a unique insight into what might be a better understanding of what people are actually thinking and feeling. The investigation of the polling event in Cortland, New York, reveals that the process of deliberation affects changes in knowledge and attitude items at both the individual and aggregate levels. This analysis includes a review of the deliberative polling literature, an explanation of our hypotheses and methodology, an examination of the data, including net changes in the participants’ opinions and gross changes in the participants’ opinions, and finally a discussion of the implications for engagement.

Literature Review 

There is a considerable body of research that attempts to assess the extent to which an individual’s opinion would differ if that person were given time to inquire about the subject matter and given information about the topic.

Deliberative Polls 

Through a number of deliberative polling events, Fishkin et al. (2000) have found that, following participation in such exercises, people are stimulated to learn more about politics and that opinion shifting is common. Luskin, Fishkin, and Jowell (2002) compiled a national probability sample of 301 British subjects who met for the world’s first deliberative poll. The deliberation focused on the root causes of crime, policing, punishment, and procedural rights, covering matters such as the rights of the accused, those of victims and the citizenry in general, as well as juvenile matters. By the conclusion of the event, participants increased their support for sending fewer criminals to prison, relaxing sentences for juvenile offenders, and alternative sentencing models for those deemed a lesser risk to society. The premise upon which such events are organized is that the resulting deliberated opinions will be more considered as a consequence of increased interest in the issues, increases in knowledge, the exposure to multiple arguments and points of view, and more careful reflection.

Hough and Park (2002) found that information and dialogue could generate important shifts in attitudes about the best ways of controlling crime. The changes they discovered were all in the identical direction, relating to a decrease in support for stronger measures such as imprisonment as a response to crime and greater support for rehabilitation. It is not unexpected that there was a quantifiable change in attitudes directly after the weekend seminar. For example, 35 percent of participants initially thought that “sending more offenders to prison” would be a very efficient way of lowering the crime rate. After the weekend, only 20 percent took this view. What is surprising is that opinion change seems to be long lasting. While 50 percent initially thought that “stiffer sentences generally” would be a very effective way of reducing crime, when followed up 10 months later only 36 percent thought the same. Support for community penalties was originally quite high and remained largely unchanged. Not all people adopted more liberal views after the event; some adopted tougher views. In general, people adopted less extreme views after the event, with a net shift in a liberal direction.

Hansen and Anderson (2004) studied the results from the Danish National deliberative poll on the single currency with a representative group of 364 Danish citizens. Between 7 and 28 percent of the participants changed their viewpoint on a number of issues related to the single currency. Before participating in the deliberative poll, 45 percent of the participants indicated that they would vote yes, 37 percent no, and 18 percent did not take a stand. At the conclusion of the poll, 51 percent revealed they would vote yes, 40 percent no, and just 9 percent had not made up their minds. The participants’ answers reveal a deliberative procedure dominated by considerable changes in opinion, an increase in knowledge, and an improved ability to form a reasoned opinion.


This analysis adds to the research concerning deliberative democracy and opinion formation. It is important to emphasize that only attitude change of some kind is predicted, rather than change in any specific direction. The theory is that the involvement of such an event will move participants from a position of ignorance and disengagement to a position of considered opinion and engagement.

H1: Participant opinions concerning the economic implications of Walmart on a small town will have a net change after deliberation on the issues.

H2: Participant opinions concerning the economic implications of Walmart on a small town will have a gross change after deliberation on the issues.


This study presents evidence from a deliberative polling event that took place at SUNY Cortland. The all-day event started at 8 a.m. and concluded at 4 p.m. Our limited budget prevented us from hosting the event for two days. The issue selected was the economic impact of a Walmart on a small town. This is particularly relevant for Cortland County as it is an economically depressed area.

Of the 76 student and community participants who attended the event, 44.7 percent were men and 55.3 percent were women; 36.8 percent were 25 years old or younger, 13.2 percent were between the ages of 26 and 40, 23.7 percent were between the ages of 41 and 55, and 26.3 percent were 56 or older. The racial makeup of participants was 96.1 percent Caucasian, 2.6 percent African-American, and 1.3 percent Native American. The income distribution for 76 participants revealed that 25 percent of participants make less than $10,000 per year, 18.4 percent make between $10,001 and $19,999, 23.6 percent make between $20,000 and $34,999, 17.1 percent make between $35,000 and $49,999, 11.8 percent make between $50,000 and $64,999, and 2.6 percent make more than $65,000 a year. The sample of participants closely reflects the demographic characteristics of Cortland.

Participants and Design 

The sampling process followed a two-stage probability design consisting of all individuals within Cortland County and all members of the SUNY Cortland student body. Volunteer students reading from a prepared script used a residential phone book covering Cortland County and called every eighth name listed. A separate college list including students was used with the understanding that this method would have the potential of selecting members of the college community who also lived locally twice. We chose every eighth name on this list as well.

One thousand forty-eight community members and students were contacted for the deliberative polling event. Of those, two hundred seven indicated that they were interested in attending. The pre-deliberation survey was given to those individuals prior to the event and they were required to bring the pre-surveys back to the event. We followed up the phone call with an additional letter and/or email reminding them of the date and time of the event. The morning of the event, the organizers received 22 phone calls from individuals informing them of their inability to attend. A total of 76 participants completed both the pre- and post-surveys for a response rate of approximately 36.7 percent (based on the number who had originally committed to attend). The lower than anticipated turnout was due, in part, to the weather. The day we held our deliberative polling event was an unusually warm and sunny early spring day in Central New York, an area noted for its long, cold, snowy winters.

Materials and Procedure 

The briefing materials provided to participants were drafted and reviewed by a panel of four faculty members from the university. The focus of the materials was information about the possible economic impacts of a Walmart store on a local economy as well as the public policies associated with this topic. The small group moderators were trained a week before the event and included faculty, staff, and members of the community. There was a six-member panel of experts, including three members of the faculty, town council members, local business people, and the mayor of Cortland.

On the day of the event, participants registered, signed the informed consent document, and then were randomly assigned to a small-group. Nine small groups contained 8-10 participants plus the moderator. The deliberations included a 60-minute small group session, a preliminary 30-minute opening session with welcome remarks by the President of the University and instructions for the day, and a 60-minute plenary session with the expert panels fielding questions, followed by a 60-minute lunch break, and another 60-minute plenary session with experts answering the remaining questions from the small groups. At the conclusion of the second plenary, the participants filled out the post survey.

The Data 

The survey instrument included 38 economic-based questions concerning Walmart. The breakdown of the questions was as follows: 11 questions concerning workers; 9 questions concerning prices; 10 questions concerning Walmart’s impact on other businesses and taxes; and 8 questions concerning economically related public policy questions and issues related to big box stores. We hoped that the number of total questions would provide a stronger incentive for participants to complete the survey. An incentives drawing was included as part of the event to increase participation. We had a number of items such as electronic equipment and items donated from the college store that were given away at the end of the event. The incentives giveaway did not begin until the moderators collected all of the surveys from each group.

A 5-point Likert scale was used to measure participants’ opinions. Likert scale items are most often used to investigate how respondents rate a series of statements by having them circle or otherwise mark numbered categories. Our scale was as follows: 1 (strongly agree), 2 (agree somewhat), 3 (don’t know), 4 (disagree somewhat), and 5 (strongly disagree). The inclusion of “don’t know” within a basic Likert scale makes an implicit acknowledgment that not all respondents will have a position or the knowledge to respond appropriately. It seemed appropriate to use don’t know for this project as we knew that not all of the participants would have formed opinions on some of the questions included in our survey.

We first analyzed our data to examine net changes in pre-survey and post-survey responses using a paired samples t-test to examine differences in mean scores. Our results are discussed below in the next section. In order to more finely differentiate types of change, we coded each pair of responses (pre-test and post-test) for each participant on each variable according to whether or not their responses indicated no change in position, a change in at least one level or degree, a change into or out of neutral (the “don’t know” category), or a change of side. For each recomputed variable, we ran a frequency distribution to analyze the percentage of respondents who exhibited no change, a change in at least one level, a change into or out of neutral, or a change in side. As a test of statistical significance, we used the chi-square test for frequency distributions.

Results: Net Change 

Following the analysis of Luskin et al. (2002), we evaluated change on two broad dimensions: net change and gross change. Net change is simply the difference between pre and post deliberation means, aggregated across individuals. These changes may be positive or negative; we were interested in magnitude of the absolute net change. By these criteria, the Walmart deliberative polling project seemed effective. The opportunity for discussion, reflection, and additional information had an impact on participants’ opinions. On nearly half of the survey items, opinions underwent statistically significant change. Table 1 presents the means, before and after participation, of the participants’ positions on every survey item, including the p-value from a paired comparison test of the significance of the differences in means. Of the 38 items, 16, or 42.1 percent, showed statistically significant change at the 0.05 level or above.

Results: Gross Change 

To get a more accurate picture of opinion change, we calculated the gross opinion change of the participants. Gross opinion change is computed in several ways. One measure of gross changes examines the percentage of the participants who changed position on the five-point scales. This measure includes any individual who moves either way on the 5-point Likert scale for a particular item to any degree at all (including moving even one position from strongly agree to agree, for example). The second measure of gross change examines the percentage of participants who change sides or who move from “neutral” to any level of agreement or disagreement with a particular item. The third measure of gross change is the percentage of participants who changed sides completely, and it does not include those who moved from neutral. Percentages of those who exhibit change on the first dimension of gross change will be higher than that of those who exhibit change on the second or third dimensions of gross change (Luskin et al, 2002). In other words, the measure of the third dimension for gross change—change in position, disregarding movement from neutral—is the most conservative measure. The measure of the first dimension, any change at all, is the most liberal. Table 2 presents several measures of gross change.


We hypothesized that participation in a deliberative polling event would be associated with a shift in attitudes and an increase in general knowledge about the subject. The observation of the net and gross attitude change of participants’ opinions would seem to support the theory of deliberative polling concerning the effect of more informed opinions. Our event was smaller than the national events that have taken place over the past twenty years. Yet, this more local focus allowed us to reduce our operational and economic costs, thereby, increasing the likelihood of a successful deliberative polling event.

The data illustrates significant opinion change by participants. In terms of net change, sixteen out of thirty-eight questions or 42.1 percent demonstrated statistically significant change at the 0.05 level or above. And while many of the changes in opinion may have cancelled each other out, we determined that overall 72.6 percent of the respondents changed position on at least one variable, that at least 63% of the respondents percent changed sides on at least one variable (including movement to and from the neutral category), and that 30.1% percent changed sides completely (excluding movement to and from the neutral category) on at least one variable.

Net Change 

The trends concerning the magnitude of net change in opinions about Walmart reflect a mixed bag with a slight majority of questions moving in a positive direction in support of Walmart. Of the 16 questions that demonstrated a statistically significant change, 10 were more negative toward Walmart and six were more positive. The questions dealing with impact on business and prices reflected a positive trend about Walmart, while the responses to the impact on workers and policy questions tended to be more negative. Why these particular changes in opinion? We can offer a few impressions, based on firsthand observations of the event.

Concerning the impact on businesses, three of the seven questions demonstrated statistically significant net change with two of the three questions shifting in favor of Walmart. With respect to whether Walmart “attracts new business to the area,” one of our experts, the Mayor of Cortland, discussed in great detail that Panera Bread, Bed, Bath and Beyond, and Lowe’s had agreed to establish themselves in Cortland if the Walmart project was approved. The Mayor also mentioned a study that the town commissioned which supported the claim that Walmart “generates a larger customer base for business” because individuals from outside Cortland County would travel to Cortland if Walmart was approved. There was also a slight shift in a positive direction for the issue “promotes growth in non growth communities.” The one question which shifted negative was the “impact of bulk purchasing” on other businesses in the area. Participants thought that existing stores would be at a disadvantage, as they had not usually offered many products for bulk purchase to save consumers’ money.

Of the four statistically significant questions concerning prices, three moved in a positive direction in support of Walmart. They included “impacts competitor’s prices,” “prices allow me to buy products I usually can’t afford,” and “brings international goods to customers.” Responses to one question moved in a negative direction, “raises the cost of living for consumers.” Of the three statistically significant questions concerning impact on business, two tended toward the positive, “attracts new business to the area” and “generates a larger customer base for business.” It appears from the data that participants’ positive opinions concerning Walmart and prices were strengthened as a result of the event and they were not as concerned about businesses that were already in Cortland.

The most significant positive shift in opinions related to prices. There was a shift for participants toward Walmart having a positive impact on competitors’ pricing, which is related to the ability of individuals to afford products that they may not usually be able to afford. The ability to afford products that would usually be considered out of range may have also influenced the question concerning cost of living. On the other hand, we noted a shift in position on “Walmart raises the cost of living for consumers” toward neutral, consistent with the shift toward agree on “Pricing leads to less competition.” Perhaps respondents recognize an immediate effect of lower prices, but a longer-term effect of higher ones if competition is diminished.

The starting positive position concerning prices may be due in part to Walmart’s marketing campaign. The “Always Low Prices” slogan was used by Walmart for 19 years and their estimated marketing budget is approximately $570 million dollars a year (Helm, 2006). Walmart has a very powerful marketing campaign that seems to have influenced the participants to some extent. Certainly the small-group discussions as well as the expert responses played some part in the overall shift of opinions.

The impact on workers section demonstrated a negative shift of opinions including “provides an increase in jobs,” “supports unions,” and “contributes to the export of jobs.” Only two of the five statistically significant questions reflected a positive trend with “treats workers fairly” and “provides more jobs for people with disabilities.” Two of the experts addressed Walmart’s exclusion of unions in response to a small-group question. There was a brief discussion about outsourcing jobs due to the increase in trade with China. Even though the Mayor of Cortland explained that new businesses had already agreed to open in Cortland, an opposing expert explained that Walmart would diminish the capacity of local businesses to employ workers leading to a net decrease in jobs. The result of the conversation was that overall job creation would not offset job loss, leaving a typical town with a net loss of jobs.

A specific question developed by one of the small-groups dealt with unions. The experts responded that Walmart does not have a union but allows workers to buy into the company with shares of stock. One statistical hint of the impact of this discussion may be seen in Table 1 in the increased negativity of participants believing that jobs were more likely to be exported overseas again leading to an overall net loss for a town. This result was not surprising given the questions asked during the plenary session focused on job creation and whether Walmart salaries were comparable to other box stores in the area. The economic downturn and loss of jobs locally may also have impacted participants’ opinions concerning middle-class workers in general. Although they do not seem to believe that Walmart creates jobs overall, neither do they seem to believe that Walmart causes net economic harm for a particular community. This may be due to the belief that the impact of Walmart on the economy of their community as a whole is positive by bringing in new business and offering lower prices.

It is important that deliberative polling also impact participants’ policy preferences. This should be the result of most deliberative polling events or these results would be revealing no more than what ordinary polling provides. The four policy questions that demonstrated a statistically significant negative response toward Walmart included “department stores should be regulated,” “the number of department stores should be limited,” “the government should establish store design guidelines,” and “communities should be able to vote on whether Walmart should be allowed.” Even though there were more positive statistically significant changes overall, all four statistically significant policy questions were negative. We were surprised by this result and are not completely convinced that this is due to the plenary session discussion.

At the same time, the deliberative experience had very little impact on participants’ policy preferences concerning the minimum wage and providing health care. As there was little to no discussion of these questions, it would make sense that opinions did not shift on these issues. The lack of change on these questions, coupled with the statistically significant changes on questions related to Walmart is consistent with our hypothesis that deliberative polling accounts for shifts in opinion. And on the items directly related to workers, there was limited opinion change concerning the treatment of workers in relation to other box stores, providing jobs with career tracks, and paying workers a living wage. On the questions concerning prices, there was also not much movement about prices at Walmart being generally fair, participants shopping at Walmart due to their prices, and consumers benefiting from the stores’ retailing system. In terms of impact of a Walmart on taxes, there was limited change in response to the questions regarding increases in the tax base of a town and benefits to individual taxpayers. All things considered, however, the overall net changes were still relatively significant.

Gross Change 

As with net change, gross change included some noticeable variations. Proposals dealing with wages and health care see relatively little gross change. This result might be expected of issues that were not addressed by the deliberative polling discussion. Another question with a lower percentage of participants changing position was Walmart “provides products at a lower price.” This is not unexpected as there might be a psychological reason for the response to this question. Once a shopper enters Walmart and notices those lower price point items, they might form the opinion that everything in the store is the lowest price in the area. This perception is reinforced with a media campaign blitz that reemphasizes this point to the consumer. Taken together, it is not surprising that participants believe that Walmart provides consumers with products at a lower price. In contrast, questions relating to the impact on workers and impact on businesses show particularly widespread gross change. Perhaps these are two areas where pre-deliberation attitudes were not as solidified and participants’ attitudes were more subject to change.

The number of participants who changed their position is between 35 and 73 percent. The percentage changing sides runs in the 11-60 percent range. The percentage of those participants changing side completely was between 3 and 30 percent. As with net change, there are some noticeable variations across policy topics. Concerning those participants who changed side completely, five out of the ten lowest percentages were found in the public policy section. On the question asking whether the “government should increase the minimum wage,” only 2.7 percent changed their position completely. Regarding the question as to communities being able to vote on whether Walmart should be allowed, 4.1 percent changed sides completely on the scale provided. And concerning the question whether the “government should pass legislation for universal health care coverage,” 4.2 percent changed sides completely.

As Luskin et al. (2002) point out, when examining gross changes, particularly those involving changing sides completely, percentages that are in the “mid-to-high single digits are impressive … and those in the twenties are astonishing” (p. 472). Those questions on which more than 20 percent of the participants changed sides (whether from the agree to the disagree range or vice versa) included, “provides an increase in jobs,” “provides jobs with career tracks,” “I shop at Walmart because of the prices,” “generates a larger customer base for business,” “draws people to the community,” “negatively impacts small business,” and “department stores should be regulated.” Of those questions, respondents tended to move from the disagree to the agree range on the variables, “I shop at Walmart because of the prices,” “generates a larger customer base for business,” “draws people to the community,” “negatively impacts small business,” and “department stores should be regulated.” Respondents tended to move from agree to disagree on the variables, “provides an increase in jobs” and “provides jobs with career tracks.” Table 2 summarizes statistically significant net changes.

There were also a number of issues where the percentage of participants who changed sides completely was statistically significant. On an issue related to workers, “provides an increase in jobs,” 30.1 percent of participants changed sides completely. The opinion shift was moving in a negative direction on this issue, weakening support for Walmart. Slightly more than 24% of participants changed sides completely about whether Walmart negatively impacts small businesses. Overall, the participants were moving in a more positive direction in response to this issue, although the mean responses mask changes in sides that could have canceled each other out. And 26.4 percent of participants changed sides completely on whether department stores should be regulated. The shift in opinion for this issue was stronger in terms of allowing the government to regulate these types of box stores. Each of these issues represents a significant shift of opinions.


There is disagreement about the efficacy of the deliberative poll as a proper research tool. It is important to note that scholars question the ability of a deliberative poll to stimulate true deliberative dialogue (Hart & Jarvis, 1999; Tringali, 1996), while others in the field have questioned the basic premise that deliberation provides the optimal format to study social outcomes (Mendelberg, 2002; Sanders, 1997).

Denver et al. (1995) in their study of one type of deliberative opinion poll found no evidence that deliberation significantly affected the quality and nature of participants’ beliefs and understanding about the issues discussed. This limitation as applied to this study is addressed in more detail under future research. They also found that participants’ knowledge of certain political “facts” had actually decreased. While the knowledge effects of deliberation are an important aspect of the arguments offered for deliberative democracy, the evidence of knowledge gains is not yet conclusive and should continue to be analyzed.

There were a number of potential limitations with our study. First, one of the common criticisms of deliberative polling is the amount of time and money necessary to successfully organize such an event. Even though we were able to save money by using local resources for the event, cost was still an issue. For example, we had to rely on volunteers to make up the expert panels. Because this issue has been debated locally, we were able to utilize local faculty members and elected officials as experts and moderators. We did, however, have a cancellation without much notice and were scrambling to find a replacement to make sure both positions were represented.

Another possible limitation concerns the reliance on the use of experts to provide information for the deliberative dialogue for the participants. In our case, we used local experts—members of the faculty at SUNY Cortland, area business people, and local politicians knowledgeable about the issue. The philosophy of the deliberative polling process is that “experts” are defined in as neutral a way as possible and the information they deliver should be factually based to aid in rational decision-making. Yet, there are clearly problems with this process. First of all, the mere fact of labeling someone an “expert” (and someone else as “not expert”) shapes how that individual’s information is received by an audience. Calling someone an expert may turn them into someone perceived to be an expert, regardless of the quality of the information that they may be presenting. Second, the personalities of the experts had a tremendous range. We had several academics who were much more reserved than other presenters, and we had several politicians who were very outgoing. Even with a moderator and a platform that included an equal amount of time to respond, it is possible that the more extroverted experts could have a slightly greater impact due to their presentation style. In addition, those speaking favorably about Walmart were fairly well known public figures in the community. Those speaking on the other side of the issue were not as well known. It is difficult, if not impossible, to perfectly balance a panel in terms of knowledge of an issue and skill at presenting one’s arguments. At the same time, labeling someone as an expert gives weight to their testimony that may not be deserved. The audience is still left to sort out which “facts” they will believe and which ones they will ignore.

Finally, the format of the deliberative polling event has been characterized as a “quasi-experiment” because it lacks the full investigational control characteristic of a laboratory experiment and because it lacks a true randomly assigned control group. The first limitation is inevitable given the setting and time treatment period. The inability to exclude extraneous influences is largely shared by all field experiments, but the deliberative polling process includes a number of elements such as the unbiased moderator and balanced briefing materials which hopefully minimize this potential problem. As for the second limitation, our deliberative polling event had a small attrition rate from those individuals who agreed to participate during the initial interview to those individuals who decided to participate. This may be explained by the smaller scope of our event. The limitations of a deliberative poll should continue to be evaluated in light of the purpose of each project and weighed against the strengths and weaknesses and costs of the available research options for the purpose at hand.

Future Research 

There are a number of questions that this project does not address that should be examined for future research. One theoretical question concerns how much of the event’s impact stems from gains in information versus increased contemplation. The questions we have to ask ourselves: are new or changed viewpoints accounted for by simply providing greater attention to the particular policy or issue presented? Do opinions shift by the mere fact of offering time for individuals to think about them? Even without acquiring more information about the economic impacts of Walmart, some participants may view the alternatives through different eyes due to different standards or to think harder about them and thus to see more clearly through the same eyes.

Another question would be to determine whether changes in opinion were derived from the briefing materials, the discussion in their small groups, and/or the plenary session in which questions were answered by the experts. It would also be interesting to determine whether changes in opinion were based on participant’s demographic characteristics such as their political affiliation or gender. Future deliberative polling events might collect data which attempt to better understand these possible influences (and further analysis of our own data might shed light on these questions as well).

One final area for further research might concern the longevity of these changing opinions or whether participants revert back because of the re-introduction of friends and family. A deliberative polling event in the future might re-poll the participants six months to a year after the initial event to determine whether or not their opinions remained the same or reverted back to their initial positions. This could help us to better understand the participants who were more or less susceptible to change during the event.


The general insight provided by this project is that deliberative polling can be a successful device for increasing the awareness and understanding of the economic impacts of Walmart amid a somewhat diverse number of participants. Our experience has shown that universities are valuable forums for these types of events, as resources can be assembled cheaply and quickly with experts and moderators such as professors, local politicians and administrators. The process of deliberation allows participants to become more informed, to realize its aptitude to solve public problems, to become engaged, to make a decision based on the virtues of an issue rather than on media sound bytes.

This deliberative polling event demonstrated that exposure to information and allowing for open discussion concerning an issue led to opinion change. Citizens who were directly involved with the deliberation were stimulated by the both the small-group and plenary session discussions. Many expressed the fact that they operated under misconceptions about Walmart, and felt that they learned a lot. At the end of the last plenary session, there was an apparent connection made between the participants. They felt comfortable in their new role as citizens deliberating on a given topic, asking key questions, and formulating new opinions.


Ackerman, J., & Fishkin, J. (2004). Deliberation day. New Haven, CT: Yale University Press.

Althaus, S. (2003). Collective preferences in Democratic politics: Opinion surveys and the will of the people. New York, NY: Cambridge University Press.

Andersen, I. E., & Jaeger, B. (1999). Scenario workshops and consensus conferences: Toward more Democratic decision-making. Science and Public Policy, 26(5), 331-340.

Beck, K.H., & Bargman, C.J. (1993). Investigating Hispanic adolescent involvement with alcohol: A focus group interview approach. Health Education Research, 8(2), 151-158.

Bennett, S.E., Flickinger, R.S., & Rhine, S.L. (2000). Political talk over here, over there, over time. British Journal of Political Science, 30, 99-119.

Bowie, C., Richardson, A., & Sykes, W. (1995). Consulting the public about health service priorities. British Medical Journal, 311, 1155-1158.

Cooke, M. (2000). Five arguments for deliberative democracy. Political Studies, 48, 947-969.

Campbell, A., Converse, P., Miller, W., & Stokes, D. (1960). The American voter. New York, NY: John Wiley and Sons.

Crandall, C., Eshleman, A., & O’Brien, L. (2002). Social norms and the expression and suppression of prejudice: The struggle for internalization. Journal of Personality and Social Psychology, 82, 359-378.

Denver, D., Hands, G., & Jones, B. (1995). Fishkin and the deliberative opinion poll: Lessons from a study of the Granada 500 television programme. Political Communication, 12(2), 147-156.

Eveland, W.P. (2004). The effect of political discussion in producing informed citizens: The role of information, motivation, and elaboration. Political Communication, 21(2), 177-193.

Fishkin, J. (1991). Democracy and deliberation: New directions for democratic reform. New Haven, CT: Yale University Press.

Fishkin, J. (1997). The voice of the people: Public opinion and democracy. New Haven, CT: Yale University Press.

Fishkin, J, & Luskin, R.C. (1999). Bringing deliberation to the democratic dialogue. In M. McCombs & A. Reynolds (Eds.), The poll with a human face: The national issues convention experiment in political communication, pp. 59-84. Inglewood Cliffs, NJ: Lawrence Erlbaum Associates.

Fishkin, J., Luskin, R.C., & Jowell, R. (2000). Deliberative polling and public consultation. Parliamentary Affairs, 53, 657-666.

Fishkin, J., Luskin, R., & Jowell, R. (2002). Considered opinions: Deliberative polling in Britain. British Journal of Political Science, 32, 455-487.

Gastil, J., & Dillard, J.P. (1999). Increasing political sophistication through public deliberation. Political Communication, 16(1), 3-23.

Gordon, S.B., & Segura, G.M. (1997). Cross-national variation in the political sophistication of individuals: Capability or choice?. Journal of Politics, 59(1), 126-147.

Gutmann, A., & Thompson, D. (1996). Democracy and disagreement. Cambridge, MA: The Belknap Press of Harvard University Press.

Helm, B. (October 9, 2006). Wal-Mart, please don’t leave me. Business Week, p. 84.

Hansen, K.M., & Andersen, V.N. (2004). Deliberative democracy and the deliberative poll on the Euro. Scandinavian Political Studies, 27(3), 261-286.

Hart, R., & Jarvis, S. (1999). We the people: The contours of lay political discourse. In M. McCombs & A. Reynolds (Eds.), The poll with a human face: The national issues convention experiment in political communication, pp. 59-84. Inglewood Cliffs, NJ: Lawrence Erlbaum Associates.

Hough, M., & Park, A. (2002). How malleable are public attitudes to crime and punishment? In J. Roberts & M. Hough (Eds.), Changing attitudes to punishment: Public opinion crime and justice, pp. 163- 183. Collompton, Devon: Willan Publishing.’

King, C. S., Feltey, K.M., & Susel, B.O. (1998). The question of participation: Toward authentic public participation in public administration. Public Administration Review, 58(4), 317-326.

Krueger, R.A., & Casey, M.A. (2000). Focus groups: A practical guide for applied research. Thousand Oaks, CA: Sage Publications.

Lyn, K., & Martin, J.A. (1991). Enhanced citizen participation: Panel designs, perspectives, and policy formation. Journal of Policy Analysis and Management, 10(1), 46-63.

Mendelberg, T. (2002). The deliberative citizen: Theory and evidence. In M. Delli Carpini, L. Huddy, & R.Y. Shapiro (Eds.), Research in micropolitics, (pp. 151-193). New York, NY: Elsevier Press.

Merkle, D.M. (1996). Review: The national issues convention deliberative poll. Public Opinion Quarterly, 60(4), 588-619.

Morris, D. (1999). Behind the Oval Office: Getting re-elected against all odds. Los Angeles CA: Renaissance Books.

Morrison, D.E. (2003). Good and bad practice in focus group research. In V. Nightingale & K. Ross (Eds.), Critical readings: Media and audiences, pp. 59- 84. Maidenhead: Open University Press.

Mutz, D. (2002). Cross-cutting social networks: Testing democratic theory in practice. American Political Science Review, 96(2), 111-126.

Neijens, P.C. (1987). The choice questionnaire: Design and evaluation of an instrument for collecting informed opinions of a population. Amsterdam: Free University Press.

Neijens, P.C., De Ridder, J.A., & Saris, W.E. (1992). An instrument for collecting informed opinions. Quality and Quantity, 26, 245-258.

Price, V., & Neijens, P. (1998). Deliberative polls—Toward improved measures of informed public-opinion. International Journal of Public Opinion Research, 10(2), 145-176.

Purnam, R. (1993). The prosperous community: Social capital and community life. The American Prospect, 13, 35-42. 

Renn, O., Webler, T., Rakel, H., Dienel, P., & Johnson, B. (1993). Public participation in decision-making: A three-step procedure. Policy Studies, 26, 189-214.

Smith, G. & Wales, C. (2000). Citizen juries and deliberative democracy. Political Studies, 48, 51-65.

About the Authors 

Christopher Latimer is an associate professor in the Department of Political Science and associate director of the Institute of Civic Engagement; Karen Hempson is Lecturer II in the Childhood Education Department; and Richard Kendrick is a professor in the Sociology/Anthropology Department and founding director of the Institute for Civic Engagement—all at the State University of New York at Cortland.

Leave a Reply

Your email address will not be published.