Asia Pacific Advances in Consumer Research Volume 3, 1998 Pages 68-73
LOTTERY ADVERTISING AND ETHNIC GROUP PLAY OF LOTTERY GAMES
Richard Mizerski, Griffith University, U.S.A.
Katherine Straughn-Mizerski, Griffith University, U.S.A.
Jakki Williams, University of Arizona, U.S.A.
Edward Forrest, Griffith University, U.S.A.
Blacks, Whites and Hispanics have reported significant differences in playing state lotteries. Recent criticisms of these games blame the ethnic targeting of lottery advertising for the disproportionately higher level of play by Blacks and Hispanics. Although income and education may account for some of these differences, the influence of promotion and cultural models of gambling have had little empirical study. This research reports on the use of a two wave survey of Florida residents that contrasts their self-reports from 6 months after the lottery introduction to a period twenty-seven months later. These surveys will be used to test for the existence of ethnic differences in lottery game play, and their relationship to alternative cultural-based models of gambling and memory of non-targeted advertising for state lottery games.
Participation in legalized gambling has increased for people from all demographic backgrounds and psychographic profiles (Edmondson 1986; McConkey and Warren 1987). In fact, legalized gambling is being hailed as "Americas No. 1 growth industry" and the hottest new "entertainment" pastime (Zuckoff and Bailey 1993). As gamblings stigma erodes, it is now possible to make a legal wager in all but two states (Utah and Hawaii) with lotteries offered in 38 U.S. states plus the District of Columbia (Griggs 1994).
State lotteries appear to be driving much of the growth in legalized gambling. The lottery industry is now considered one of the 30 largest industries in the United States (Kaplan and Blount 1990). Lotteries provided revenue-strapped states $21.8 billion in 1992 (Griggs 1994), with a substantial additional amount allocated to lottery management. Welles (1989) reports that total lottery wagering increased 229% over a five-year period from 1983 to 1988. The growth in wagering for state lotteries far exceeds the growth rate in wagering for casinos (slot machines and table games), pari-mutuels (horses and dog racing), charitable games (bingo), and illegal gambling (Green 1991).
Subsequently, this growth of state lottery play has generated much criticism from opponents of legalized gambling (Clotfelter and Cook 1989; Karcher 1989). They argue that state lotteries exert regressive taxation (Kaplan and Blount 1990), and contribute to increased gambling addiction (Lorenz 1990). It is also claimed that lotteries disproportionately exploit vulnerable segments of the states population such as ethnic groups and the poor with "misleading" and "targeted" lottery advertising (Karcher 1989; Selinger 1993).
The disproportionate lottery play of Blacks and Hispanics has been well documented (Green 1991). In general, both ethnic groups [The term ethnic group is subject to varied popular interpretation. The comparison of Black and White racial groups along with various Hispanic groups such as Mexican, Cuban and Jamaican actually combines racial and ethnic comparisons, but will be called ethnic in this paper.] exhibit above average participation in the various games offered (e.g., instant, draw games and Lotto), spend more per game and have a greater relative prevalence of heavy play members in their respective groups (Clotfelter and Cook 1989).
The three major reasons provided for this disproportionate play by Blacks and Hispanics are cultural influences or the deliberate targeting of these groups with advertising and distribution (Karcher 1989). Even if not targeted, some suggest that these minorities are more vulnerable to persuasive commercial messages. Of course, some combination of these and other influencing factors may best explain lottery play.
The objective of this study is to examine the effects of racial/ethnic background on lottery play in a promotion environment absent of any specific group targeting. This would eliminate the influence of targeted creative or media efforts from the perceptions and behaviors of lottery players. Any differential in game play by group will be compared to a measure of promotion effectiveness to explore differential persuasiveness. Finally, competing models of why individuals gamble will be tested.
Ethnic Differences in Gambling and Lottery Play
Research findings indicate that gambling is very popular in large urban areas that have a sizeable ethnic population (Cook and Clotfelter 1993; Kallick-Kaufmann 199; Karcher 1989). Unfortunately, little published theoretical research exists on ethnic differences in lottery play (Clotfelter and Cook 1989; Kaplan and Blount 1990; Karcher 1989). Most of the existing research is descriptive (i.e., Kallick-Kaufmann 1979) rather than explicitly theory-driven. For example, Clotfelter and Cook (1989) compare the rates of lottery play for Hispanics, Whites, Blacks, and Asians from an unpublished Los Angeles Times poll. For one week for the California lottery during 1986, participation rates for the various ethnic groups found were 56% for Hispanic, 35% for Whites, 34% for Blacks, and 32% for Asians.
Also, most existing research focuses on group differences in illegal gambling and numbers play (Kaplan and Blount 1990). For example, Kallick-Kaufmanns (1979, p. 16) findings indicate that "that proportion of illegal gamblers among blacks and those of Hispanic origin is double that among other gamblers." Similarly, a 1973 study in Black Enterprise Magazine reports:
"There are few reliable national statistics on policy gambling but police across the nation agree that blacks are more frequent players of numbers than any other group. However, in New York, the city with the heaviest play, reportedly more Whites wager than Blacks, although 40 percent of the Black and Puerto Rican population play the game as opposed to 20 percent of the citys White population (p. 41)."
One exception to the trend of descriptive research is Light (1977), who proposes that numbers gambling has become institutionalized among many minority groups in the United States. This illegal gambling industry provides many group members with jobs and opportunities for upward mobility for themselves and their families that otherwise would not usually be made available to them.
PERSPECTIVES ON LOTTERY PLAY
Cultural Reasons for Differences in Game Play
As an alternative explanation to those previously suggested, four models of lottery play have been proposed to explain why different racial or ethnic groups show differences in gambling activity. These models (pathology, economic, stimulation and socio-cultural adaption) have emerged from the gambling literature, and are primarily from the fields of psychiatric medicine, psychology, sociology and economics. Surprisingly, little has been reported in the consumer behavior literature (e.g., Burns et al. 1990).
The four models primarily differ as to whether lottery play is deemed noxious or innocuous, and whether the locus of stimulation is internal or external to the player (see Figure 1).
The pathology model (e.g., Custer 1980; Cohen 1988) views gambling as an addictive illness, where the player is driven by an overwhelming impulse to gamble that often destroys everything that is meaningful in the persons life.
The stimulation model (e.g., Welles 1984; Kallick-Kaufmann 1979) views lottery play as a healthyinnocuous escape from the frustrations of modern society. Some people are viewed as playing for excitement and challenge, and others play for pleasure or to eliminate boredom. The area of hedonic consumption would apply here.
The socio-cultural model (e.g., Light 1977; Campbell 1976) suggests that lottery play provides an opportunity for the social interaction that is missing from existing structures and networks. It also offers many players an opportunity to emulate traditional entrepreneurial practices and provides an alternate form of investment.
Finally, the economic model (e.g., Li and Smith 1976; Smith 1976) provides several explanations for the gambling differences associated with ethnicity, income and social class. Gambling is suggested to be a function of player knowledge, utilities for winning and losing, and the players use of gambling as consumption expressing status.
There is no previous empirical evidence that ethnic groups will conform to different models, although Blacks and Hispanics have been anecdotally associated with the pathology (e.g., Wells 1989) and socio-cultural adaption (Cohen 1988) perspectives.
Selinger (1993) notes that advertising is an essential element of lottery marketing as evidenced by the combined advertising budget for state lotteries of $286 million for fiscal 1992. This would rank state lotteries among the top 50 advertisers in the U.S.A. Lorenz (1990) reports that lottery advertising has come under review by various state legislators for truth in advertising, "questionable taste" in hard-sell appeals, and the use of lottery advertising to promote other forms of gambling. Executives from the American Advertising Agency Association suggest that much of this sentiment is like that of the opponents of tobacco and alcohol who "tend to blame advertising for the social ills surrounding the products rather than focusing on the products themselves." Gambling advertising is the next hot issue with Griggs (1994) reporting that the major U.S. states legislative battle in the 1994 sessions were about restrictions to lottery advertising.
While only a handful of states have established policies that specifically restrain aggressive advertising, the number is growing. Karcher (1989) argues for more regulation of lottery advertising based on charges that lottery agencies intentionally target minorities and the poor with advertising campaigns. Some argue that non-intentional targeting to vulnerable groups may in itself be unfair and deceptive.
Although lottery game play can be traced to a retail location with sales data, individual player demographics and purchase patterns must be obtained with a survey. A set of phone interview surveys were collected from the state of Florida lottery, one of the largest and most successful U.S. state lotteries (Green 1991). The total data set incorporates a nine wave telephone survey (n=7,401) of state residents 18 years or older (legal age to play). The sample was stratified to match and be projectable to the resident makeup of the state population in the appropriate year surveyed. Each wave is independent in that few individuals would be included in more than one survey. A research firm that specializes in collecting survey data conducted the surveys.
The initial quarterly survey (n=1,001) was conducted nine months after the first instant game was introduced, and six months from the first 3 number draw game and Lotto. One month before the survey, one individual won $55,000. The ninth wave (n=800) was collected twenty-seven months later, and one year after a five number draw game had begun. These two surveys would appear to capture the lottery in its growth and its maturity stage (Heiens 1993). It would be expected that the effects of advertising and public relations should be felt by the last survey although Heiens (1993) failed to find any effects of advertising in his examination of the Colorado lottery.
PERCENT OF ETHNIC GROUP EVER PLAYED
The questions on the survey addressed the demographic and psychographic background of each respondent, their playing of lottery games, perceptions of the games and lottery, motives for playing and memory of the lottery advertising. Not all questions were asked on each survey. The questions were pre-tested for clarity and freedom from biasing the responses.
The period that will be reported had the same top administrators and advertising agency. Of particular importance is that this state lottery adhered to a written policy of not targeting based on income, racial or ethnic affiliation, nor the use of creative appeals that disparaged work or employers. Compliance was enforced by the Florida Governors office with a Governor that is opposed to having a state lottery. All campaigns and media purchases were based on a mass marketing strategy. None of the bordering states offered a lottery product during the time period studied.
Report of Ever Playing Each Game
The previous research reports that Blacks and Hispanics tend to have higher rates of gambling in both illegal (Kallick-Kaufmann 1973; Kaplan and Blount 1990) and lottery games (Clotfeller and Cook 1989). Given these previous findings, one would expect these two ethnic groups to report greater play than whites on all games. Table 1 presents respondent self-reports of Florida state lottery games ever played (yes or no) by race and wave surveyed (1988 and 1991). Note that the Fantasy 5 game was not available (NA) until nine months after the first wave of respondents were surveyed.
All Florida lottery games cost $1, but the payoff odds and immediacy of possible payoff differ. The games are ranked in Table 1 from lowest (instant) to highest odds (Lotto). Ease and immediacy of payoff are negatively associated with the odds.
In the introductory stage, there was a significant difference in the percent of each ethnic groups participation for two of three of the lottery games when using the X2 statistic (3 ethnic groups x yes/no). However, these differences were not consistent. Whites reported the highest incidence of "ever playing" the instant game (78.4%) and were significantly different (p<.05) from Hispanics (60.9%), but not Blacks (75.0%). For Cash 3, Blacks (63.9%) reported the highest "ever playing", and were significantly different (p<. 05) from both Hispanics (43.8%) and Whites (41.2%). There were no differences between ethnic group play of Lotto, with all three groups reporting over 80% previous participation.
In the last wave, twenty-seven months later, only two of four (Fantasy 5 now available) games showed significant differences in ever played by race. In some cases, the reported play of games went down. There are no longer ethnic differences in the play of the instant game, and both Blacks and Whites report lower levels of past play. However, the levels in the two waves were not significantly different by race.
The second wave again has Blacks reporting significantly more of that group having played Cash 3 (69.6%, p< .05) than Whites (35.4%). Hispanics showed a significantly (p<.05) increased degree of play (61.1%) than when the lottery was being introduced (43.8%) and were no longer reporting less previous play than Blacks. These patterns were similar to the new Fantasy 5 game where Blacks (75.0%) and Hispanics (77.8%) were significantly more apt (p<. 05) to have played the game than Whites (51.9%). As in the first wave, there were no significant differences in ethnic group play of Lotto, with all groups reporting near 90% previous participation.
Looking across these two waves of surveys it appears that Blacks and Hispanics have increased their previous participation in the lottery games relative to Whites. Blacks reported significantly higher past participation of Cash 3 than Whites in both waves. Hispanics showed significant increases in Cash 3 and lead all groups in past Fantasy 5 play. Nonetheless, Lotto showed no ethnic differences in past participation in either wave/stage, and Whites continue to lead or equal Black and Hispanic in previous instant game play. Due to the categorical nature (yes/no) of the previous play response, the competing association of a reason for gambling model could not be determined for this measure.
Reasons for Gambling Tested
In order to determine if the various models (e.g., economic, socio-cultural, etc.) could account for the differences in play, a composite measure was arrived at for each model. Table 2 provides the questions contained in wave 2 that were used to develop the model score. The pathology model could not be tested because there were no questions contained in the survey, which were applicable to this model. All responses were on a 1 to 5 scale with 1 labeled very high priority, to 5 representing very low priority. A mean score was arrived at for each subject based on their responses to the questions contained in Table 2. Respondents were categorized as either high or low on each of the models based on the median for the sample (see Table 2).
QUESTIONS USED IN DEVELOPMENT OF MODEL
Despite the previous suggestion that ethnic group behavior toward the lottery would differ for each of the models, this was not found to be the case in 2 of the 3 models tested. Contrary to what has been previously suggested, the socio-cultural model had no significant main effect or interaction with race on any measures of lottery play. The only model that provided a significant effect in the examination of lottery play was the stimulation model. Therefore, only the results of that model will be discussed further.
Number of Tickets Bought / Usage
One method of determining usage level relative to game play is to total up the number of lottery game tickets purchased over some time period. This is the technique used by many data providers such as Simmons Market Research Services. The first wave asked respondents how many tickets they had purchased of each game since the game began (6 months). The last wave asked how many tickets of each game had been purchased in the last two weeks. Although the two waves cannot be directly compared, viewing relative differences by race in each wave is appropriate. The total number of tickets for all games combined that each respondent purchased was calculated and is also reported by race for both waves. Table 3 presents the mean number of tickets purchased by ethnic group and survey period (1988 and 1991).
Looking at the means shows that Black players reported larger numbers of tickets purchased than Whites and Hispanics for every game and total in each wave. In order to ascertain if these differences were significantly different, ANOVA was used. For the first wave, a one group (3 levels of ethnic) with age, education level and level of income were applied to the data. The last survey also provided responses that could be used to test the association of the stimulation model of gambling discussed earlier. Because this data was only on this last survey, those responses will be tested with a two group (three ethnic groups, and a high or low stimulation score) ANOVA with the same covariates noted in analyzing the first wave of number of tickets purchased data. The parametric nature of this data enables one to use ANOVA to see if the stimulation model of gambling adds a significant additional explanation beyond ethnic group game play activity. The covariates have been suggested as a significant source of variance in previous research.
Black players mean instant ticket sales (10.41) in the first survey were significantly greater (1,731 df, F=6.4, p<.01) than Whites (4.20) or Hispanics (4.45). Respondent education was a significant covariate (1,731 df, F=5.84, p<.02) inversely related with number of tickets purchased. The mean number of Cash 3 tickets purchased by Blacks (6.71) was marginally greater (p<.08) than Hispanics (4.59) and Whites (3.58), but education was again a significant covariate (1,731 df, F=6.94, p<.01). There were no other statistically significant differences associated with ethnic background.
The second survey covered the number of tickets purchased in the previous two weeks, and had responses that could test a stimulation need rationale for gambling. Whereas only one in four comparisons showed a significant main effect by ethnic group in the first survey, the second survey showed Blacks with a significantly higher number of tickets purchased in three of five comparisons. Blacks reported almost twice the number of instant tickets per player (18.06) than Hispanics (10.57) or Whites (8.32). Only the main effect of ethnic group (2,156 df, F=4.36, p<.01) was significant. There was also a main effect of ethnic group (2,640 df, F=4.53, p<.015) only on the number of Cash 3 ticets purchased and the total number of tickets for all lottery games purchased (2,377 df, F=8.72, p=.001). The responses concerning need for stimulation failed to be a significant effect in any of the analyses.
Memory of Lottery Product Advertising
Although the state of Florida lottery may not consciously target Blacks and Hispanics, it is clear that their propensity to play the game is often higher than Whites in percent group participation and mean number of tickets per group member. It may be that the advertising appeals more to relatively less affluent Blacks and Hispanics. Arguably, a poor man will be more responsive to the chance to win big than those more well off in terms of income. Therefore, it has been argued that Blacks and Hispanics would pay more attention and be more persuaded to purchase lottery products due to advertising.
The surveys asked all respondents if they had recently heard or seen any advertisements for lottery products. Although memory of advertising is one of several responses to advertising, it has often been used as a surrogate for effectiveness (e.g., DiFranza et al. 1991). If advertising was associated with game play, one would expect Blacks and Hispanics to have higher proportions of their groups having remembered some advertising. Table 4 provides the percent of each group who had reported to have "heard or seen any advertisements."
NUMBER OF TICKETS PURCHASED
PERCENT WHO HAVE HEARD OR SEEN ANY ADVERTISEMENTS
In the first survey, Blacks reported a lower level of recall (80.0%) than either Hispanics or Whites. These differences were not significant using a X2 analysis.
In the last survey, Blacks again report less recall (59.7%) than Hispanics or Whites. These last responses prove to be significantly lower (X2=6.62, 2 df, p<.04) than the other two groups (79.5% for Hispanics, 72.8% for Whites). Indeed, all groups recall rates are down from the first survey that was conducted early in the lotterys life cycle. Because Blacks tend to exhibit high levels of lottery game play, their low recall fails to support the expected favorable association of advertisement recall. Although all groups play (number of tickets purchased) appears up from the first survey to the last, ad recall is down for all groups. This pattern also fails to support the alleged influence of even non-targeted advertising.
SUMMARY AND CONCLUSIONS
Two surveys of Florida residents were used to investigate the lottery game play of Blacks and Hispanicsas compared to White players. The first survey was conducted early (6 months) in the product life cycle of the lottery. The second survey was done twenty-seven months later, in what has been described as the lotterys maturity stage (Heiens 1993). By law and official oversight, the Florida state lottery did not use advertising, sales promotion nor distribution targeted at Black or Hispanic groups during this period. This environment would appear to offer a controlled test of the association of reported game playing with the recall of non-targeted advertising estimated to cost $30 million in media for each year tested.
Contrary to earlier research, the percent of black respondents lottery participation was significantly greater than Whites in only one of the three games available in the first six months of the lottery. Generally, Hispanic play was no different than Whites except for their low instant game participation, where Whites reported the highest level of group play.
When usage is viewed in average number of tickets purchased, Blacks report the highest average number of tickets for all four measures (three games and total for all games). However, only in the instant ticket game do they show a significant difference. As was the case in participation level, the early play (number of tickets) of Blacks and Hispanics is not in line with previous reports that suggested they would be heavier users. Group saw no differences in advertising recall, as well.
The effort to see if select lifestyle orientations and values could help explain game play did not provide additional insight. Although the stimulus-seeking measure appears most promising, new measurements (Maso-Fleischman 1997) may better capture the possible influence of these
The ethnic groups memory of the lottery advertising did not positively correspond to their reported game play over the two periods. For example, the reported mean number of tickets purchased by the black respondents went up significantly, but they reported significantly lower recall of lottery product advertising. Although recall is only one measure of advertising effect (c.f., Straughn and Mizerski 1997), it is often viewed as an indicant of potential effect in public policy making agencies. There is no support for this ad effect being positively associated with gambling. At least at this point, advertising cant be blamed for increased ethnic lottery play. Ethnic targeted advertising may better reach and potentially influence Black and Hispanic groups, but their game play is high absent this emphasis.
Black Enterprise (1973), "Legalized Numbers." 3 (April): 41-5.
Burns, Alvin C., Peter L. Gillett, Marc Rubinstein, and James W. Gentry (1990), "An Exploratory Study of Lottery Playing, Gambling Addiction and Links to Compulsive Consumption," in Jerry C. Olson (ed) Advances in Consumer Research, Ann Arbor, Mi: Association for Consumer Research, 298-305.
Campbell, Felicia (1976), "Gambling: A Positive View," in Gambling and Society, William Eadington (ed), Charles Thomas: Springfield, IL.
Clotfelter, Charles T. and Philip J. Cook (1989). Selling Hope: State Lotteries in America. Cambridge, MA: Harvard University Press.
Cohen, John (1988), "Explaining Gambling Behavior," in Gambling Without Guilt: The Legitimation of an American Pastime, John Rosecreance (ed), Brooks/Cole Publishing: Pine Grove, CA.
Cook, hilip J. and Charles T. Clotfelter (1993), "The Peculiar Scale of Economies of Lotto," The American Economic Review, 83 (3), 634-643.
DiFranza, Joseph, John Richards, Paul Freeman, Nancy Wolf-Gillespie, C. Fletcher, R. Jaffe and D. Muncy (1991), "R.J.R. Nabiscos Cartoon Camel Promotes Cigarettes to Children," Journal of the American Medical Association, (Dec. 11), 3149-3153.
Edmondson, Brad (1986), "Demographics of Gambling," American Demographics, 8 (10), 39-41.
Green, Philip R. (1991). The $65 Billion Game: The Whole World Lottery Guide. Toronto Canada: World Media Brokers.
Griggs, Robyn (1994), "Political Ploys in the Lottery Game," Agency, (Fall) 12-13.
Heiens, Richard (1993), "The Influence of Newspaper, Radio and Television Publicity on Sales of the State of Colorados Lottery Product, Lotto," Unpublished Doctoral Dissertation, Florida State University.
Kallick-Kaufmann, Maureen (1979), "The Micro and Macro Dimensions of Gambling in the United States," Journal of Social Issues, 35 (3), 7-26.
Kaplan, H. Roy and William R. Blount (1990), "The Impact of the Daily Lottery on the Numbers Game: Does Legalization Make a Difference?," Journal of Gambling Studies, 6 (3), 263-274.
Karcher, Alan J. (1989), Lotteries. New Brunswick: Transaction Publishers.
Light, Ivan (1977), "Numbers Gambling Among Black: A Financial Institution," American Sociological Review, 42 (December), 892-904.
Lorenz, Valerie C. (1990), "State Lotteries and Compulsive Gambling," Journal of Gambling Studies, 6 (4), 383-396.
Maso-Fleischman, Roberta (1997), "Archetype Research for Advertising: A Spanish Language Example," Journal of Advertising Research, (September / October), 81-84.
Mcconkey, C. William and William E. Warren (1987), "Psychographic and Demographic Profiles of State Lottery Ticket Purchases," Journal of Consumer Affairs, 21 (2), 314-327.
Selinger, Iris Cohen (1993),"The Big Lottery Gamble," Advertising Age, .
Straughn, Katherine and Richard Mizerski (1997), "Deception by Omission:" Social Marketing Quarterly,
Welles, Chris (1989), "Americas Gambling Fever," BusinessWeek, April 24, 112-120.
Zuckoff, Mitchell and Doug Bailey (1993), "These days, everybodys taking the gamble," Tallahassee Democrat, October 31, p. 1A, 4A.