Content-Type: text/html

Media Effect on Race and Immigration - Page  of
Media Effect on Race and Immigration - Page  of









Media Effect on Race and Immigration:
Testing the Link











By
Cleo Joffrion Allen
Doctoral student in media and public affairs
Louisiana State University
Manship School of Mass Communication
Baton Rouge, Louisiana

1120 E. Palmview Street
Gonzales, LA 70737
225.644.5790 (home)
225.644.578.7095 (office)
225.936.4492 (cell)
<a href="/cgi-bin/wa?LOGON=A3%3Dind0209C%26L%3DAEJMC%26E%3D8bit%26P%3D887907%26B%3D--%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D_3360455%253D%253D_%26T%3Dtext%252Fplain%3B%2520charset%3Diso-8859-1" >[log in to unmask]</a>




A/V needs: Overhead projector








Abstract



Martin Gilens concludes in his book Why Americans Hate Welfare (1999) that racial stereotypes play a central role in whites' attitudes about welfare, crime, and immigration. His content analysis suggests a link between the "darkening" of poverty in news and public perceptions, but fails to empirically connect the two. I test the putative link between race and immigration using 2000 NES data - specifically, whether media use is positively correlated to racial attitudes and attitudes about immigration spending.
 Three years ago, Martin Gilens culminated years of research on race and poverty (1995, 1996a, 1996b) with his book Why Americans Hate Welfare (1999). He analyzed a wide range of surveys and experiments to explain the paradox of public opinion about the welfare state: Americans consistently say they want more government effort and higher levels of spending for almost every aspect of the welfare state. The exceptions are the means-tested programs of cash benefits - formerly Aid to Families with Dependent Children (AFDC), now Temporary Assistance for Needy Families (TANF) - and, to a lesser degree, the food-stamp program. Gilens argues that the white public is angry about these two programs because citizens view welfare as a program that rewards the undeserving poor, and they associate welfare with blacks. He says racial stereotypes play a central role in generating this opposition, and performs content analyses on visual images in newsmagazine articles and television segments to su
ggest a link between the "darkening" of poverty in the news and public perceptions. He further links whites' racial perceptions to the issues of crime and immigration.
        Gilens' analysis appears very thorough with one major exception: He fails to empirically connect public perceptions and media use. In fact, he concedes that, "Unfortunately we lack direct evidence that the proportion of African Americans pictured in news stories on poverty shapes the public's perception of the racial composition of the poor." (p. 136) Corey et al (2000) attempted to remedy Gilens' oversight by analyzing American National Election Studies data from 1988, 1992, and 1994 to explore the putative link between media use, welfare attitudes, and attitudes towards blacks. They found that media appears to be unrelated. The purpose of this paper is to examine the putative link between race and immigration using 2000 NES data. Specifically, I explore whether media use is positively correlated to racial attitudes and attitudes about spending to control illegal immigration. I propose an amended replication of  the model of Corey et al in which attitudes on spending to control
illegal immigration is a function of attitudes towards blacks, media usage, an interaction variable for black attitudes and media usage, a media-black interaction variable, and other control variables.

Literature review
Using data from the 1991 National Race and Politics study and others, Gilens shows that while blacks are a minority of both poor people (27 percent) and of welfare recipients (36 percent), the average American believes that blacks make up half of all the poor. The average American also thinks blacks comprise 23 percent of the total population, nearly twice the actual figure of 12 percent. As for welfare, he says, the popular view is that the rolls are bloated with undeserving recipients: Two of three Americans say most people on welfare are taking advantage of the system; only one in three believes most recipients really need help. Gilens offers a causal model with 11 predictors of welfare spending preferences: age, sex, region, education, marital status, family income, ideological identification, party identification, individualism, the perception of blacks as lazy, and the perception of welfare recipients are undeserving. The strongest predictor, he suggests, are whites' attitudes about blacks' work ethic, or the perception that blacks are lazy.
Gilens traces changes in popular images of the poor on network news and in Time, Newsweek, and U.S. News and World Report magazines from the 1950s to the 1990s  to suggest the link between changes in whites' attitudes about the poor and media coverage. While scientists began studying poverty in the late 19th century, Gilens says, blacks were ignored for the most part in research until the 1960s and were excluded in great numbers from assistance rolls in many states through a "suitable home" clause. Even when President John F. Kennedy implemented several antipoverty programs, Gilens says the dominant image of poverty was that of the rural Appalachian poor white. Poverty came to be racialized and popular images of poverty changed as the result of two trends, he says. One was the widespread migration of rural Southern blacks to the North. The second was the changing racial composition of AFDC as President Lyndon Johnson implemented his War on Poverty and civil rights leaders pushed f
or economic justice in the second half of the 1960s.[1] In an arresting chart based on his results, Gilens shows the poverty rate declining for the most part from 1950 to 1990 as the number of poverty stories increase, "darken," and become more negative.
From the beginning of the content analysis through 1964, most visual images of the poor were white. From 1964 to 1966, the proportion of blacks' pictures in poverty stories increased to 27 percent to 53 percent, then to 72 percent in 1967. Over 45 years through 1992, an average of 57 percent of the pictures in poverty stories were black, nearly twice as many as the true proportion. Perhaps more importantly, pictures of the black poor dominate when poverty coverage is most negative while the more sympathetic coverage uses pictures of nonblacks, Gilens' research indicates.
Gilens also reports the results of a "survey experiment" in which respondents were given a scenario of a hypothetical single mother welfare recipient; some were told the recipient was white, some black. The responses indicated the black version of the experiment was a greater predictor of welfare opposition that the white version.
Gilens is not the first to make such a claim about "race-coded" topics, though Reese (2001) says few have examined the issue in such systematic detail. Quadagno (1994) looks at race-coding from a historical perspective to suggest the link between race and social policy. Like Gilens, she says other social programs also have been racialized. Gilens mentions crime and immigration. Quadagno cites urban renewal, job training, and school choice as eliciting similar connotations.
Sears et al (1997) offer the general perspective of symbolic politics theory, which assumes people have strong, longstanding predispositions to certain attitudes because of socialization that can be evoked by appropriate political symbols. This perspective has largely displaced the "old" racism that was based on the idea of biological superiority. Instead, "presenting whites with racially targeted policies or black candidates should evoke that common antiblack element" (1997, p. 18). Three concepts are involved: racism is phrased in abstract and ideological terms (Sears and Kinder, 1971); it includes the beliefs that blacks should just work harder since  racial discrimination is largely a thing of the past (Sears, 1988); and the perception of blacks' lack of work ethic blends with an antiblack effect (Kinder and Sears, 1981). Using four surveys over a decade in three policy areas, Spears et al found that controls on ideology don't weaken the effects of symbolic racism, and ideolog
y had insignificant effects in general. They found the effect stronger than nonracial dispositions even among the college-educated. In direct opposition, however, are Sniderman and Piazza (1993), who contend that prejudice no longer is the dominant problem of racial politics, but rather ideology and other nonracial considerations.
Gilens' thesis is in line with Gerbner's cultivation theory, suggesting strong media effects. Other experimental research has found race to be a predictor of frames (Kinder and Sanders, 1996, and Valentino, 1999). Previous studies have shown that visual elements in news are highly salient (Graber, 1990; 1987) and that the race of the pictured individual is a salient visual cue (Iyengar, 1991; Iyengar; 1987; Iyengar and Kinder, 1987). Iyengar and Kinder say the power of television rests on agenda-setting and "priming," or "defining criteria underlying the public's judgment" (p. 117). The two say television news is powerful (but limited) in setting some viewers' standards for judgments by priming certain aspects of national life while ignoring others. Iyengar also says individuals are influenced by how news presentations "frame" issues. He says, "When poverty is framed as a societal outcome, people point to societal or governmental explanations; when poverty is framed in terms of pa
rticular victims of poverty, particularly the homeless, people point instead to dispositional explanations." The findings of other researchers, including Domke (1999) and Pan and Kosicki (1996), support this view as well.
Entman (1995, 1994, 1992, 1990) defines framing as "selecting and highlighting some elements of reality and suppressing others, in a way that constructs a story about a social problem, its causes, its moral nature, and its possible remedies" (p. 142). Entman explores the concept of modern racism by examining television news and suggests that the way blacks are covered may have a different impact on whites than it has on blacks. He found nearly 60 percent of network news stories centered on negative news about blacks. In his latest work (2001), he describes the pattern of crime and sports coverage of blacks in national news and blacks as victimizers and whites as victims in local news. Similarly, Van Dijk (1988) found that the media portrayals indirectly favor the stereotypical frameworks of interpretation. In a study examining newspaper articles about poverty and welfare since the 1996 implementation of TANF, Bullock et al (2001) found most articles to be neutral in tone, but with
 little to contextualize poverty or illuminate its causes.
But when it comes to gauging effects, results vary by methodology. Bartels (1993) notes the most convincing demonstrations of media exposure effects come from laboratory experiments, but concedes the method is limited with respect to external validity. On the other hand, nonexperimental literature on media effects usually report negative findings, he says. Using survey data, Bartels himself found media effects to be modest when adjusted for the effects of measurement error and that media exposure only occasionally produces strong, unidirectional opinion changes.

Race and Immigration
        America has seen four major waves in immigration (Pedraza, 2000): the first of northwest Europeans in the mid-19th century; southern and eastern Europeans from the end of the 19th century to the beginning of the 20th century; the migration from South to North of African-Americans, Latinos and native Americans over the two world wars; and immigrants from Latin America, from 1965 to the present.[2] What essentially had been an open-door policy to immigrants gradually closed, starting with barring the Chinese in 1882, Asian Indians in 1917, Japanese in 1924, and Filipinos in 1934 (Lowe, 1996). King (2000) uses archival materials to suggest "the finely filtered regime of selection" (p. 1) was based on racial quotas and eugenic categories. Timmer and Williams (1998), who evaluated immigration policies in six countries, said the ethnic composition of the immigrants was clearly a factor in the politics of restriction, not only in America, but in other Western countries as well.
Yet, typically the research on immigration policy has split into two perspectives, with the dominant being political economy (Portes, 1997; Zolberg, 1981, 1987, 1989), or as New and Petronicolos  put it, "reduce(d)_to simple cost-benefit analysis" (1997, p. 1) of issues like labor supply versus wages and limited welfare provision. The alternative is the examination of the sociology of the political system, in which race obviously plays a role.[3] Li et al (2001) found support for both sides - national economic concerns and prejudice and racism - in an experiment examining illegal immigration in the aftermath of California's Proposition 187.
The Immigration and Naturalizaton Act of 1965 (revised in 1986 and 1990), with its emphasis on family reunification, resulted in a major shift in the point of origin for immigrants away from Europe and towards Latin America and Asia. Each year, about 800,000 people immigrate legally to the United States, and at least another 200,000 illegal immigrants stay, according to the National Research Council (Smith et al, 1997). In 1990, 43 percent of immigrants came from Latin America and the Caribbean, 26 percent from Canada and Europe; 25 percent from Asia, and 6 percent from other countries. The report, commissioned by the U.S. Commission on Immigration Reform, estimates that immigration will account for two-thirds of U.S. population growth by 2050 if the current level continues, with Hispanics increasing to 25 percent of the total population and replacing blacks as the largest minority group. Eight percent is expected to be Asian.
Sanchez (1999) says immigration historians have started to recognize the critical role of race in "facilitating the adaptation of certain European newcomers to American society" (p. 1272), impelled by the growing field of "whiteness" studies (Roediger, 1991; Barrett and Roediger, 1997; Rogin, 1996; Sacks, 1994; King, 2000). Noting the relational concept of race rather than a biological or cultural one, these researchers contend that European immigrants assimilated by positioning themselves as "white." King says debates leading up to 1929 were couched in terms of desirable and undesirable immigrants, and immigrants such as the Irish and Jews eventually entered the labor market by exploiting whiteness as a source of power. Hochschild (1995) says these transformations were taking place as late as the 1920s: "(D)escendants of old-stock immigrants thought of southern and eastern European immigrants as a different race. But that language disappeared over the next few decades, in favor o
f an increasingly general category of 'white' or 'American' " (p. 243).
Although a few content analyses have been performed of news content about immigration (Hufker and Cavender, 1990; Miller, 1994; Simon and Alexander, 1993), Domke notes a gap in scholarship related to the link between mass media  and race. Indeed, an issue devoted to  "Immigration and Immigration Research in the United States" in the American Behavioral Scientist (1999) through the Social Science Research Council's International Migration Committee offers no view of the role of media or race in its eight studies.
Methodology
        Data from the 2000 American National Election Studies were analyzed for this paper. The NES, conducted every two years by the Center for Political Studies of the Institute for Social Research at the University of Michigan, provides individual-level data from a cross-section of the electorate in pre- and post-election interviews on a variety of topics. Because Gilens' thesis related to whites' beliefs about welfare, only the responses from whites, 1,393 cases, were analyzed.
        I am replicating the model of Corey et al (with variations) in which attitudes about spending to control illegal immigration are a function of attitudes towards blacks, media usage, an interaction variable for black attitudes and media usage, a media-black interaction, and control variables. I hypothesize that respondents' feelings towards blacks should be inversely and strongly related to their views about spending to control illegal immigration. Additionally, if Gilens is correct, media use should be positively related to attitudes towards immigration control spending. The coefficient for the interaction variable between media and attitudes towards blacks should be positive.
        Demographic and political attitude measures are included as controls in the models. Demographic variables include education, family income, age, and gender, and include the lifestyle characteristics of marital and homeowner status. Political variables include party identification, ideological intensity, political trust, political efficacy, and government activity (support for public-sector activity over free-market activity). I expect these variables will have significant effects on attitudes about spending to control illegal immigration.
Dependent variable
        The respondent's attitude towards spending to control illegal immigration is the dependent variable, measured on a three-point scale. The question is "federal spending to tighten border security to prevent illegal immigration" should be  increased, coded as 1; kept the same, 0; or decreased or cut out, coded as -1. The dependent variable is an ordered, three-point scale, so ordinal regression (ordered logit) was used for this analysis.

Independent variables: Media and blacks
        Five independent variables are the focus of this analysis: Two questions on attitudes towards blacks, media use, and two interaction variables are designed to capture the effect of media on feelings towards blacks. The black feeling thermometer is measured on a scale that ranges from 0 to 100, with high scores representing positive feelings towards blacks. I hypothesize that a negative relationship will be seen between attitudes towards blacks and attitudes towards spending to limit illegal immigration; as attitudes towards blacks decline, the desire for more spending to limit illegal immigration should go up.
        A second attitudinal question was used on whether blacks are "hard-working" or "lazy," similar to a measure used by Gilens, who argued that whites' attitudes on welfare depend on whether they believe the clients are "deserving" of help or "lazy." The variable ranges from 0 ("Blacks are lazy") to 6 ("Blacks are hard-working"). Because whites appear to link welfare to blacks, the attitude on blacks' work ethic should be negatively related to the desire to increase spending to limit illegal immigration.
        Media is an additive scale of media usage, ranging from 0 (no media usage) to 14 (high media usage), based on responses to the questions: "How many days per week do you read the newspaper? Watch television news?" Additionally, the media/black interaction variable is designed to capture the effect of feelings towards blacks as media use increases. Finally, the media/hard-working interaction is designed to capture the effect of perceptions of blacks as hard-working as media use increases. I hypothesize negative relationships will exist between these three variables and the desire to increase spending to limit immigration.

Independent variables: Demographic characteristics
        Six demographic and lifestyle variables are included as controls to measure preexisting attitudes toward immigration: education, family income, age, gender, marital status, and homeowner status. Education is measured as the highest level of schooling completed, ranging from 1 (eight grades or less) to 7 (advanced degree).  Family income is measured on a 21-point scale, ranging from 1 (income below $4,999) to 21 ($200,000 per year or more). Age is measured in years, ranging from 18-97, with 97 representing 97 years or more. Gender is a dichotomous variable, with women coded as 1 and men as 0. Marital status is coded as 1 for married and 0 for all other respondents. Homeowner status is measured dichotomously, with 1 for homeowners and 0 for all other respondents. I expect a negative relationship between two of these and the independent variable - education and females - since education should have a liberalizing impact and because females are expected to have more liberal views. I expect positive relationships between the dependent variable and family income and homeowner status because of a possible desire to avoid additional taxation. I predict positive relationships between age and marital status and the dependent variable since conservatism tends to increase with age and when one weds.

Independent variables: Political characteristics
Political variables include party identification, ideological intensity, political trust, political efficacy, and support for public-sector activity over free-market activity. Party identification and ideological intensity should be strong predictors of  immigration attitudes since conservatives and Republicans are more apt to want higher spending to limit illegal immigration; I expect a positive relationship. Party identification is measured on a seven-point scale of 0 for strong Democrat to 6 for strong Republican. Ideological identification is measure on a seven-point scale, with 0 denoting strong liberal to 6 for strong conservative.
Four questions were combined through factor analysis to create the political trust variable: 1) How much of the time do you think you can trust the government in Washington to do what is right? 2) Do you think that people in government waste a lot/some/not very much of the money we pay in taxes? 3) Would you say the government is pretty much run by a few big interests looking out for themselves or that it is run for the benefit of the people? 4) Do you thing that quite a few/not very many/hardly any of the people running the government are crooked? The single variable resulting from factor analysis has an eigenvalue of 1.937, with a variance explained of 0.484. I hypothesize that confidence in government's ability to allocate resources will be positively related to support for increased spending to limit illegal immigration.
The political efficacy variable combines another four questions from the NES survey: 1) How much attention do you feel the government pays to what people think when it decides what to do? 2) How much do you feel that having elections makes the government pay attention to what people think? Respondents also were asked to rate their level of agreement with the following statements: 3) Public officials don't care much what people like me think. 4) People like me don't have any say about what the government does. The responses were coded as 0, a good deal; 1, some; and 2, not much. The eigenvalue of the combined variable is 2.187, with a variance explained of 0. 547. I hypothesize that higher efficacy levels will be positively related to immigration spending attitudes.
Finally, a factor analysis of three questions provides one variable to measure general attitudes towards government activity. Respondents were asked to choose between: 1) The less government the better vs. Government should do more. 2) We need a strong government to handle today's complex economic problems vs. The free market can handle these problems without government involvement. 3) The main reason government has become bigger over the years is because it has gotten involved in things that people should do for themselves vs. Government has become bigger because the problems we face have become bigger. The eigenvalue is 1.983, with a variance explained of 0.661. I hypothesize that support for an active government will be positively related to attitudes on spending to limit illegal immigration.
A significant amount of the variance in attitudes about spending to limit illegal immigration should be accounted for in models with these variables. Testing Gilens' conclusions, media use should have a negative effect on the link between attitudes towards blacks and attitudes towards spending to limit illegal immigration.

Results/Discussion
        Table 1 provides the results for the baseline model of support for spending to limit illegal immigration as a function of attitudes towards blacks, media usage, and control variables, excluding the two interaction effects. The pseudo-R2, or the coefficient of determination, which represents the fit of the model to the data, is 0.121. Thus, 12.1 percent of the variance in support for spending to limit illegal immigration is explained by the model.
        The intercept, or a, is -2.184, with a t of -4.051. It is highly significant at the 0.000 level. The intercept is the predicted value on spending to limit illegal immigration when all of the independent variables are zero.
        The slope coefficient - indicating the relative impact of the independent variables on the dependent - for  the black feeling thermometer is significant at the more modest level of 0.10 (b = -0.051, t = -1.319), but the hard-working blacks variable is highly significant at the 0.01 level (b = -0.168, t = -2.464). Both relationships are negative, which supports my hypotheses. This suggests that attitudes toward blacks are predictive of attitudes about spending to control illegal immigration. However, the media variable is not significant (b = -0.046, t = -0.526) although the direction is negative as expected, offering no support for Gilens' link of media to black attitudes and those on immigration spending. All are in the expected direction.
        Three other control variables are in the expected direction and highly significant - education, age, and ideological intensity (liberal-conservative) - at the 0.01 level and one, party identification, is moderately significant at the 0.05 level. Education has a slope coefficient of       -0.257 and a t of -4.724. Age has a slope of 0.019 and a t of 3.352. And liberal-conservative's slope is 0.213 with a t of 3.239. Party identification has a b of 0.072 and a t of 1.583. Thus, whites with less education, more income and those who are older are significantly more supportive of anti-immigration spending.
        The findings for three other control variables were in the opposite direction from that hypothesized, but the results on these three were not significant, at any rate. Gender was positive while I hypothesized a negative relationship, indicating females were not as liberal as I had supposed. The homeowner variable results were negative when I suggested positive, suggesting those who did not own their residences were more supportive of anti-immigration spending. And the direction of the political efficacy variable was negative, when I had predicted positive.
        Table 2 includes the two media interactions to no avail. Neither the media-black interaction nor the media-hard-working interaction is significant (b = 0.085, t = 0.203 for the former; b = 0.023, t = 0.032 for the latter). Additionally, the two interaction variables are in unexpected directions; the findings show a positive direction when I had hypothesized a negative relationship. Media usage alone also remains insignificant although the direction is negative as expected.
The significant variables remain the same as before - education, age, and ideological intensity - at the 0.01 level and party identification at the 0.05 level. All are in the predicted direction.
Once again, gender and political efficacy had results in the opposition direction from the hypotheses. Gender's relationship was positive when I expected negative, and efficacy the reverse.
        Table 3 shows the parameter estimates of comparisons between high and low media use. This model offers a slightly different view and a better goodness of fit. The pseudo-R2, or coefficient of determination, is 0.136 for high media use and 0.179 for low media use, indicating 13.6 percent and 17.9 percent of the variation of y are explained respectively. These are higher percentages than were found in Tables 1 and 2. Some variables are significant for high media use, but not for low media use, and vice versa. Only age and ideological intensity show up as significant in both high and low media use while media is significant in neither.
With high media use, education and age were found to be in the expected direction and highly significant at the 0.01 level (b = -0.294, t = -2.978; and b = 0.027, t = 2.800). Three other variables were significant at the more relaxed 0.10 level: hard-working blacks (b = -0.203, t = -1.595, and in the expected negative direction); liberal-conservative (b = 0.151, t = 1.291, and in the expected positive direction); and political trust (b = 0.255, t = 1.590, and in the expected positive direction).
It should be noted that political trust changed to a negative direction in low media use, although the finding was not significant. Another direction switch occurred with gender, in which the finding for low media use was negative as expected, but in the opposite direction for high media use. These findings also were not significant, however. Finally, the direction for political efficacy in both high and low media use was opposite of the positive relationship hypothesized, although the findings were not significant, either.
Neither the black feeling thermometer nor the media results was significant in either the high or low media use categories. All the relationships were negative as expected, however.
With low media use, education remained highly significant (0.01) with a b of -0.353 and a t of -3.344, and in the expected direction. Three others are moderately significant at the 0.05 level: family income (b = 0.101, t = 1.971, and positive direction as expected); homeowner (b = 0.550, t = 1.668, and positive as expected); and ideological intensity (liberal-conservative)  (b = 0.234, t = 1.880, and positive as expected). Party identification results were in the expected positive direction and modestly significant (0.10) with a b of 0.127 and a t of 1.391.
Realizing that many believe newspaper readers to be different in characteristics that television viewers, I decided to test a fourth model separating the two media, with the hypothesis that television use would be negatively related while newspaper use would be positively related. This expectation is based on the common view that regular newspaper readers are generally more well-educated and elite than are television viewers. The results are shown in Table 4.
The pseudo-R2, or coefficient of determination, is 0.122, meaning that 12.2 percent of the variation of y is explained by the model. (This is basically the same as in Tables 1 and 2.)  However, while five variables are in the highly significant range in this model and one is modestly significant, all in the expected direction, neither television usage nor newspaper usage is significant at all. Party identification is modestly significant (0.10) with a b of 0.075 and a t of 1.640, with a positive relationship. Education (b = -0.252, t - -4.618) was highly significant at 0.000, with the expected negative relationship. Both the black feeling thermometer and the hard-working blacks variable are negatively related and significant at 0.01 (b = -0.054, t = -1.369; and b = -0.168, t = -2.462, respectively. Two others also were in the expected directions and highly significant at 0.01: age (b = 0.019, t = 3.395), and ideological intensity (b = 0.211, t = 3.203).
Yet no significant relationships were found between either television usage nor newspaper usage and attitudes on anti-immigration spending. Additionally, a positive relationship was found with newspaper use when I predicted a negative direction. And television usage, which is expected to attract more plebian tastes, was in the unexpected positive direction.
Once more, the variables for gender, home ownership, and political efficacy had findings in unexpected directions. Thus, males, non-homeowners, and those who presume less political efficacy were found to be more supportive of anti-immigration spending. But none of the three relationships was significant.


Conclusions
Ordered regression analysis (ordered logit) was performed on variables taken from the 2000 American National Election Studies to test Gilens' claim that media use is positively correlated to racial attitudes and attitudes about immigration, using an NES question on spending to control illegal immigration as the dependent variable. Attitudes about blacks were found to be significant in three of the four models on spending to limit illegal immigration, lending support for Gilens' position that negative attitudes about blacks are a strong predictor of "race-coding" issues such as welfare, crime, and immigration. The most consistent control predictors beyond black affect in all models were education, age, and ideological intensity. White citizens with less education, who are conservative in their political outlook, and older were found to be more supportive of increased spending to control illegal immigration.
However, media use was not found to be significant, whether tested separately (television and newspaper); jointly (media); between high and low users; or as an interaction with attitudes about blacks. I did find some difference in the directions of the relationships between television use and newspaper use, but these differences were not significant, either. Thus, I find fail to find support for Gilens' link of media portrayals of blacks to black attitudes and attitudes on "race-coded" issues.
I must concede that the question on immigration spending might not have been the best possible choice as a dependent variable, but I was limited to what had been used in a national study. That question was deemed to be the better of the two on immigration in the latest NES study. Subsequent research could be attuned to finding a better national survey question choice. Additionally, more finely tuned media-use questions could be used.
Despite the evidence from previous researchers of the priming effect of television news and the correlated movement over time between attitudes about blacks and attitudes about welfare (and theoretically, to other "race-coded" issues), I, like others, have been unable to find direct evidence linking these characteristics with media use. This is not the final word on the subject, however. The key is probably in refinement of the survey instrument. Sears et al use factor analysis to look at equal opportunity, federal assistance, and affirmative action. Perhaps we should do the same with immigration, welfare, and crime. Further study also should include the Hispanic feeling thermometer since that group is the fastest-growing minority population, or perhaps a factor analysis of the Hispanic and black feeling thermometers. The possibility also cannot be overlooked that Bartels may be correct in saying that experiments are a better tool to measure effects. Perhaps, to extend his argumen
t, an experiment is needed to seek the direct evidence as well.
 Appendix 1: Description of variables



Variable                                        Description


(Dependent)
Immigration control attitudes   Measured on the following scale: Federal spending to tighten border security to prevent illegal immigration should be increased, 1; kept the same, 0; or decreased or cut out, -1.

(Independent)
Black feeling thermometer       Measured on a scale that ranges from 0 to 100, with high scores representing positive feelings toward blacks.

Media/black interaction Media * black feeling thermometer; designed to capture the effect of feelings towards blacks as media usage increases.

Blacks are hard-working Seven-point scale, ranging from 0 (blacks are lazy) to 6 (blacks are hard-working).

Media/hard-working interaction  Media * Blacks are hard-working; designed to capture the effect of perceptions of blacks as hard-working as media use increases.

Media   Additive scale of media usage, ranging from 0 (no media usage) to 14 (high media usage) based on responses to "How many days per week do you read the newspaper? Watch television news?"

Television use  Additive scale of television usage, ranging from 0 to 7 (days per week).

Newspaper use   Additive scale of newspaper usage, ranging from 0 to 7 (days per week).

Education       Highest level of schooling completed, ranging from 1 (eight grades or less) to 7 (advanced degree).

Family income   A 21-point scale of family income, ranging from 1 (income below $4,999 per year) to 21 ($200,000 per year or more).

Age     Age in years, ranging from 18-97, with 97 representing 97 years or more.


Gender                                          1 = women; 0 = men.

Homeowner                                       1 = homeowner; 0 = all other respondents.

Married                                 1 = married; 0 = all other respondents.



Variable                                        Description

Government activity scale       Scale of support for government activity over the private sector or market activity, based on a factor analysis of responses to the following questions: 1) support for government doing more vs. government doing less; 2) support for free-marking handling complex economic problems vs. strong government
handling those problems; 3) whether government has become bigger over the years because it is involved in things that people should do themselves vs. government has gotten bigger because the problems we face have become bigger. (Eigenvalue = 1.983; variance explained = 0.661).

Partisan Identification Measured on a scale from 0 (strong Democrat) to 6 (strong Republican).

Ideological identification      Measured on a scale from 0 (strong liberal) to 6 (strong conservative).

Political efficacy      Scale of feelings of political efficacy, based on a factor analysis of responses to the following items: 1) How much attention do you feel the government pays to what people think when it decides what to do? 2) How much do you feel that having elections makes the government pay attention to what people think? Additionally, respondents were asked to rate their level or agreement or disagreement with the following statements: 3) Public officials don't care
much what people like me think. 4) People like me don't have any say about what the government does. (Eigenvalue = 2.187; variance explained =0.547).

Political trust Scale of feelings of political trust, based on a factor analysis of the following items: 1) How much of the time do you think you can trust the government in Washington to do what is right? 2) Do you think that people in government waste a lot/some/not very much of the money we pay in taxes? 3) Would you say that the government is pretty much run by a few big interests or that it is run for the benefit of all the people? 4) Do you think that quite a few/not very many/hardly any of the people running the government are crooked? (Eigenvalue = 1.937; variance explained = 0.484).



 Table 1. Parameter estimates for 2000 model of immigration attitudes without interactions



Variable                                b               se              t               Prob.


Black feeling thermometer [-]           -0.052          0.004           -1.319*         0.094
Blacks are hard-working [-]             -0.168          0.068           -2.464***       0.007

Media [-]                               -0.046          0.087           -0.526          0.299

Demographic/lifestyle characteristics
Education [-]                   -0.257          0.054           -4.724***       0.000
Family income [+]               0.017           0.023           0.735           0.232
Age [+]                         0.019           0.006           3.352***        0.001           Gender [-}                      0.137           0.153           0.897           0.185
Homeowner [+]                   -0.073          0.185           -0.392          0.348
Married  [+]                    0.153           0.168           0.910           0.181

Political attitudes and characteristics
        Government activity [+]         0.109           0.090           1.210           0.113
        Party identification [+]                0.072           0.046           1.583*          0.057
        Liberal-conservative [+]                0.213           0.066           3.239***        0.001
        Political efficacy [+]          -0.019          0.088           -0.217          0.414
        Political trust [+]             0.063           0.086           0.729           0.233


Constant                                -2.184          0.539           -4.051***       0.000
N = 758
Pseudo-R2 = 0.121
Chi-Square = 97.349
Prob Chi-Square = 0.000

***     prob. ( 0.01 (one-tailed test)
**      prob. ( 0.05 (one-tailed test)
*       prob. < 0.10 (one-tailed test)
 Table 2. Parameter estimates for 2000 model of immigration attitudes with interaction effects




Variable                                b               se              t               Prob.


Black feeling thermometer [-]           -0.052          0.004           -1.304*         0.096
Blacks are hard-working [-]             -0.168          0.068           -2.467***       0.007

Media [-]                               -0.108          0.291           -0.373          0.355

Demographic/lifestyle characteristics
Education [-]                   -0.257          0.055           -4.705***       0.000
Family income   [+]             0.017           0.023           0.730           0.233
Age [+]                         0.019           0.006           3.354***        0.001           Gender [-]                      0.135           0.153           0.881           0.189
Homeowner [+]                   0.073           0.185           0.395           0.347
Married  [+]                    0.153           0.169           0.910           0.182

Political attitudes and characteristics
        Government activity [+]         0.109           0.090           1.215           0.112
        Party identification [+]                0.073           0.046           1.589*          0.056
        Liberal-conservative [+]                0.214           0.066           3.241***        0.001
        Political efficacy [+]          -0.020          0.088           -0.224          0.412
        Political trust [+]             0.062           0.086           0.724           0.235

Interactions
        Media-black interaction [-]     0.085           0.004           0.203           0.420
        Media-hardworking
Interaction [-]         0.023           0.071           0.032           0.488

Constant                                -2.177          0.541           -4.026***       0.000
N = 758
Pseudo-R2 = 0.121
Chi-Square = 97.405
Prob Chi-Square = 0.000

***     prob. ( 0.01 (one-tailed test)
**      prob. ( 0.05 (one-tailed test)
*       prob. < 0.10 (one-tailed test)
 Table 3. Parameter estimates for 2000 model of immigration attitudes with high and low media use



                                High Media Use          Low Media Use
Variable                        b               t               b               t


Dependent variable
        Attitudes about spending to control illegal immigration

Intercept

Black feelings [-]      -0.039          -0.495                  -0.046          -0.670
Blacks hard-workg [-]   -0.203          -1.595*                 -0.069          -0.564

Media [-]               -0.285          -0.893                  -0.283          -0.637

Demographic/lifestyle characteristics
Education [-]           -0.294          -2.978***               -0.353          -3.344***
Family income [+]       0.042           1.106                   0.101           1.971**
Age [+]                 0.027           2.800***                0.068           0.485
Gender [-]              0.178           0.645                   -0.136          -0.447
Homeowner [+]           0.023           0.063                   0.550           1.668**
Married  [+]            0.286           0.932                   0.079           0.032

Political attitudes and characteristics
Party ID [+]            0.054           0.669                   0.127           1.391*
Liberal-conserv  [+]    0.151           1.291*                  0.234           1.880**
Political trust [+]     0.255           1.590*                  -0.105          -0.650
Political efficacy [+]  -0.085          -0.055                  -0.125          -0.748
Gov't activity [+]      0.158           0.962                   0.035           0.210



N                                       259                             226
Pseudo-R2                               0.136                           0.179



***     prob. ( 0.01 (one-tailed test)
**      prob. ( 0.05 (one-tailed test)
*       prob. < 0.10 (one-tailed test)

 Table 4. Parameter estimates for 2000 model of immigration attitudes with newspaper and television use



Variable                                b               se              t               Prob.


Black feeling thermometer [-}           -0.054          0.004           -1.369***       0.009
Blacks are hard-working [-]             -0.168          0.068           -2.462***       0.007

Television use  [-]                     0.015           0.029           0.524           0.300
Newspaper use   [+]                     -0.036          0.028           -1.272          0.102

Demographic/lifestyle characteristics
Education [-]                   -0.252          0.055           -4.618***       0.000
Family income [+]               0.017           0.023           0.745           0.228
Age [+]                         0.019           0.006           3.395***        0.001           Gender [-]                      0.122           0.153           0.794           0.214
Homeowner [+]                   -0.057          0.185           -0.310          0.378
Married  [+]                    0.138           0.169           0.820           0.206

Political attitudes and characteristics
        Government activity [+]         0.112           0.090           1.243           0.107
        Party identification [+]                0.075           0.046           1.640*          0.051
        Liberal-conservative [+]                0.211           0.066           3.203***        0.001
        Political efficacy [+]          -0.028          0.088           -0.311          0.378
        Political trust [+]             0.076           0.086           0.884           0.188


Constant                                -2.254          0.516           -4.368***       0.000
N = 758
Pseudo-R2 = 0.122
Chi-Square = 98.863
Prob. Chi-Square = 0.000

***     prob. ( 0.01 (one-tailed test)
**      prob. ( 0.05 (one-tailed test)
*       prob. < 0.10 (one-tailed test)


 References

Barrett, J.R., and Roediger. D. (1997). Inbetween peoples: race, nationality and the 'new immigrant' working class. Journal of American Ethnic History, 16, no. 3, 3-44.
Bartels, L.M. (1993). Messages received: the political impact of media exposure. American Political Science Review, 87, no. 2, 267-285.
Bullock, H.E., Wyche, K.F., and Williams, W.R. (2001). Media images of the poor. Journal of Social Issues, 57, no. 2, 229-246.
Corey, E.C., Garand, J.C., and Plaisance, K.A. (2000). The linkage between racial attitudes toward welfare spending: Are the media responsible? Paper presented at the annual meeting of the Southern Political Science Association, Atlanta, Georgia, November 8-11.
Domke, D., McCoy, K., and Torres, M. (1999). News media, racial perceptions, and political cognition. Communication Research, 26, no. 5, 570-607.
Entman, R.M. (2001). The Black Image in the White Mind: Media and Race in America. Chicago: University of Chicago Press.
Entman, R.M. (1994). Representation and reality in the portrayal of blacks on network television news. Journalism Quarterly, 71, no. 3, 509-20.
Entman, R.M. (1992). Blacks in the news: television, modern racism and cultural change.  Journalism Quarterly, 69, no. 2, 341-61.
Entman, R.M. (1990). Modern racism and the images of blacks in local television news. Critical Studies in Mass Communication, 7, no. 4, 332-45.
Entman, R.M. (1994). Television, democratic theory, and the visual construction of poverty. Research in Political Sociology, 7, 139-60.
Gilens, M. (1996). Race and poverty in America: public misperceptions and the American news media. Public Opinion Quarterly, 60, no. 4, 515-41.
Gilens, M. (1995). Racial attitudes and opposition to welfare. Journal of Politics, 57, no. 4, 994-1014.
Gilens, M. (1996). 'Race coding' and white opposition to welfare. American Political Science Review, 90, no. 2, 593-604.
Gilens, M. (1999). Why Americans Hate Welfare. Chicago: University of Chicago Press.
Graber, D. (1990). Seeing is remembering: how visuals contribute to learning from television news. Journal of Communication, 40, no. 3, 134-55.
Graber, D. (1987). Television news without pictures? Critical Studies in Mass Communication, 4, 74-78.
Hochschild, J. (1995). Facing Up to an American Dream. Princeton, New Jersey: Princeton University Press.
Hufker, B., and Cavender, G. (1990). From the freedom flotilla to America's burden: the social construction of the Mariel immigration. Sociological Quarterly, 31, no. 2, 321-35.
Immigration and immigration research in the U.S. A special topic of the June/July 1999 issue. American Behavioral Scientist, 42, no. 9.
Iyengar, S. (1991). Is Anyone Responsible? How Television Frames Political Issues. Chicago: University of Chicago Press.
Iyengar, S. (1987). Television news and citizens' explanations of national affairs. American Political Science Review, 81:3, 815-32.
Iyengar, S., and Kinder, D.R. (1987). News That Matters: Television and American Opinion. Chicago: University of Chicago Press.
Kinder, D.R., and Sanders, L.M. (1996). Divided by Color: Racial Politics and Democratic Ideals. Chicago: University of Chicago Press.
King, D. (2000). Making Americans: Immigration, Race, and the Origins of the Diverse Democracy. Cambridge, Massachusetts: Harvard University Press.
Lee, Y.T., Ottati, V., and Hussain, I. (2001). Attitudes toward "illegal" immigration into the United States: California Proposition 187. Hispanic Journal of Behaviorial Sciences, 23, no. 4, 430-43.
Lowe, L. (1996). Immigrant Acts: On Asian American Cultural Politics. Durham, N.C.: Duke University Press.
Miller, J. (1994). Immigration, the press, and the new racism. Media Studies Journal, 8, 19-28.
Muller, T., and Espenshde, T.J. (1985). The Fourth Wave: California's Newest Immigrants. Washington: Urban Institute.
Neuman, W.R., Just, M.R., and Crigler, A.N. (1992). Common Knowledge: News and the Construction of Political Meaning. Chicago: University of Chicago Press.
New, W.S., and Petronicolos, L. (1997). Rereading the record: the rhetoric of anti-immigration legislation and education. Paper presented at the annual meeting of the American Educational Research Association, Chicago.
Pan, Z.D., and Kosicki, G.M. (1996). Assessing news media influences on the formation of whites' racial policy preferences. Communication Research, 23, no. 2, 147-178.
Pedraza, S. (2000). Beyond black and white: Latinos and social science research on immigration, race, and ethnicity in America. Social Science History, 24, no. 4, 697-726.
Pedraza, S., and Rumbaut, R.G., (eds.). (1996). Origins and Destinies: Immigration, Race, and Ethnicity in America. Belmont, California: Wadsworth Press.
Portes, A. (1997). Immigration theory for a new century: some problems and opportunities. International Immigration Review, 12, 469-84.
Quadagno, J.S. (1994). The Color of Welfare: How Racism Undermined the War on Poverty. New York: Oxford University Press.
Reese, E. (2001). Review of Why Americans Hate Welfare. Contemporary Society, 30, no. 2, 181-83.
Roediger, D.R. (1991). The Wages of Whiteness: Race and the Making of the American Working Class. London: Verso.
Robinson, J.P., and Davis, D.K. (1990). Television news and the informed public: an information-processing approach. Journal of Communication, 40, no. 3, 106-119.
Rogin, M. (1996). Blackface, White Noise: Jewish Immigrants in the Hollywood Melting Pot. Berkeley, California: University of California Press.
Sacks, K.B. (1994). How did Jews become white folks? In S. Gregory and R. Sanjek, (eds.) Race. New Brunswick, New Jersey: Rutgers University Press.
Sanchez, G.J. (1999). Race, national and culture in recent immigration studies. Journal of American Ethnic History, 18, no. 4, 66-84.
Sears, D.O. (1988). Symbolic racism. In P.A. Katz and D.A. Taylor, (eds.) Eliminating Racism: Profiles in Controversy. New York: Plenum, 53-84.
Sears, D.O., and Kinder, D.R. (1971). Racial tensions and voting in Los Angeles. In W.Z. Hirsch, W.Z., (ed.)  Los Angeles: Viability and Prospects for Metropolitan Leadership. New York: Praeger.
Sears, D.O., et al. (1997). Is it really racism? The origins of white Americans' opposition to race-targeted policies. Special issue on race. Public Opinion Quarterly, 61, no. 1, 16-53.
Sciortino, G. (2000). Toward a political sociology of entry policies: conceptual problems and theoretical proposals. Journal of Ethnic and Migration Studies, 26, no. 2, 213-28.
Simon, R.J., and Alexander, S.H. (1993). The Ambivalent Welcome: Print Media, Public Opinion, and Immigration. Westport, Connecticut: Praeger.
Smith, J.P., et al. (1997). The New Americans: Economic, Demographic, and Fiscal Effects of Immigration. Report prepared by the National Research Council for the U.S. Commission on Immigration Reform. Washington: National Academy Press.
Sniderman, P.M., and Piazza, T. (1993). The Scar of Race. Cambridge, Massachusetts: Harvard University Press.
Timmer, A.S., and Williams, J.G.. (1998). Immigration policy prior to the 1930s: labor markets, policy interactions, and globalization backlash. Population and Development Review, 24, no. 4, 739-71.
Valentino, N.A. (1999). Crime news and the priming of racial attitudes during evaluations of the president. Public Opinion Quarterly, 63, no. 3, 293-320.
Van Dijk, Teun A. (1988). News Analysis: Case Studies of International and National News in the Press. Hillsdale, New Jersey: Erlbaum.
Zolberg, A. (1981). International migrations in political perspective. In M. Kritz, C. Keeley, and S. Tomasi (eds.) Global Trends in Migration. New York: Center for Migration Studies, 3-27.
Zolberg, A. (1987). Wanted but not welcome: alien labor in Western development. In W. Alonson (ed.) Population in an Interacting World. Cambridge, Massachusetts: Harvard University Press, 36-73.
Zolberg, A. (1989). The next waves: migration theory for a changing world. International Migration Review, 23, 403-30.
[1]  Proportionately more white widows had been moved from Assistance to Dependent Children rolls starting in 1939 after the establishment of Social Security survivors' benefits.
[2]  For a fuller account, see  Muller and Espenshade ( 1985) and Timmer and Williamson (1998).
[3]  Sciortino (2000) provides a critical review of immigration policy research that discusses the perspectives in detail.