Changes in American Attitudes toward Immigrant-Native Job Competition

It is often perceived that the American public has been concerned about immigrantnative job competition for at least the last two to three decades. Less is known about the changing attitudes of Americans in this regard. This paper examines changes in American attitudes toward competition of immigrants with the native-born and changes in the determinants of such attitudes, using data from General Social Surveys 1996, 2004, and 2014. It is found that the percentage of Americans who rejected the statement that “Immigrants take jobs away from people who were born in America” had actually increased from 28.4% in 1996 to 34.6% in 2004 and to 42.8% in 2014. Results of multiple regression reveal that nativity, education, race, and region were consistent predictors of attitudes toward immigrant-native job competition across the three points in time, but subjective class standing and political party affiliation were significant predictors only in 2004 but not in the other years. The implications of findings are discussed.


Introduction
Immigration is again at the center of national discourses in the United States and across Europe. Despite many aspects of the immigration debate, the economic impact of immigration on the host society and population often takes the center stage. In particular, job competition between immigrants and native-born citizens has been one of the main issues in the debate over the economic impact of immigration in the United States. However, it is surprising to see how little research that focuses on attitudes toward immigrant-native job competition has been conducted. Furthermore, while it is often perceived that recent American public opinion on immigration in general and immigrant-native job competition in particular has become increasingly negative, especially after the 9/11 terrorist attacks [1][2][3][4], hard evidence, especially longitudinal evidence, is lacking. Lack of evidence is likely to lead to mis-assessment of public mood and therefore inappropriate policy decisions. Attitudes toward immigration and immigrants can also influence individuals' behaviors toward the foreign-born. To fill this lacuna, this paper seeks to provide evidence on changes in American attitudes toward immigrant-native job competition. Specifically, this study addresses two research questions. (1) How have American attitudes toward immigrant-native job competition changed in recent decades? (2) How have the determinants of American attitudes toward immigrant-native job competition changed in recent decades?
The next section of this paper briefly reviews the literature pertinent to the research questions. This is followed by the hypotheses to be tested in this study. Data and methods of data analysis are then described. The remaining sections present the results and discuss their implications.

Literature Review
An extensive literature search reveals no single study devoted to American attitudes toward immigrant-native job competition, although there are many studies of American attitudes toward immigration and immigrants in general. Gallup polls have surveyed American opinions about immigration and immigrants over time, but we cannot find a single question about how Americans view job competition between immigrants and native-born Americans. The closest question asked in Gallup polls in June 2006 and June 2008 was: "Which comes closest to your view --[Rotated: illegal immigrants mostly take jobs that American workers want, (or) illegal immigrants mostly take lowpaying jobs Americans don't want"]? A large majority of the respondents (74% in 2006 and 79% in 2008) stated that illegal immigrants mostly take low-paying jobs that American workers don't want, compared to 17% in 2006 and 15% in 2008 of them who said they mostly take jobs that American workers want with the rest stating "neither/both/no opinion." However, this question was about job competition between the illegal segment of the immigrant population and natives, not between all immigrants and natives. The American Election Study included a question "How likely immigration will take away jobs?" However, this question does not specify who is competing with whom for jobs and is therefore not very useful.
There is no study of the determinants of American attitudes toward immigrant-native job competition in the literature, although there are quite a few studies of determinants of American attitudes toward immigration and immigrants in general [5][6][7][8]. With regard to the forces that shape changes in American attitudes toward immigration and immigrants over time, there are two competing approaches. The political economy approach contends that American attitudes toward immigration and immigrants change contingent upon changes in the economy and labor market conditions [6,7,[9][10][11]. Specifically, when the economy is bad and actual job competition between immigrants and the native-born intensifies, negative attitudes toward immigration and immigrants will rise. When the economy improves and actual immigrant-native job competition slows down or declines, negative attitudes will also decrease. On the other hand, the political psychological approach argues that American attitudes toward immigration and immigrants change independent of changes in the economy and labor market conditions but dependent on sociotropic concerns about its cultural impacts [12][13][14]. That is, the attitudes are not much correlated with the actual economic and labor market conditions and are shaped more by the psychology of the nation. It is important to emphasize that these approaches address what factors drive American attitudes toward immigration and immigrants, not American attitudes toward immigrant-native job competition. However, they may be useful for the current study. I would argue that these two approaches may not be mutually exclusive. Both economic and labor market conditions and sociotropic concerns could influence American attitudes toward immigrant-native job competition simultaneously, but sometimes a particular force could play a more important role than the other.

Hypotheses
For research question 1, I expect that American attitudes toward immigrant-native job competition have changed in the last couple decades in response to the changing economic and political circumstances. Specifically, the American attitudes are expected to have become more negative in at least a short period of time after the 9/11 attacks but more positive in the current decade of the twenty-first century. There are two plausible reasons for the more negative shift in the attitudes in the period after the 9/11 attacks. First, the sense of insecurity and uncertainty toward foreign invasion in the American public may have risen as a result of the 9/11 attacks. Second and perhaps more importantly, the economic recession in the aftermath of the 9/11 attacks greatly increased the level of anxiety about job prospects of Americans and concerns about immigrant-native job competition [3]. Nativism tends to increase with economic hard times [6]. However, as the U.S. economy improved after the great recession of 2007-2009 and deceasing unemployment rates in the first several years of the current decade, American concerns about job competition between immigrants and natives may have flagged significantly, leading to a more positive shift in the attitudes.
With regard to the second research question, I expect that many determinants of American attitudes toward immigrantnative job competition should show a consistent relationship with the dependent variable, but some determinants may gain or lose some importance because of changing national and international environment. For instance, the effect of education is expected to remain stable over time because of the long established relationship between education and attitudes toward immigration and immigrants. Nonetheless, the effect of race on the attitudes may vary over time, depending on the perceived job competition between certain racial groups and immigrants during specific time periods.

Data and Methods
The data for this analysis come from the GSS of 1996, 2004, and 2014. Since the focus of this study is on the changes in American attitudes toward immigrant-native job competition, longitudinal data are required. These three years of GSS were selected because the question on the dependent variable-attitude toward the statement "Immigrants take jobs away from people who were born in America" was surveyed only in those three years. The three years were spaced roughly once every decade since the 1990s. We can also see the American attitudes before and after the 9/11 attacks and in the latest available year (2014). It is well known that the GSS is a nationally representative sample of the U.S. adult population aged 18 or older. Since only one adult in each household was selected for survey, people in larger households may have a smaller probability to be selected. To correct for this selection bias, the data were weighted so that the results can accurately reflect the total population. I restricted the analysis to the respondents who provided a Madridge J Behav Soc Sci. ISSN: 2638-2032 valid answer to the dependent variable. After the restrictions, the sample sizes for three years were 1,278 for 1996, 1,185 for 2004, and 1,218 for 2014.

The dependent variable
The dependent variable for this study is attitude toward immigrant-native job competition, which is based on the survey statement "Immigrants take jobs away from people who were born in America. "This is a Likert scale ordinal measurement with five categories (1=agree strongly, 2=agree, 3=neither agree nor disagree, 4=disagree, 5=strongly disagree). This variable was reversed coded so that a higher score indicates a higher level of agreement with the statement, or a more negative attitude toward immigrant-native job competition.

Independent variables
Based on relevant studies in the literature and the availability in the GSS, nine demographic, socioeconomic, and political variables are included as predictor variables for each of the three years. Male is a dummy variable coded 1 for males and coded 0 for females. Age is a continuous variable measured by years. The nativity variable is also a dummy variable with 1 indicating that the respondent was born in the U.S. and with 0 denoting foreign-born. I created two dummy variables for race, one for black (black=1, otherwise=0) and the other for other race (other race = 1, otherwise = 0), with white as the reference category. Three dummy variables were created for region: Northeast, Midwest, and West, and the reference category is South. I created a dummy variable for urban coded 1 for urban areas and coded 0 for rural areas. Education is a continuous variable measured by years of schooling. Social class is an ordinal variable with four categories: 1= lower class, 2= working class, 3=middle class, and 4=upper class. Finally, a dummy variable was recoded for Republican (Republican=1, otherwise=0). Income is not included partly because prior studies of immigration attitudes [15,8] found an insignificant effect of income on immigration attitudes and partly because income has a large number of missing cases.

Limitations of data
The GSS data are not without limitations. First, continuous data on attitudes toward immigrant-native job competition would be ideal to observe the complete trend, but only three years of data are available. Secondly, some significant predictors of the attitudes, especially economic variables pertinent to labor market conditions and job competition, are absent in the GSS data. For example, some variables related to economic prospects or threat and variables germane to cultural threat used by Chandler and Tsai [15] are only available in 1994. A constraint is that a variable must be available in all three survey years in order to be included in this study. Nevertheless, these limitations notwithstanding, the GSS remains the best data available to answer the research questions of this study. In particular, GSS is the only survey that asked the question about job competition between immigrants and native-born Americans.

Methods
To determine changes in American attitudes toward immigrant-native job competition, I performed trend analyses across the three years in three decades. To assess changes in the determinants of the attitudes toward immigrant-native job competition, I conducted regression analyses. Since the dependent variable is ordinal, it appears that ordinal logistic regression is most appropriate. I tested ordinal logistic regression models for each of the three survey years. However, the parallel line assumption is not met for 2004 and 2014. In order to make the results comparable, I decided to use ordinary least squares regression model instead because the dependent variable has five categories and approximate normal distributions.    Table 1 provides the means and standard deviations of all the variables used in the analysis. Consistent with the results in Figure 1, the means for the 5-point ordinal dependent variable display a pattern of continuous declines from 3.29 to 3.15 and to 2.93, a movement from a more negative attitude (agree) to a more positive attitude (disagree). Table 1 also shows changes in the independent variables over time. Note that the mean of a dummy variable can be interpreted as percentage after multiplying it by 100. The percentage of males in the samples appeared to decrease slightly from 47% in 1996 to 45% in 2004 and to 44% in 2014.

Trends in attitudes toward immigrant-native job competition
Madridge J Behav Soc Sci. ISSN: 2638-2032 The respondents became slightly older over time on average from 43 years old in 1996 to 44 years old in 2004 and to close 46 years old in 2014. The percentage of the native-born witnessed a 7-point decline from 1996 (91%) to 2014 (84%). The blacks in the sample increased slightly in 2014 to 13% from 12% in the earlier years, and the other race rose significantly from 7% in 1996 and 2004 to 12% in 2014. There were slight increases in the population in the West but a decline in the population in the Northeast. The proportion of the population that was urban somewhat decreased in 2004 and 2014. The average years of education increased slightly with fluctuations and remained at the level of some college. The average class ranking appeared to decrease a bit in 2014. The share of Republicans declined significantly to 21% in 2014 from 28% and 30% in 1996 and 2004, respectively.

Changes in the determinants of attitudes toward immigrant-native job competition
The results of OLS regression models predicting American attitudes toward immigrant-native job competition in the three survey years are presented in Table 2. Education appeared to be associated with a more positive attitude toward immigrant-native job competition in all three years. On the other hand, being native-born was consistently correlated with a more negative attitude toward the immigrant-native job competition in all the three years. Age had no effect on the dependent variable at all three points in time.
The effects of other predictors varied from time to time. Gender has no significant effect on the dependent variable in 1996 and 2014, but men were significantly more likely to have a negative attitude toward immigrant-native job competition than women in 2004. Blacks had a more negative attitude toward immigrant-native job competition than whites although the black-white difference was insignificant in 2004.
In contrast, people of other race tended to have a more positive attitude than whites though the difference between whites and other races was not significant in 1996 and 2014. As expected, people in the Northeast, Midwest, and West tended to have a more positive attitude toward immigrantnative job competition than the southerners, but statistical significance can be observed for the West-South difference at all the time and for the Northeast-South difference in 1996. Urban population had a consistently more positive attitude toward immigrant-native job competition than rural population, although the difference did not reach statistical significance at the .05 level in 2014. A higher class status was consistently associated with a more positive attitude at all time, but the effect was only significant in 2004. Finally, Republicans tended to have a more negative attitude toward immigrant-native job competition than non-Republicans, but the difference reached statistical significance at the .05 level only in 2004. In addition, I tested the effects of potential interactions between class and gender and between education and gender on attitudes toward immigrant-native job competition, but no significant interaction effects were detected.

Conclusion
A key finding of this study is that partly incongruent with my expectation, American attitudes toward immigrant-native job competition had shifted toward the more positive direction from 1996 to 2014. An important implication of this finding is that the general perception about Americans' negative attitudes toward immigration and immigrants especially after the 9/11 attacks claimed in the academic writings and media reports may be inaccurate. The anti-immigrant rhetoric couched by some politicians or political candidates may not represent the thinking of the majority of Americans. A plausible reason for the positive swing in 2004 was that perhaps the worst period of insecurity and economic mishaps after the 9/11 terrorist attacks was already over by 2004 as the GDP growth rate reached 3.8% in that year, the highest since 2001 and in the decade of 2001-2009. As a result, the fear of competition with immigrants may have lessened, leading to a positive trend in the immigration attitudes. Similarly, one possible reason for the positive shift in 2014 was that the economic recovery was in full swing with a GDP growth rate of 2.4% and a decline of unemployment rate to 6.2% in 2014; thus, the concerns about immigrant job competition with the natives had declined.
Another major finding is that the determinants of American attitudes toward immigrant-native job competition had remained stable although the effects of some predictors had changed. For example, education consistently reduces negative attitude toward immigrant-native job competition while being U.S.-born is more likely to have a negative attitude toward immigrant-native job competition at all time. The consistent relationship between education and a more positive attitude toward immigrant-native job competition suggests that more education makes people more openminded and more rational in judgement about immigration and immigrants. On the other hand, the consistent relationship between being native-born and a more negative attitude toward immigrant-native job competition suggests that nativism remains deep among the natives. The inconsistent effects of gender, race, region, urban residency, class, and political party affiliation on attitudes toward immigrant-native job competition caution us in making generalized statements about their relationships.
If possible, future research should use continuous timeseries data to analyze the complete trend in American attitudes toward immigrant-native job competition. Future research should also use measurements of economic threat and cultural threat of immigrants as predictors if available. Better data and techniques will ensure a fuller understanding of American attitudes toward immigrant-native job competition. More generally, as the Trump administration is moving to broaden control from illegal immigration to legal immigration, further research on American attitudes toward, and impact of, legal immigration is a desideratum.