Green Critical Research Review Assignment | Online Assignment

I will upload both my paper and the CRR document

The Trump Hypothesis: Testing Immigrant Populations as a Determinant of Violent and Drug-Related Crime in the United States.

The author David Green presented the research question of whether Trump’s Hypothesis about immigrant Mexicans lead to an increase in violent and drug-related crimes in the U.S. or not. This announcement was claimed by Mr. Trump during his presidential campaign speech in summer 2015. The main dependent (Y) variable is the high level of violent and drug-related crimes in the U.S. The main independent (X) variable is immigrants specially Mexicans who are coming to the U.S. However. Dr. Green believes that X is not associated with Y. In addition, Dr. Green believes there are multiple independent variables which are: all immigrants, Mexican nationals, all undocumented immigrants and/or undocumented Mexican nationals.  The author succeeded in providing a clear definition of the operationalized variables based on the information Mr. Trump provided in his announcement that allows the readers to understand how the author will intend to measure the conceptual variables that he is interested in. Trump referred to the type of immigrants that enter the U.S as “drug dealers and rapists, defined the timeframe as immigrant who recently immigrated to the U.S. and defined the type of crime as both killing and raping as well as drug-related crime as arrestation during sales and possession”.

Green (or Dr. Green) hypothesized that, if the Trump hypothesis is true, immigrant populations have a positive effect on violent crime rates and on the drug-related arrest rates in the United States. The author aimed to reject the null hypothesis (Ho) and stated that there is no significant relation between immigrants including especially Mexican, documented and undocumented living in the U.S. and the level of violent and drug-related crimes. The author aimed to negate the Trump hypothesis through running an empirical statistical analysis. In order to ensure that the sample is nearly identical to the selected population on one or more specific variables; such as race, education, partisanship, gender, age range, and income that are important to this study, the author used stratified sampling (divided the population into two sub-groups-strata). In my opinion this is the best of selective the sampling methods to avoid biasing the results in a particular direction. Since the author could not identify from which countries the immigrants are coming from, the author decided to test for immigrants in general during the years of 2012-2014. Therefore, to test Mr. Trump’s Hypothesis, the author first used a basic correlation coefficient and simplified regression analysis by comparing the crime statistics to documented and undocumented immigrant population size using recent data. The author tested the association between the immigrant population, violent crime and drug-related at the state level.

In this analysis the author used four immigrant classifications: the overall foreign population, Mexican population and undocumented immigrant population, and undocumented Mexican population. Then, Dr. Green measured the dependent variable by comparing the difference in percentage in each sub-category between different groups and see whether their rates of crime differ. The author will test the Mr. Trump hypothesis through the following: First, running correlations with immigrant populations and crime rates whether they are related. Then, regressing immigrant populations data against violent crime rates and drug-related arrests the factor strength does not affect the results. Finally, running expanded regressions controlling for a variety of economic and demographic factors at the state level.

The research designs that Mr. Green obtained to see if there exists any evidence that immigrants are the causes of the crime in the U.S., is cross-sectional observational study. In this study Dr. Green examines a cross-section of social reality by focusing on obtaining data on immigrant population and crime data from different legislators at a single point in time. For example; the author obtained state-level foreign population and Mexican immigrant population data from the U.S ACS averaged for the year of 2012-14, Overall undocumented immigrant population estimates and undocumented Mexican immigrant population estimations were obtained from the non-partisan think tank MPI data of 2015, and 2008 Survey of Income and Program Participation data. As for the crime data, he obtained them from the Federal Bureau of Investigation’s UCR statistic by state for the year of 2012-14.

Some of the strengths of the author’s choice of research methods is that data can be very consistent and reliable. Also, data on all variables can be collected at once. The researcher will also be able to measure prevalence for all factors under investigation. On the other hand, there are few weaknesses. One is that data it may not be rigorous enough to explain complex issues and as well as data could be bias due to the low response and misclassification.  Dr. Green is specifically looking at the differences in percentage between the two outcomes of the hypothesis and the results were statically significant which shows a p-value lower than 0.05. In order the generate a strong support hypothesis in the quantitative research p-value method must be used, which in my opinion indicate it is unlikely that the positive relationship between X and Y is truly by chance. As a result, for the correlations, Dr. Green predicted about the direction of the relationship is that no significant association between any of the measures of (IV) crime listed and the overall (DV)foreign population which he tested to shows difference in proportions. For example; the comparison of foreign population and rape correlations are negative although not significant. Unemployment shows a strong association with murder at 99% confidence, while the poverty has strong and positive correlation with all violent crime. The higher (gender) male concentrations are negatively correlated with both violent crime and murder at 99% confidence, while (education) completion of high school is negatively associated with murder at 99% confidence. Age is the only variable showing significant correlation with drug-related crime at negative and 90-95% confidence. On the other hand, as a result for the regression, the highest level of significance achieved is at the 90% confidence level showing a weak positive association between undocumented immigrants and violent crime, and a weak negative relationship between foreign population size and rape. For violent crime, expanded regression indicate either non statistically significant relationships between immigrant population size and crime or a significant negative relationship on the overall foreign population and the murder rate. As for the drug-related crime, all foreign population associations are positive and statically significant at 95% confidence level.

There are a few limitations that were discussed in this study. One, the analysis for the crime was using only government statistics, which some viewed it as subjective and bias. Two, the variables selection was considered to be limited and very controlled. Last, using the state as the level of analysis results in a sample size. A good extension of the study is to be aware about the ecological fallacy which could resulted giving the right indication about that X effects Y even thought there were significant relations. The last interesting future research to be considered is to refine some of the tests by using municipal or neighborhood-level data.  In conclusion, I believe the statistical results strongly support Mr. Green’s main argument and conclusion. Mr. Green took the initiative and only tested Mr. Trump’s presidential speech hypothesis announced in 2015. I would still be interested to know what the results or the finding would be if similar research were conducted during the other pervious presidency campaigns for other presidents who announced similar hypothesis as Mr. Trump did. I believe if researchers conducted similar study, I think they would have saved time and effort.

The Trump Hypothesis: Testing Immigrant
Populations as a Determinant of Violent and
Drug-Related Crime in the United States∗
David Green, Nagoya University
Objectives. To test the “Trump Hypothesis”: whether immigrants are responsible for higher levels
of violent and drug-related crime in the United States, as asserted by Donald Trump in his 2015
presidential campaign announcement. This is achieved using recent crime and immigration data,
thus testing the common public perception linking immigrants to crime, and providing an updated
assessment of the immigrant-crime nexus. Methods. Rates of violent crime and drug arrests by
state are pooled for 2012–2014. These are compared against pooled statistics on foreign-born
and Mexican nationals living in the United States, as well as estimates of undocumented foreign
and undocumented Mexican population by state. The data are analyzed using correlation and
multivariate regressions. Results. Data uniformly showno association between immigrant population
size and increased violent crime. However, there appears to be a small but significant association
between undocumented immigrant populations and drug-related arrests. Conclusions. Results
largely contradict the Trump Hypothesis: no evidence links Mexican or undocumented Mexican
immigrants specifically to violent or drug-related crime. Undocumented immigrant associations
with drug-related crime are minimal, though significant. The Trump Hypothesis consequently
appears to be biased toward rhetoric rather than evidence.
When do we beat Mexico at the border? They’re laughing at us, at our stupidity. And now
they’re beating us economically . . . The US has become a dumping ground for everyone else’s
problems. These aren’t the best and finest. When Mexico sends its people, they’re not sending
their best . . . They’re sending people that have lots of problems, and they’re bringing those
problems to us. They’re bringing drugs. They’re bringing crime. They’re rapists. And some, I
assume, are good people.
Donald Trump, June 16, 2015 campaign speech (Washington Post, 2015)
To officially announce his presidential bid on June 16, 2015, Donald Trump gave a
speech likely to remain a part of his lasting legacy. Although he touched on a number of
topics, particularly memorable was his claim that Mexican immigrants, seemingly coming
to the United States at the behest of the Mexican government, are drug dealers, criminals,
and rapists. But to what extent is such a characterization true? While it may be largely
accepted in academia that immigrants at worst show no greater propensity toward crime
than native populations, popular dogma often identifies immigrant groups as criminal
threats that society needs to protect itself from. Trump’s declaration of a large criminal
element coming primarily from Mexico, hereafter dubbed the “Trump Hypothesis,” is the
latest in a long line of assertions attempting to conflate immigration with criminal activity.
∗Direct correspondence to David Green, Nagoya University Graduate School of Law, Furo-cho, Chikusaku,
Nagoya, Aichi 464–8601, Japan david.green@law.nagoya-u.ac.jp. The author shall share all data and
coding for replication purposes. The author wishes to thank the anonymous reviewers for their insightful
feedback.
SOCIAL SCIENCE QUARTERLY, Volume 97, Number 3, September 2016
C
2016 by the Southwestern Social Science Association
DOI: 10.1111/ssqu.12300
The Trump Hypothesis 507
In a polarizing political era coupled with a polarizing campaign season, it is useful to
consider once again to what extent immigrant populations may be associated with crime.
Aside from testing the assertions of a prominent political candidate, this article addresses
the underlying and deep-seated notion that immigrants are disproportionately responsible
for crime, updating previous findings with recent data. Although the Trump Hypothesis is
a contemporary iteration, the question of the immigrant-crime nexus is an old one. Given
the considerable degree of attention the Trump Hypothesis has received (Murray, 2015;
Sanneh, 2015; Victor, 2015), it is useful and necessary to reconsider immigrant associations
with crime using recent data.
This article consequently seeks to examine the “Trump Hypothesis” in detail, considering
how migrants, including Mexican nationals and documented and undocumented
immigrants, affect crime rates. Looking at state-level data, this article questions whether
immigrants are associated with violent crime including murder and rape specifically, as
well as drug-related crime including sales and possession. The Trump Hypothesis is tested
first by running correlations with immigrant populations and crime rates, then regressing
immigrant population data against violent crime rates and drug-related arrests, and finally
running expanded regressions controlling for a variety of economic and demographic factors
at the state level. Results are largely in line with prior research, indicating no significant
association between immigrant population and violent crime, and a small, albeit significant,
association between undocumented immigrant population and drug-related arrests.
Such results provide evidence contradicting the major tenet of the Trump Hypothesis:
that immigrants, by any measure of the group included here, are largely responsible for
crime in the United States. Consequently, this article asserts that the Trump Hypothesis is
driven much more strongly by rhetoric and ideology, rather than a careful consideration of
immigrant-crime data.
This article begins with a brief review of prior immigrant-crime research, followed by a
discussion of how the Trump Hypothesis is operationalized in this instance and the data
used to test it. Data analysis follows. The study concludes with a discussion of the results,
their implications, and some of the limitations of this study.
Background
The association between immigration and crime has been a recurring theme in American
popular culture and media for much of the country’s history. That immigrants are often
blamed for crime, and have been for a considerable period of time in the United States
and in other countries, is perhaps not a surprising revelation. The immigrant-crime association
has also been subject to considerable empirical investigation, particularly since the
early 20th century. Early government-sponsored commissions such as the 1901 Industrial
Commission, the 1911 U.S. Immigration Commission, and the 1931 Wickersham Commission,
for example, investigated the link between immigration and crime (Hagan, Levi,
andDinovitzer, 2008;McDonald, 1997). Although these commissions found no significant
relationship between immigrants and propensity to commit crime, the immigrant-crime
connection has remained popularly compelling even through modern times. Evidence may
point to little association between immigrants and crime, but public opinion continues to
link them together.
Recent events attempting to play off the immigrant-crime connection include the 1994
passage (and later repeal) of Proposition 187 in California, requiring among other things
immigration status checks for individuals in police custody (Martin, 1995). Following
the 1995 Oklahoma City bombing, perpetrated by a natural-born U.S. citizen no less,
508 Social Science Quarterly
Congress severely tightened immigration laws in the name of fighting terrorism (Johnson,
2003:47). Pat Buchanan’s 1996 presidential bid was predicated on building a wall across
the U.S.-Mexico border. The Patriot Act, passed in the wake of the September 11, 2001
terrorist attacks, gave the attorney general broad powers over immigrant residency, effectively
legitimizing indefinite detention for any foreign national labeled a “terrorist” under
a broad definition with little legal review (McCarthy, 2002:449). Arizona’s 2010 immigration
law allows police to check the immigration status of anyone stopped or arrested,
in spite of strong opposition from the Obama Administration (Liptak, 2012). Similarly,
Joe Arpaio, long-serving Sheriff of Maricopa County, Arizona, has garnered considerable
national attention with his immigration-focused policing efforts (Romero, 2011). Donald
Trump’s presidential campaign, and more specifically the Trump Hypothesis, is now a new
addition to the list.
While conservative groups often point to the real costs associated with crimes committed
by immigrants (see, for example, FAIR, 2015; Family Security Matters, 2015; Hahn,
2015), the larger and more fundamental question is whether immigrants, documented or
undocumented, are disproportionately more likely to commit crime than the native population.
This is the heart of the assertion of the Trump Hypothesis as well: that immigrants
are either entering the country already as criminals or they are highly criminally prone. All
groups in a society commit crimes to some extent, immigrants being no exception. However,
if immigrants are indeed disproportionately responsible for crime, perhaps greater
investment in border security would be worthwhile, as Trump suggests.
Empirical investigations considering associations between immigration and crime typically
show there to be no significant relationship. Early authors writing on the topic, such
as Sutherland (1924), in fact proposed that immigrants to the United States actually have
significantly lower crime rates compared to the native population. Subsequent generations
of immigrants, as they become more accultured, may have higher rates of crime but they
remain below native population levels. Others, such as Shaw andMcKay (1942), posit that
the socially disorganized urban neighborhoods where immigrants usually settle, characterized
by poverty, residential mobility, and structural disadvantage, are largely to blame for
criminal propensities rather than immigration itself. McKay (1965) later went on to more
strongly assert the neighborhood connection, claiming place of residence plays a major role
in delinquency, while immigrant status reduced propensity toward crime if anything.
A number of scholars have since taken on the immigrant-crime question, producing
very similar results (see, for example, Butcher and Piehl, 1998; Bradshaw et al., 1998;
Hagan and Palloni, 1998; Hagan, Levi, and Dinovitzer, 2008;McDonald, 1997). Namely,
researchers usually find either no significant relationship between crime and immigration,
or they find that immigrants are less likely to commit crimes compared to natives. These
results hold steady across country-level analyses (Tonry, 1997; Yeager, 1997), as well as
national-level (Chen and Zhong, 2013; Mears, 2001), city-level (Bradshaw et al., 1998;
Martinez and Lee, 2000;Ousey andKubrin, 2009), and neighborhood-level studies (Alaniz,
Cartmill, and Parker, 1998; Sampson, Morenoff, and Raudenbush, 2005). A number of
possible explanations for this phenomenon have likewise been suggested, such as immigrant
self-selection resulting in hard-working individuals making a positive contribution to the
host economy (Borjas, 1993; Cobb-Clark, 1993; Model, 1995), immigrant optimism and
determination in the face of hardship and disadvantage (Kao and Tienda, 1995; Martinez,
2006), and close family and community ties reducing the propensity to commit crime
(Ousey and Kubrin, 2009; Sampson, Morenoff, and Raudenbush, 2005).
For the sake of this article, however, the main concern is to determine whether or not
the Trump Hypothesis holds up to empirical scrutiny. The focus is not on the underlying
The Trump Hypothesis 509
causes of or antecedents to immigrant crime. Instead, the concern is whether Donald
Trump, in fact, uncovered the determining factors significantly linking immigrant status
to violent and drug-related crime in contradiction to the previous research. Does foreign,
specificallyMexican nationality or undocumented status, play a role in criminal propensity
as the Trump Hypothesis suggests? This is where we next turn our attention.
Data
Since immigrants “bring crime,” as Trump notes, we first need to define both immigrants
and crime in order to analyze the connection. TheTrumpHypothesis is both brief and rather
simple. It does not specify explicitly whether immigrants coming to the United States are
documented or undocumented. Rather, Trump asserts thatMexico “sends its people.” This
could encompass both documented and undocumented immigration. Although Trump
singles out Mexican nationals, he later notes in his speech that such criminally-prone
immigrants are “coming from more than Mexico.” In fact, Trump contends that they are
“from all over South and Latin America, and probably from the Middle East. But we do
not know because we have no protection and we have no competence” (Washington Post,
2015). Since we do not know where these immigrants are coming from, and we apparently
have no competence to discern this, it is probably a good idea to test for immigrants in
general as well.
Crime is likewise vaguely defined in the Trump Hypothesis. However, Trump does note
that immigrants are “drug dealers” and “rapists.”This provides some indication of the nature
of crime Trump has in mind. According to the Trump Hypothesis, immigrants would not
be prone to committing white-collar crime, for example. A reasonable assumption would
be that the Trump Hypothesis refers mainly to violent and drug-related crime.
Additionally, timing should be considered. The Trump Hypothesis makes no historical
assertions nor mentions prior waves of immigration. Rather, the Trump Hypothesis uses
the present tense to describe how immigrants are “bringing crime.” It can consequently
be assumed that the Trump Hypothesis refers to the current immigration context. Recent
data should suffice for the purpose outlined here.
To test the Trump Hypothesis, we then need to compare crime statistics to documented
and undocumented immigrant population sizes using recent data. Here, we test for an
association between immigrant populations, violent crime, and drug-related crime at the
state level. This is an aggregate measure of the immigrant-crime relationship, one that
excludes individual-level factors. Essentially, we are comparing crime rates in states with
relatively large levels of immigration to states with relatively small levels. Because the Trump
Hypothesis paints a broad picture of immigrants as criminals, this is an appropriate level
of analysis to test the hypothesis. The Trump Hypothesis is, after all, a blanket assertion of
a national phenomenon. National-level data are thus required to validate or refute it.
Because the Trump Hypothesis, as presented, is simple and makes no explicit qualifications
for outside indicators or mitigating factors, the analysis here begins with basic tests
of association between immigrant populations and crime: namely, a zero-order correlation
and simplified regression analysis. The main concern here, after all, is with immigrant populations
and any relationship they may have to crime rates. However, the literature notes
a host of measures that may have some bearing on crime, including a state’s mean age,
gender balance, and general education level (Butcher and Piehl, 1998, Sampson,Morenoff,
and Raudenbush, 2005). These factors are subsequently controlled for as a more nuanced,
additional test of the Trump Hypothesis.
510 Social Science Quarterly
Four immigrant classifications are used in this analysis: the overall foreign population,
Mexican population, undocumented immigrant population, and undocumented Mexican
population. To more clearly define immigrant classifications, the foreign population category
includes any individual who was not a U.S. citizen at their time of birth. This includes
naturalized citizens, refugees, legal immigrants, some categories of nonresident immigrants,
as well as undocumented immigrants. The Mexican national category is a subset of the
foreign population grouping, considering only individuals having Mexican citizenship at
birth. Undocumented immigrants include immigrants of any nationality residing in the
country without legal documentation, either through undocumented entry or a loss of
valid visa status. The undocumented Mexican immigrant population is likewise a subset
of the total undocumented foreign population.
State-level foreign population and Mexican immigrant population data were obtained
from the U.S. Census Bureau’s American Community Survey (ACS), averaged for the years
2012 to 2014 (USCB, 2014). Obtaining the pooled result for three years rather than a
single year helps to avoid any individual year anomalies and presents a somewhat more accurate
picture in aggregate (Wadsworth, 2010:539). The ACS is a nationally representative
survey used for a variety of purposes, particularly to understand changing demographics in
the country and to assist in the allocation of federal and state funds. Overall undocumented
immigrant population estimates and undocumented Mexican immigrant population estimates
were obtained from Migration Policy Institute data (2015). The Migration Policy
Institute, a nonpartisan think tank, calculated illegal immigrant population estimations,
including breakdown by state and in many cases by nationality, based on ACS 2009–2013
data and 2008 Survey of Income and Program Participation data.
On the other side, crime data were obtained from the Federal Bureau of Investigation’s
Uniform Crime Reporting (UCR) statistics by state, averaged for the years 2012 to 2014
(FBI, 2012a, 2013a, 2014a). The analysis here uses statistics from the “violent crime” category,
consisting of murder and nonnegligent manslaughter (hereafter noted as “murder”),
rape, robbery, and aggravated assault. Overall violent crime rates are considered, as well
as breakdowns of the murder and rape subcategories. Since Trump explicitly mentions
rape, it warrants inclusion here. Trump also conflated illegal immigration with murder
in a July 6, 2015 speech, where he attributed the murder of a San Francisco woman by
an undocumented immigrant to poor immigration enforcement (Schleifer, 2015). Being
one of the most serious violent crimes and connected to immigration by Trump, we test
for specific connections to murder here as well. Rape statistics in the overall violent crime
category as well as the rape category use the “legacy” UCR definition. A somewhat looser
calculation of rape statistics was introduced in 2013, termed the “revised” definition (FBI,
2014b), although for the sake of comparability with previous years the legacy definition
was included in the aggregate statistics. We use the legacy definition here to be consistent,
although results are highly similar with either definition.
Drug-related crime data also come from UCR statistics, pooled again using 2012–2014
data. UCR data only provide actual arrest rates for drug crime, rather than total incidents
as with violent crime. Drug arrests include both sales and possession, broken down by
various categories of narcotics (FBI, 2012b, 2013b, 2014c). Subcategories of drug sales
and drug possession are provided by UCR at the regional and national level, but not at
state level. Because Trump explicitly noted that immigrants are “drug dealers,” this analysis
breaks down drug arrests into the sales and possession subcategories. Regional-level ratios
were applied to state-level arrest numbers to obtain an estimate of the sales/possession
distinction.
The Trump Hypothesis 511
TABLE 1
Variable Definitions
Variable Definition
Violent crime Incidents of murder, rape (legacy definition), robbery, and
aggravated assault per 100,000 individuals
Murder Incidents of murder and nonnegligent manslaughter per 100,000
individuals
Rape Incidents of rape (legacy definition) per 100,000 individuals
Drug arrests Number of drug arrests per state, including sales and possession,
per 100,000 individuals
Drug sales Number of arrests per state for drug sales, per 100,000 individuals
Drug possession Number of arrests per state for drug possession, per 100,000
individuals
Foreign population Size of state foreign-born population per 100,000 individuals
Mexican population Size of state Mexican-born population per 100,000 individuals
Undocumented
immigrants
Estimated number of state undocumented immigrants, per 100,000
individuals
Undocumented
Mexican immigrants
Estimated number of state undocumented Mexican immigrants, per
100,000 individuals
Poverty rate State poverty rate
Unemployment Percentage of workers unemployed in state
Median income State median household income
Age 15–24 Percentage of state residents, ages 15 to 24
Male Percentage of male state residents
HS education Percentage of state residents over age 25 with at least a high school
diploma
Additionally, to control for other factors commonly associated with crime rates, in
the expanded regression analysis we introduce poverty rates, unemployment rates, median
income, young adult concentration (individuals aged 15–24 years old), gender balance, and
percentage completing at least a high school education at the state level. While additional
measures may work as determinants of violent crime, we narrow our analysis here for the
sake of simplicity. The primary concern is the effect of immigration on crime, controlling
for some of the other major factors associated with crime. Poverty, median income rates,
young adult concentration, gender balance, and percentage completing high school were
obtained through U.S. Census reporting, based again on ACS data and pooled 2012 to
2014 (USCB, 2014). State unemployment rates 2012 through 2014 come from the Bureau
of Labor Statistics (BLS, 2012, 2015). Table 1 provides the full list of variables and their
definitions.
Crime, immigrant population data, poverty rates, unemployment rates, median income,
young adult concentration, gender balance, and percentage completing high school were
thus calculated for all 50 states, as well as the District of Colombia. Undocumented
immigrant and undocumented Mexican immigrant statistics were not available in every
instance, although 94 percent of states (48 of 51) had undocumented immigrant estimates,
and 80 percent (41 of 51) had estimates for undocumented Mexican immigrants. To
account for variations in state population size, crime rates and immigrant population
sizes were standardized to incidents/individuals per 100,000 residents. Table 2 lists the
descriptive statistics for all variables.
512 Social Science Quarterly
TABLE 2
Descriptive Statistics
Variable Mean Std. Dev. Min. Max. Obs.
Violent crime 356.64 179.70 114.9 1281.9 51
Murder 4.26 2.52 1.4 15.9 51
Rape 40.42 17.05 12.6 125.4 51
Drug arrests 381.46 162.17 5.87 676.22 51
Drug sales 67.08 29.09 1.02 124.27 51
Drug possession 313.93 135.18 4.85 559.46 51
Foreign population 8960.26 6070.53 1419.45 26899.35 51
Mexican population 2135.2 2540.39 54.05 11041.21 51
Undocumented immigrants 2385.13 1623.71 269.64 7914.95 48
Undocumented Mexican immigrants 1440.28 1215.56 29.88 5514.90 41
Poverty rate 15.16 3.36 8.7 24 51
Unemployment 6.66 1.55 2.9 9.3 51
Median income 53130.24 8842.84 37963 72483 51
Age 15–24 14.03 0.75 12.4 16.17 51
Male 48.86 0.76 46.00 51.38 51
HS education 87.61 3.21 81.20 92.27 51
Methodology
Thus far we have outlined a plan to test the Trump Hypothesis by comparing immigrant
population sizes against violent crime and drug-related arrest rates by state. We first run
simple tests, looking for correlations and simple associations, and later controlling for
additional economic and demographic indicators. To reiterate the Trump Hypothesis in
more explicit, testable terms, we start with the following hypothesis:
H1: Immigrant populations (all immigrants, Mexican nationals, all undocumented immigrants,
and/or undocumentedMexican nationals) have a positive effect on violent crime
rates in the United States.
To specify somewhat narrower parameters of violent crime, we make two subhypotheses:
H1a: Immigrant populations (all immigrants, Mexican nationals, all undocumented immigrants,
and/or undocumented Mexican nationals) have a positive effect on murder
rates in the United States.
H1b: Immigrant populations (all immigrants, Mexican nationals, all undocumented immigrants,
and/or undocumented Mexican nationals) have a positive effect on rates of
rape in the United States.
In order to additionally consider associations with drug-related crime, the Trump Hypothesis
should similarly be put into more testable terms in relation to drugs:
H2: Immigrant populations (all immigrants, Mexican nationals, all undocumented immigrants,
and/or undocumented Mexican nationals) have a positive effect on drug-related
arrest rates in the United States.
Similarly, two subhypotheses follow based on the drug sales or drug possession subcategories:
The Trump Hypothesis 513
TABLE 3
Correlations with Violent Crime and Drug Arrests
Violent Drug Drug Drug
Crime Murder Rape Arrests Sales Possession
Foreign 0.1660 −0.0583 −0.2569 0.1017 0.1008 0.0977
Mexican 0.0568 −0.1071 0.2401 0.1443 −0.0198 0.1797
Undocumented 0.1865 0.0620 −0.1831 0.2103 0.1483 0.2174
Undoc Mexican 0.0275 −0.1148 0.1593 0.1693 0.0078 0.2043
Poverty 0.3620∗∗ 0.5350∗∗∗ 0.3714∗∗ −0.1109 −0.2028 −0.0929
Unemployment 0.2843∗ 0.446∗∗∗ −0.1436 −0.1549 −0.1410 −0.1591
Income 0.0231 −0.1580 −0.2743∗ −0.0368 0.0220 −0.0551
Age 15–24 0.2430 0.2866∗ 0.2787∗ −0.2784∗ −0.3388∗∗ −0.2708∗
Male −0.6431∗∗∗ −0.6658∗∗∗ 0.1135 0.0968 0.0366 0.1255
HS education −0.2362 −0.4193∗∗∗ 0.0551 −0.1391 −0.0828 −0.1504
∗p < 0.1; ∗∗p < 0.05; ∗∗∗p < 0.01.
H2a: Immigrant populations (all immigrants, Mexican nationals, all undocumented immigrants,
and/or undocumented Mexican nationals) have a positive effect on drug sales
arrest rates in the United States.
H2b: Immigrant populations (all immigrants, Mexican nationals, all undocumented immigrants,
and/or undocumented Mexican nationals) have a positive effect on drug
possession arrest rates in the United States.
The preceding hypotheses thus allow for empirical tests of the Trump Hypothesis. In
order to test these assertions, we first conduct correlations between the various independent
variables: foreign population, Mexican population, undocumented population, and
undocumented Mexican population, with the dependent variables: rates of violent crime,
the murder and rape subcategories, as well as drug arrest rates and the sales and possession
subcategories. Also included in the correlations are the economic and demographic control
variables: poverty rate, unemployment rate, median household income, young adult
concentration, gender balance, and percentage completing high school. We subsequently
run two sets of regression analyses: first a simple test of the Trump Hypothesis comparing
immigrant populations to violent crime and drug-related arrests, then a more complex test
incorporating controls for the economic and demographic factors noted above.
Results
Table 3 provides the results of the zero-order correlation analysis.Most apparent, perhaps,
is that there is no significant association between any of the measures of crime considered
and any foreign population shown here. In some cases, such as the comparison of foreign
population and rape, correlations are in fact negative, although not significant. Unemployment
shows a strong association with murder (99 percent confidence), while poverty has a
strong and