


































Microsoft Word - 1TerryHolt_Reviewed_SG_TD.docx


 

CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

The Effect of Race on Poverty  

and the Equality of Minorities   
 

 
Terry Holt 

Department of Sociology 
California State University, Los Angeles 

 
 

Using General Social Survey (GSS) data from 1972 to 2004, this 
study considers if racial discrimination continues to represent a 
significant barrier for the economic advancement of African 
Americans and other minorities in the U.S. compared to whites. 
Specifically, does race affect income equality? Bivariate analysis 
and Logistic Regression analysis indicate that the “total family 
income average” for Blacks is significantly less compared to 
Whites, and Blacks are more likely to live in poverty (income less 
than $25,000) than Whites. Conversely, when it comes to other 
minority groups, there is no significant difference in “total family 
income average” compared to Whites. In addition, the findings 
also suggest that women are far more likely to live in poverty than 
men, as one’s gender has a significant effect on one’s income 
earning potential. 
 

 
INTRODUCTION 

 
1Racism has existed throughout human 
history, and it continues to represent 
significant problems for many people in 
the United States. Racism is the belief 
that one’s race is primarily the 
determining factor that reflects human 
traits and capacity. Racist ideology 
generally supports the premise that a 
particular race is either superior or 
inferior to another, and that a person’s 

                                                           

1. Terry Holt is a graduate student at the 
Department of sociology. His paper was 
originally submitted to SOC 410 Advanced 
Statistics, taught by Professor Hyojoung Kim in 
Summer 2009.   

social and moral qualities are pre- 
determined by his or her inborn 
biological characteristics. The distinction 
of racial differences gives way to the 
belief of an inherent superiority of a 
particular race while simultaneously 
ordering other races in a hierarchy. 
Institutional racism causes a large 
numbers of individuals who are deemed 
inferior to be denied even the most basic 
human rights. Conversely, individuals 
who belong to the group that is deemed 
superior have historically been elevated 
to positions that allow them to enjoy 
preferential treatment. Why do people 
from one social group oppress and 
discriminate against people from other 
social groups; and why is it so difficult 



INCOME INEQUALITY by Terry Holt   2 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

to eliminate? This study investigates 
whether racial discrimination continues 
to represent a significant problem for 
African Americans and other ethnic 
minorities in the U.S. Some race 
theorists assert the pessimistic view that 
racism is permanent and that policy 
reforms will not curtail the development 
of racial distinction and antagonisms.  
    Racial inequality has become an 
enduring and deeply regimented way of 
knowing and organizing the social 
world, and thus is unlikely to be 
completely eliminated. The African- 
American experience in the United 
States has enriched the fabric of 
American history and society in a 
myriad of ways, many of which have 
only recently been recognized. However, 
the overarching theme of African- 
Americans and other minority groups’ 
experience has been one of misery, 
exploitation, inequality, and discrimi- 
nation. It is to this end, that those who 
wish to understand the minority 
experience in America ask themselves 
following question: Are minorities 
making progress in the United States? Is 
there any significant gap in the income 
of African Americans and other 
minorities compared to those of 
Caucasians in the United States, and 
what ethnic minorities are more likely to 
live under the poverty level? If we find 
that a particular ethnic group is more 
likely to live in poverty compared to 
Whites, then, one might conclude that 
racism continues to remain a significant 
obstacle for the economic advancement 
of the aforementioned minority groups. 
In addition to race, this study will also 
consider the impact of age, gender, 
religion, political affiliation, educational 
attainment, and the number of years 

spent obtaining formal education and 
training with respect to relative income 
and those who live in poverty. Poverty is 
defined as the total family income level 
of respondents that falls at or under 
25,000 annually. 
 
LITERATURE REVIEW 

 

Recent battles regarding civil rights and 
race discrimination in the United States 
were fought on two fronts: legal, and the 
public’s perception of race. Legal fronts 
consisted of lawsuits and amended 
legislation prompted institutions such as 
schools, banks, and government agencies 
to lessen race discrimination. Brown vs. 
Board of Education, the Civil Rights Act 
of 1964, and other subsequent battles 
brought race discrimination to the 
attention of the American public. The 
former front involves the public’s 
perception of race. Henry and Sears 
(2002) argue that public sentiments 
concerning African- Americans are 
governed by a psycho- logical blend of 
negative feelings and conservative 
values, particularly, the belief that 
African-Americans violate cherished 
American values. The public’s per- 
ception of African-Americans is rooted 
in an abstract system of early learned 
moral values and ideas that typically 
view African-Americans as social 
misfits.  

     Racial conflict has plagued the 
United States since its inception; in 
particular, it has been primarily driven 
by racial prejudice of African-Americans 
(Allport 1979). While overt forms of 
racial discrimination such as “Jim Crow” 
segregation have been eliminated in the 
United States, and whites’ opinions 
regarding racial issues have become 



INCOME INEQUALITY by Terry Holt   3 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

more liberal, racial discrimination 
remains a significant difficulty for many 
ethnic minority groups. Moreover, 
current research shows that racism has 
evolved from these overt forms of Jim 
Crow segregation (older belief systems 
which incorporated social distance 
between the races). One form of research 
has been developed around the basic 
idea that new forms of racism have taken 
root in America, which is the symbolic 
racism theory (Sears 1988). According 
to Kender and Sears (1981), symbolic 
racism is commonly described as a 
coherent belief system, which supports 
concepts that racial discrimination is no 
longer a valid point of contention for 
African Americans, and that their 
disadvantage stems from personal ir- 
responsibility, and thus their continual 
demand for equal treatment is not valid.  

     Proponents of liberal optimism, 
on the other hand, contend that viable 
solutions to our nation’s race problems 
are possible. Robert Parks (1950) clearly 
articulates key concepts of the race 
relation cycle. He argues that race 
relations develop in a four cycled stage: 
contact, conflict, accommodation, and 
assimilation. The first stage occurs when 
two or more different races of people 
come together, and they are obliged to 
interact with each other. Competing for 
scarce resources they fall into conflict, 
which eventually gives way to accom- 
modation, where a stable but antagonist 
social order fosters a social hierarchy. 
Finally, Parks asserts that accom 
modation is attained when different 
races assimilate through a process of 
cultural and physical merging. The end 
result of such a merger is the 
development of one homogenous race, 
where class supersedes race as the 

primary focal point of social distinction. 
Parks ascertains that race relations 
invariably pass through the previously 
mentioned stages, and that the present 
location of a particular race of people 
offers strong evidence to suggest the 
kind of social experiences that a 
particular race of people will encounter 
not only their past but also the future 
path.  

     The American society, like many 
others throughout the world, is organized 
by powerful dynamics that are often very 
difficult to interrupt. Privilege is a 
predictable precursor for such things as 
race distinction, because the privileged 
group must distinguish itself from other 
groups. Distinctions based on race may 
not always be carried out with malicious 
intent, however, to suggest that the 
effects of such characterizations are 
inconsequential, definitely deserves 
examination. But, how are we to 
understand the realities that both 
produce such distinctions and the 
ensuing consequences that they in- 
variably produce? Do we view them as 
purely accidental, or as oddities that 
simply seem to happen? Or is race, in 
fact, reflective of designed dynamics that 
are sown into the very fabric of our 
society?  
 

DATA AND VARIABLES 

 

In order to empirically examine whether 
or not race remains a significant barrier 
for the equality of ethnic minorities in 
the United States, this researcher uses 
General Social Survey (GSS) data. The 
GSS was designed as part of a data 
diffusion project in 1972. The GSS 
replicated questionnaire items and 
wordings are used in order to facilitate 



INCOME INEQUALITY by Terry Holt   4 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

time and trend studies. This data 
collection includes a cumulative file that 
merges all data collected as part of the 
General Social Services Surveys from 
1972 to 2004. The 2004 survey was 
composed of permanent questions that 
appeared on two out of every three 
surveys and a small number of 
occasional questions that occurred in a 
single study. There were a total of 2,812 
respondents   

 
RESEARCH HYPOTHESES 

 

The Dependent Variable: Income 

Level 

 
A comparative level of income between 
whites (the comparison group), African 
Americans and other ethnic minorities 
over time will demonstrate whether 
racism remains a central hindrance to the 
advancement of minorities in the United 
States. That is, I hypothesize that 
Caucasians will show a higher mean 
income from that of minorities and, 
therefore, a lowered propensity for 
having a total family income of 25,000 
dollars or less. If racial equality is 
present between races, then, we can 
expect to see a somewhat uniform 
distribution of income between the 
different ethnic groups and an average 
number of people in different races 
living in poverty. However, if we see a 
significant difference between mean 
incomes of different ethnic groups, then, 
we assume that there is no real equality. 
The continuous variable income was 
converted to a dichotomous variable 
(because of a skewed distribution of 
income), where if respondent’s total 
family income is 25,000 or less, then, 
they are considered to live in poverty; 

conversely, if the respondent’s total 
family income was above 25,000 dollars 
per year, then they are coded as not 
being in poverty.  
Income level is measured by the GSS 
variable (VAR: INCOME). Respond- 
ents were asked, “In which of these 
groups did your total family income, 
from all sources, fall last year before 
taxes that is?” A fifteen point response 
category was used to capture 
respondent’s answers: under $1,000; 
$1000 to 1,999; $2,000 to 2,999; $3,000 
to 3,999; $4,000 to 4,999; 5,000 to 
5,999; $6,000 to 6,999; $7,000 to 7,999; 
$8,000 to 8,999; $9,000 to 9,999; 
$10,000 to 14,999; $15,000 to 19,999; 
$20,000 to 24,000; $25,000 or over; 
refused; don’t know, no answer; not 
applicable. The variable “INCOME” 
was converted into a dichotomous 
variable: 1) 1= living in poverty (income 
$25,000 or less) 2) 0= not living in 
poverty (income above $25,000).  

 
The Key Independent Variable: Race 

 

The mere distinction of individuals by 
race invariably gives way to the belief 
that slight biological differences between 
certain groups of people predetermines 
the worth, intelligence, value, and other 
aspects of a person’s being. As a 
consequence, race distinction is typically 
followed by the formation of preset 
stereotypes regarding a particular group 
of people and the creation of a racial 
hierarchy. Distinction by race has been 
the catalyst throughout mans history for 
wars as well as hate-crimes, and it has 
caused untold human suffering not only 
in the U.S. but, indeed, throughout the 
entire world. It is this author’s 
hypothesis that race continues to plague 



INCOME INEQUALITY by Terry Holt   5 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

minorities in the U.S. 
Race is measured by the GSS variable 
(VAR: RACE). Respondents were 
asked, “What race do you consider 
yourself?”  Respondents were asked to 
select their appropriate race from a 
three-point scale: White, Black, or other 
(specify). The key independent variable 
“RACE” was dichotomized as follows: 
1) Black or not, and 2) Other race or not.  
 

Other Independent Variables: 
Age 

  
It is my hypothesis that the working age 
of an individual will be positively 
correlated with a higher mean income. 
That is, when people begin to work they 
will often start at the low end of the pay 
scale in their respective occupations. 
However, as they gain more experience 
on the job, their worth to their employer 
increases, and thus they can demand 
higher incomes.  
Age is measured by the GSS variable 
(VAR: AGE). Respondents were asked 
to indicate their age by selection from 
the approximate eight point choice 
category. The categories are listed as 
follows: 10-19 years of age (y.o.a.), 
20-29 (y.o.a.), 30-39 (y.o.a.), 40-49 
(y.o.a.), 50-59 (y.o.a.), 60-69 (y.o.a), 
70-79 (y.o.a), 80 or over, and No 
answer/don’t know. 
REMARKS:  
Respondent’s age: Data has been 
recoded into actual age in cols. 92 and 
93. See Appendix D, and Appendix E. 
Age distribution, for the detailed 
response. The distribution for the first 
digit, col. 92 is given below. See 
Appendix N for changes. 
 
 

Gender 

 
Not only is income level stratified along 
racial dimensions, but also by gender. 
Traditionally, the U.S. has always 
exercised patriarchal domination, and, as 
such, men have characteristically held 
more prestigious employment positions 
that typically pay more. Therefore, I 
hypothesize that the mean income of 
men will be higher than that of women.  
Gender is measured by the GSS variable 
(VAR: SEX). Code respondent’s sex: 
they were asked to indicate their gender 
by using the following two point 
response category: “Male, Female;” 
Male=1, female=2. 
 

Education Level 

 
I hypothesize that higher levels of 
education will be positively correlated 
with a higher mean income. Individuals, 
who have higher levels of education, 
will be more valuable to their employers 
because of special training, job skills, 
and knowledge, which will allow them 
to perform specialized tasks.  
Education is measured by the GSS 
variable (VAR: DEGREE). Respond- 
ents were asked, “What is your highest 
level of education?” Respondents were 
asked to select their appropriate 
education level from a six-point scale 
which is listed as follows: 1) Less than 
high school, 2) High school, 3) 
Associate/ junior college, 4) Bachelor’s, 
5) Graduate, and 6) Don’t know. 
 

Religion One was Raised with 

 
I hypothesize that those individuals who 
were raised in families that regularly 
attended religious services as children 



INCOME INEQUALITY by Terry Holt   6 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

will have a stronger work ethic, than 
those who did not; consequently, those 
individuals who were raised with 
religion will possess a higher mean 
income than those respondents who were 
not raised in a family that attended 
religious services.  
Religion is measured by using the GSS 
variable (VAR: RELIG16). 
Respondents were asked, “In what 
religion were you raised?” Respondents 
were asked to select the religion they 
were raised in by making a selection 
from the following five-point choice 
selection category: Protestant, Catholic, 
Jewish, none, other (specify religion, 
and/or church denomination. The 
variable “RELIGION” was dicho- 
tomized as follows: 1) Jewish or not, 2) 
Catholic or not, 3) Protestant or not 4) 
Other religion or not, and “None” or not 
raised with any religion is the 
comparison group.   
 
Political Affiliation  

 

I hypothesize that those respondents who 
are Republicans (who typically hold 
more traditional values) will be more 
positively correlated with higher mean 
income levels than those of other 
political affiliations. This argument is 
not strong enough   
Political affiliation is measured using 
the GSS Variable (VAR: PARDYID). 
Respondents were asked, “Generally 
speaking do you usually think of 
yourself as Republican, Democrat, 
Independent, or what?” The variable 
“PARDYID” was dichotomized as 
follows: 1) Democrat or not 2) 
Independent or not 3) Other political 
party or not. 
 

THE FINDINGS 

 
Figures 1, 2, and 3 report the frequency 
distributions for the main variables used 
in Analysis. 
 
A. Frequency Distribution of the D.V. 

and the Key I.V. 

 

The main hypothesis of this paper 
singles out the dependent variable as 
total family annual income (VAR: 
INCOME), and this continuous variable 
was converted into a dichotomous 
variable: 1) respondents whose total 
family income was 25,000 dollars or less 
are considered to be in poverty, and they 
were coded as “1”; and 2) those 
respondents whose total family income 
is above 25,000 dollars were coded as 
“0,” and they are considered not in 
poverty. There were a total of 2,812 
respondents. 1, 764 respondents (71.1%) 
reported that their total family income 
was $25,000 or more (not in poverty), 
and 718 respondents (28.9%) reported 
that their total family income was below 
$25,000 (in poverty), and 330 (11.7%) 
respondents showed missing data for this 
question. 

 
The Key Independent Variable 

Race 

 

The key independent variable for this 
study is race (VAR: RACE), and this 
nominal variable was dichotomized as 
follows: 1) Black or not, 2) Other race or 
not (White is the comparison group). 
The frequency distribution for “Black or 
not” is as follows: there were a total of 
2,812 respondents in this study, 377 of 
whom (13.4%) reported that they were 
Black; 2,482 respondents indicated that 



INCOME INEQUALITY by Terry Holt   7 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

they were not Black, and 333 
respondents failed to answer the 
question. The frequency distribution for 
“Other race or not” is as follows: 2812 
respondents took part in the survey, and 
201 persons reported that their race was 
“Other” (7, 1%), and 2,611 respondents 
reported that their race was not “Other.”  

 
B.  Univariate Statistics 

 

Table 1 reports the frequency distribu- 
tions for all the variables used in 
Analysis. The dependent variable 
“INCOME” was converted to a dicho- 
tomous variable (in poverty or not) and 
it has a frequency of 2482, a mean of 
0.2893, and a standard deviation of 
0.45352. Most respondents in the survey 
had incomes that were 25,000 or more 
(not in poverty). 
    The key independent variable 
“RACE” was dichotomized as follows: 
Black or not, and Other race or not 
(Whites are the comparison group). The 
frequency for “Black or not” is 2812 
with a mean of 0.1341 and a standard 
deviation of 0.34079. The frequency for 
“Other race or not” is 2812 with a mean 
of 0.0715 and a standard deviation of 
0.25767. 

“AGE” has a frequency of 2803, 
a mean of 45.96, and a standard 
deviation of  
16.1801. 

“RESPONDENTS’ SEX” has a 
frequency of 2812, a mean of 1.54, and a 
standard deviation of 0.498.  

“RS HIGHEST DEGREE” has a 
frequency of 2811, a mean of 1.61, and a 
standard deviation of 1.207. 

   “RELIGION One was RAISED 
With” was dichotomized as follows: 1) 
Jewish or not, 2) Catholic or not, 3) 
Protestant or not, and 4) Other religion 
or not (No religion is the comparison 
group). “Jewish or not” has a frequency 
of 2809, a mean of 0.0228, and a 
standard deviation of 14924. “Catholic 
or not” has a frequency of 2801, a mean 
of 0.2960, and a standard deviation of 
45656. “Protestant or not” has a 
frequency of 2801, a mean of 0.5598, 
and a standard deviation of 49650. 
“Other religion or not” has a frequency 
of 280, a mean of 0.0421, and a standard 
deviation of 20092.  

“PARTYID” was dichotomized 
as follows: 1) Democrat or not, 2) 
Independent or not, and 4) Other 
political party (no political affiliation is 
the comparison group). “Democrat or 
not” has a frequency of 2800, a mean of 
0, 3425, and a standard deviation of 
0.47463. “Independent or not has a 
frequency of 2800, a mean of 0.3539, 
and a standard deviation of 0.47827. 
“Other political party or not” has a 
frequency of 2800, a mean of 0.0104, 
and a standard deviation of 0.10126. 

 
 

 

 

 

 

 

 

 



INCOME INEQUALITY by Terry Holt

   CALIFORNIA SOCIOLO

 

 

 

 

 

 

 

Terry Holt   

CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 

Figure 1 

Figure 2 

8 

1-17) 

 

 



INCOME INEQUALITY by Terry Holt

   CALIFORNIA SOCIOLO

Terry Holt   

CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 

Figure 3 

9 

1-17) 



INCOME INEQUALITY by Terry Holt   10 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

Table 1 Frequency Distribution 

 

 

 
 

Variables Survey Questions Response Categories Freq Mean S.D. 

Income    In which of these 
groups did your 
total family income 
fall? 

0= Not in Poverty 
1= In Poverty 

2482 .2899 .4535 

Race    What race do you 
   Consider yourself? 

1= White or not 
 
 

2812 1.28 0.586 

  2= Black or not 2812 .1341 .3408 

  3= Other Race or not 2812 .0715 .2577 

Age    Select you age 1=10 to 19 
2=20 to 29 
3=30 to 39 
4=40 to 49 
5=50 to 59 
6=60 to 69 
7=70 to 79 
8=80 or over 

2803 45.96 16.18 

Sex   Select your sex 1=male 
2=female 

2812 1.54 0.498 

Degree What is your highest 
Level of education? 

1=Less than High school 
2=High school 
3=Junior college 
4=Bachelors 
5=Graduate 
6=Don’ know 

2811 1.61 1.207 

Relig16 In what religion were 
you raised? 

1= Jewish or not 2809 
 

.0228 1492 

  2= Catholic or not 
2801 .2960 4566 

  3= Protestant or not 
2801 .5598 4565 

  4= Other religion or not 
2801 .0421 2009 

Pardyid In which political 
party are you affiliated 
with? 

1=Democrat 
2800 .3425 .4746 

  2= Independent or not 2800 .3530 .4783 

  3= Other Pol. Party or not 2800 .0104 .1013 



INCOME INEQUALITY by Terry Holt   11 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

C. Pearson’s Correlation Analysis 

 

 Table 2 reports the bivariate associations 
(Pearson’s correlation) for all the 
variables used in Analysis.  

     As hypothesized, Pearson’s 
correlation supports the premise that the 
total family income of Blacks, on 
average, is likely to fall at or below 
25,000 dollars (poverty). Pearson’s 
correlation is 0.175 (very significant); 
therefore we reject the null hypothesis 
and accept the alternative. That is, there 
is a significant difference between the 
average total families incomes of Blacks 
compared to Whites. 

        Other races appeared to earn 
slightly more than Whites. Pearson’s 
correlation for Other is -0.003. 
Therefore, we fail to reject the null 
hypothesis for Others, and conclude that 
there is no significant difference 
between the average total family 
incomes of others as opposed to Whites. 

     Pearson’s correlation for 
income and other political party is 
-0.029, indicating a slight decrease in the 
number of other political party members 
who have a total family income of 
25,000 dollars or less. Therefore, we fail 
to reject the null, and conclude that other 
political party affiliation does not 
significantly affect the total family 
income of these respondents, as opposed 
to Republicans. 

     Pearson’s correlation between 
poverty and Democrats shows .064. 
Therefore, we reject the null, and accept 
the alternate hypothesis; Democrats are 
significantly more likely to show a 
family income of 25, 0000 or more a 
year. Democrats are less likely to have a 

total family income of 25,000 dollars or 
less.  

     Pearson’s correlation between 
Independent party and poverty is -0.064, 
which is significant; therefore, we reject 
the null hypothesis and accept the 
alternative hypothesis. Independent party 
membership has a significant effect on 
poverty. That is, Independent party 
members are more likely to have total 
family incomes of 25,000 dollars or less. 

     Pearson’s correlation between 
the Jewish religion and poverty is -.052 
which is significant. Therefore, we reject 
the null hypothesis and accept the 
alternative hypothesis. We conclude by 
stating that respondents who are Jewish 
are more likely to have a totally family 
income that is above 25,000 dollars. 

     Pearson’s correlation for 
Catholics is -0.55, which is significant. 
Therefore, we reject the null hypothesis, 
and conclude that Catholics are less 
likely to show a total family income of 
less than 25,000 dollars. 

        Pearson’s correlation for 
Protestants is -0.028, which is insig- 
nificant. Therefore, we fail to reject the 
null hypothesis and reject the alternative 
hypothesis. We conclude that respond- 
ents who are Protestant are not likely to 
have incomes below 25, 0000 dollars. 

     Pearson’s correlation between 
other religions and poverty is 0.017, 
which is statistically insignificant. 
Therefore, we fail to reject the null 
hypothesis and reject the alternative 
hypothesis. Respondents who are coded 
as having Other political affiliation are 
not likely to have total family incomes 
below 25,000 dollars.  

 

 



INCOME INEQUALITY by Terry Holt   12 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

Table 2. Pearson’s Coefficients of Variables Used in the Analysis  

*P<0.05; **P<0.01 (two-tailed tests) 
 

 
    The main limitation of Pearson’s 
correlation is that the observed 
relationship between the dependent 
variable “In poverty or not” may be 
spurious due to the effects of other 
variables in the analysis. The 
multivariate analysis enables us to 
address such spurious relationships that 
may arise due to other control variables 
in the model. I used a dichotomous 
logistic analysis because the dependent 
variable “Income” was converted to a 
dichotomous value (In poverty or not). 
Logistic regression analysis was 
performed using a dichotomous 
dependent variable: 1= “In poverty and 
0= “Not in poverty.” An OLS binary 
regression test was performed and it 
yielded the following results (see table 
3). The likelihood ratio test is 
statistically significant at .05 levels. 

Therefore, we reject the null hypothesis; 
that is, none of the independent variables 
has a significant effect. The reported 
R-square is .2093, meaning that roughly 
21-percect of the variance in the latent 
dependent variable is captured by the 
model. 
    The regression coefficient of 
“Black or not” is positive and significant 
at the .05 level, net the other variables in 
the model. Therefore, we conclude that 
being Black does significantly affect 
one’s income earning potential, and thus 
Blacks are more likely to live in poverty 
than Whites. 
    The regression coefficient of 
“Other race or not” is positive and 
insignificant, net the other variables in 
the model. Therefore, we conclude by 
stating that there is no significant 
difference between Whites and Other 

 1 
In 
Poverty 
or Not 

2 
Black 
or Not 

3 
Other 
Race 

4 
Other 
Pol. 
Party 

5 
Demo. 
or not 
 

6 
Indep. 
or not 

7 
Jewish 
or Not 

8 
Catholic 
or not 

9 
Protestant 
or not 

10 
Other 
Religi
on 

In Poverty 
or not 

Delete This  Row        

Black or not .174**          

Other Race -,003 -.109**         

Other 
Pol.Party 

-.029 -.020 -.015        

Democrat .064** .269** .000 -.074**       

Independent .061** .059** 089** -.076** -.534**      

Jewish or 
not 

-.052** -.060** -.024 .032 .071** -.038*     

 Catholic 
or Not 

-.055** -.191** .121** -.012 .001 .029 -.096**    

 Protestant 
or not 

-028 .201** .198** .-.009 -.063 -.063** -.167** -.731**   

Other 
Religion 

.017 .001 .225** .-.004 .031 .031 -.031 -.136** -.231**  



INCOME INEQUALITY by Terry Holt   13 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

races in their probability of being in 
poverty. Persons who are of other racial 
groups are not significantly more likely 
to live in poverty than Whites. Of 
course, this offers an interesting paradox 
that is exactly why some minority 
groups fair better finically than others.  
    The regression coefficient of 
“Age” is positive and insignificant. The 
effect of age was found to have no 
significant effect on one’s income 
earning potential; therefore, we fail to 
reject the null hypothesis, and reject the 
alternative hypothesis. We conclude by 
observing that one’s age does not 
significantly affect one’s earning 
potential.  
    The regression coefficient of 
“Sex” is positive, and it was found to be 
highly significant. Therefore, we reject 
the null hypothesis and accept the 
alternative hypothesis. We conclude by 
noting that one’s gender has a very 
significant impact on earning potential. 
Finally, women are statistically much 
more likely to live in poverty than men.    
    The regression coefficient of “In 
poverty or not” and “RS Highest degree” 
is negative and quite significant at the 
.05 level, net the other variables in the 
model. Therefore, we reject the null 
hypothesis and accept the alternative 
hypothesis. We conclude by asserting 
that respondent’s educational level has a 
very significant impact on one’s income 
earning potential. Finally, we conclude 
that respondents who obtained higher 
levels of education are less likely to live 
in poverty than the less formally 
educated person. 
    The regression coefficient of 
“Jewish or not” is negative and 
statistically insignificant. Therefore, we 
fail to reject the null hypothesis, and we 

reject the alternative hypothesis. Further, 
we conclude that the probability that 
one’s income is not statistically 
impacted by the fact that they are 
Jewish. That is, being Jewish does not 
appear to affect one’s chances of living 
in poverty, as opposed to individuals 
who were not raised in any religion. 
    The regression coefficient of 
“Catholic or not” is negative and very 
statistically significant. Therefore, we 
reject the null hypothesis and accept the 
alternative hypothesis. The relationship 
between living in poverty and 
Catholicism shows that individuals who 
were raised catholic are less likely to 
live in poverty than those who were not 
raised in any religion. As hypothesized, 
individuals who were raised in Christian 
based religion have a stronger work ethic 
than people who were raised with no 
religious belief.   
    The regression coefficient of “In 
poverty or not” and “protestant or not” is 
negative and significant. Therefore, we 
reject the null hypothesis and accept the 
null hypothesis. We conclude by 
asserting that people who were raised in 
the Protestant religion are significantly 
less likely to live in poverty than 
individuals who were not raised in any 
religion. Again, we see that people who 
were raised with religious belief are less 
likely to live in poverty than those who 
were not; this could of course be 
explained by my hypothesis that 
individuals who were raised with strong 
religious beliefs are more likely to have 
a stronger work ethic, thereby making 
them less likely to live in poverty. 
    The regression coefficient of “In 
poverty or not” and “Other religion” is 
negative and statistically insignificant. 
Therefore, we fail to reject the null, and 



INCOME INEQUALITY by Terry Holt   14 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

we reject the alternative hypothesis. We 
conclude by stating that there does not 
appear to be a significant relationship 
between living in poverty and 
individuals who were raised in other 
religious beliefs, as opposed to those 
who were raised with no religious belief. 
This finding could be explained by the 
fact that many other religions are not 
Christian based, which would support 
my hypothesis. 
    The regression coefficient of “In 
poverty or not” and “Other Political 
Party” is negative, and it is statistically 
insignificant. Therefore, we fail to reject 
the null hypothesis and reject the 
alternative hypothesis. We can conclude 
by stating that other political party 
members are not more likely to live in 
poverty than Republicans.  

    The regression coefficient of “In 
poverty or not” and “Independent Party 
or not” is positive and highly statistically 
significant at the .05 level. Therefore, we 
reject the null hypothesis and accept the 
alternative hypothesis. We conclude by 
asserting that people who are 
Independent party members are much 
more likely to live in poverty than 
Republican (as hypothesized).  
    The regression coefficient of “In 
poverty or not” and “Democrat or not” is 
positive and very statistically significant. 
Therefore, we reject the null hypothesis 
and accept the alternative hypothesis. 
We conclude by stating that people 
whose political preference is associated 
with the Democratic Party are more 
likely to live in poverty than their 
Republican counterpart. 

 
 
D. Logistic Model Selection for Dichotomous D.V. 

Table 3. Model Fit Statistics of Binomial Analysis 

Independent Variables Coefficients Standard Error 

Constant 

Black or not 

Other Race or Not 

Age 

Sex 

RS Highest Degree 

Jewish or Not 

Catholic or Not 

Protestant or Not 

Other Religion 

Other Political Party 

Independent or Not 

Democrat or Not 

-0.786 
0.744 
0.063 
0.005 
0.404 
-0.728 
-0.724 
-0.507 
-0.382 
-0.058 
-0.081 
0.481 
0.406 

(0.274)** 
(0.142)** 
(0.195) 
(0.003) 
(0.098)** 
(0.052)** 
(0.464) 
(0.186)** 
(0.176)* 
(0.282) 
(0.595) 
(0.127)** 
(0.133)** 

Likelihood Ratio Chi-Sq. 390.482**  

Nagelkerke R-Square .209  

*P<0.05; **P<0.01 
 
 

 

 

 



INCOME INEQUALITY by Terry Holt   15 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

 

 

CONCLUSION 

 

To examine how the independent vari- 
able “In Poverty or Not” is affected by 
the independent variables (race, age, 
educational level, religion, and political 
affiliation), we first examined the 
frequency distribution of all variables in 
the study, see table I. Table II shows the 
relationship between the dependent 
variable and each of the independent 
variables using Pearson’s correlation. 
Pearson’s correlation indicates the 
strength and direction of a linear 
relationship between two random 
variables; in this case, the dependent 
variable “in poverty or not” and all 
independent variables are compared 
while negating the affect of other 
variables 
 
Logistic Regression 

 

Logistic regression was performed on all 
the variables in the model (see table 3) 
    The findings above can be 
viewed as both encouraging and as a 
source of concern. On one hand, we can 
clearly see that most people in the 
United States have total family incomes 
above $25,000—that is most people in 
the U.S. are not living in poverty. On the 
other hand however, as hypothesized, 
there is reason for concern regarding the 
relative income of African-Americans in 
the U.S. Despite the efforts of civil 
rights leaders in the turbulent years of 
the 1960s and laws enacted to promote 
racial equality, African-Americans are 
currently far more likely to live in 
poverty than Whites or members of other 
races. Considering the fact that many 

minority groups who recently migrated 
to the United States are not living in 
poverty at the same rates as 
African-Americans suggests that 
African- Americans may be dis- 
criminated against based on preset 
stereotypes that were popularized during 
slavery. That is, in order for White slave 
owners to justify their inhumane 
treatment of slaves, they dehumanized 
them. Insisting that slaves were 
subhuman, Whites were able to justify 
their maltreatment. It is my belief that 
lingering racism against African- 
Americans stems from these 
preconceptions; consequently, there 
remains a high degree of covert racism 
in America with regards to African- 
Americans. In addition, capitalist 
nations, such as the U.S., thrive on 
exploitation. Therefore, one could 
assume that some group(s) may be 
prevented from acquiring equality. It is 
obvious that racism involves complex 
dynamics, and more research is needed 
to shed light on this issue, but it is clear 
that African-Americans are far more 
likely to live in poverty than Whites or 
other races. 
    In this study, one of the most 
troubling findings is the statistical 
significance of gender and one’s income 
earning potential. That is, women are far 
more likely to live in poverty than men, 
as initially hypothesized. The U.S., from 
its inception, has been a male dominated 
society. As such, women are often seen 
as second class citizens. Thus, women 
are relegated to positions that are 
deemed not as desirable as those that are 
typically occupied by men. These less 
desirable positions, typically, pay less 
than the more desirable positions of 
men. Therefore, we expect to see women 



INCOME INEQUALITY by Terry Holt   16 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

being more likely to live in poverty than 
men. Changing attitudes towards women 
and laws enacted to bring about gender 
equality would help to ameliorate this 
problem. 
    The respondent’s highest degree 
of formal education was also found to 
have a very significant effect on the 
probability of living in poverty. As 
hypothesized, respondents who have 
attained more formal education are far 
less likely to live in poverty than those 
with less education. Respondents who 
are more educated are generally more 
valuable to their respective employers 
because of specialized training and 
knowledge. Therefore, the more 
educated person is more likely to a hold 
higher paying position than a less 
educated person. As a consequence, the 
more educated is less likely to live in 
poverty than the less educated person. 
    Respondents who were raised in 
the Catholic and Protestant religion were 
also found to have a significant effect on 
their income earning potential. That is, 
people who were raised Catholic or 
Protestant are far less likely to live in 
poverty than people who were raised 
with no religious belief. As hypothe- 
sized, people who are raised in a strong 
Christian based belief system are more 
likely to have a stronger work ethic than 
those who were raised with no religious 
belief. Traditional values of Christian 
based religious belief systems encourage 
its members to work hard to gain God’s 
favor. 
    As hypothesized, respondents 
who were affiliated with the Independent 
political party were found to be far more 
likely to live in poverty than people who 
were affiliated with the Republican 
party. Independent party members can 

be seen as those who want to radically 
change government. Therefore, we 
would expect them to be ostracized by 
the powers that be; as a result, 
Independent Party members might have 
restricted employment opportunities.  
    The independent variables in this 
model that were not found to be 
statistically significant at the .05 level 
are as follows: other race, age, Jewish, 
other religion, and other political party. 
    Respondents who are African- 
American and women continue to be the 
prime victims of inequality, which 
presents significant barriers for them and 
can be the catalyst for impoverished life 
opportunities. Sadly, many Blacks and 
women are still plagued by apparent 
discrimination; consequently, these 
groups are often mired in poverty, 
deprivation, and despair. Racism, no 
doubt, involves a complex set of factors 
that are beyond the scope of this study. 
For example, some factors that were not 
considered in this study that may 
contribute to lingering racism in 
America are reduced federal funding of 
social programs, economic factors, 
cultural, and psychological damage that 
past racial discrimination has inflicted 
on minorities, especially African- 
Americans. Thus, while this study does 
support my initial hypothesis that ethnic 
discrimination remains a central hind- 
rance to the equality of opportunities for 
many in the United States, the subject of 
discrimination needs further investi- 
gation.   
   
 
 
 
 
 



INCOME INEQUALITY by Terry Holt   17 

   CALIFORNIA SOCIOLOGY JOURNAL, 2010, VOL. 2 (ISSUE 1: 1-17) 

REFERENCES 

 
Allport, Gordon. 1991. “The Nature of 

Prejudice” Political Psychology       
12(1):125-157. 

Sears, David O. 1995. “The Scar of 
Race.” American Journal of 
Sociology. 100(5)1351-1354  

 

 
 

Sears, David O., and P.J. Henry. 2003. 
“The Origins of Symbolic Racism.”  

 Journal of Personality and Social 
Psychology 85(2):259-275. 

Parks, Robert. 1950. “Race and 
Culture.” American Sociological 
Review 15(3)447.  

 
 
 

 


