13 minute read
RACE AND CONFIDENCE IN AMERICA’S EDUCATION
Sophia Nguyen
Major: Government and International Politics
Concentration: International Relations
Minor: Global Affairs
Class of 2023 ABSTRACT
This paper analyzes the question of whether one’s race impacts one’s confidence level in America’s education system. Specifically, if black Americans have less confidence in America’s education system versus white Americans. Using Statistical Package for the Social Sciences (SPSS) and scholarly articles, we statistically analyze the two variables to determine if there is a correlation. The end result of my data analysis did not support my hypothesis that black Americans have less confidence in America’s education system versus white Americans, however, it does support the fact that race has a relationship with one’s level of confidence in education.
Introduction
Do African Americans show less confidence in the United State’s education system? This has been complex yet continuous research, but the answer to this question can further acknowledge a problem in America’s education. This paper theorizes that race is a predominant factor as to why African Americans are more likely to show less confidence in the education system and the following theoretical concepts support this hypothesis:
1. Better educational opportunities in white institutions during the 1960s:
During the 1960s when African Americans first were granted the opportunity to have an education, black students began to enroll in predominantly white schools rather than choosing to attend historically black institutions. This was generally because of the greater opportunities that were accomplished through these white institutions, and how it was widely known that black institutions during the 1960s were purposely underfunded, understaffed, and established to separate the achievement gap between white and black students (Hartney & Flavin, 2013).
2. Reduced fundings towards education in low-income cities: Low-income cities, most of which are predominantly black, tend to have less funding towards their education, compared to its higher-income, more predominantly white counterparts. This creates an achievement gap between black and white students, furthering racial disparities in education (Hartney & Flavin, 2013).
3. Racial Prejudice in Academics: Not only do underserved black families in America face racial prejudice inside of academic settings, but many black Americans still face racial discrimination in higher levels of education like being subjected to slurs and jokes, being seen as if they were not smart enough compared to their peers, and being unfairly suspected by their teachers, peers, and polices on campus. (Anderson; 2019).
While I theorize that race is a determinant of an individual’s confidence level in America’s education system, there are also numerous intervening variables that can alter the way we see the dependent variable. For example, income is a good intervening variable to act as a control group as it will be used to interpret the independent and dependent variables in SPSS. Despite my hypothesis, I hypothesize that higher income black Americans will have more confidence in America’s education system rather than lower-income black Americans.
I am aware of my limitations throughout this paper and acknowledge the fact that race cannot stand on its own in comparison to one’s level of confidence in education, as there are many complex variables that surround racial bias and inequality. That being said, this paper does not attempt to “prove” the hypothesis to be true, but rather uses existing research data to support the once questioned hypothesis.
Literature Review
Public trust in education has fluctuated throughout the years. While there are many variables to choose from for an independent variable, it is hopeful that supporting variables like age, income, ideology, achievement rate, and demographics are supportive to the independent variable of race and how a person’s color affects their level of confidence in education. This literature review attempts to look back at past research and add to the question of if race determines one’s level of confidence in education and if African Americans tend to have less confidence in education.
Joshua Klugman and Jun Xu (2008) reviewed confidence levels in education between white and black Americans from 1974-2002. They theorized and tested that presidency, income, and higher education levels were associated with one’s level of confidence in their education. In contrast to my theory, Klugman and Xu hypothesized white Americans were more likely to have less confidence in education, as they tended to be more conservative, more religious, and have more income than black Americans, making them more critical towards education. Klugman and Xu’s data stated that white Americans’ confidence in education was always lower than black Americans, but while both white and black confidence had lowered between the years 1974-2002, many variables caused it to fluctuate (Klugman & Xu; 2008).
For example, the variable presidency was correlated with the fluctuating levels of confidence. They claimed that during Republican presidencies, whites were more confident in their education than blacks, rather than when Democrat Jimmy Carter and Democrat Bill Clinton came into office in 1977 and 1992, black confidence significantly increased as white confidence lowered. Overall, Klugman and Xu concluded that white Americans had lower confidence in education than black Americans (Klugman & Xu; 2008).
David R. Johnson and Jared L. Peifer (2015) succeeded Klugman and Xu’s theory and tested whether political ideology, religion, income, and demographics were associated with the confidence level of education (Johnson & Peifer; 2015). They both argued against Klugman and Xu and claimed that blacks had an overall lower level of trust in education compared to whites. Using income as a controlled variable, they both found that higher-income, more conservative blacks had a higher level of trust. Johnson and Peifer stated that with using blacks and education as covariates, they were the only variables that had significantly low levels of confidence. It was concluded that simply due to inequality within secondary and primary education for black students, blacks were more likely to be untrusting towards education, especially in higher-level institutions (Johnson and Peifer; 2015).
Jean W. Yeung and Dalton Conley (2008) explored the black-white achievement gap to see if there was any correlation between family income and the gap between black-white achievement in education. Three variables, occupation, income, and socioeconomic status, were the main determinants to one’s achievement rate. (Yeung & Conley; 2008). Based on the results, black students had significantly lower levels of academic achievements compared to white students. It was concluded that racial disparity in wealth was clear between black and white families with the way that white families were more likely to have a higher level of academic achievement because of their higher level of income and status.
In conjunction with Yeung and Conley’s hypothesis, Michael T. Hartney and Patrick Flavin (2013) furthered the belief that political ideology and political foundations separated the black-white academic achievement gap, thus causing black families to have a lack of trust in their education. They theorized that due to the states’ legislators lack of attention to improving school resources and teacher quality for underserved, black families, this was a determinant to their lagging achievement rates. It was concluded that not only were education reform policies more responsive towards white communities rather than black communities, but whites were more likely to dismiss the identity of a racial achievement gap. (Hartney & Flavin; 2013). As a result, systematic political inequalities distinguished the academic achievement gap between blacks and whites.
James P. Huguley, Eric Kyere, and Ming Te-Wang (2018) agreed with Hartney and Flavin in terms of black Americans’ lower levels of confidence, yet investigated that parental confidence in their children’s long-term education determined educational outcomes within black students. They theorized that because black parents were more likely to have lower confidence in both their children’s long-term and shortterm education, they were more likely to have less achievement in their education. It was suggested that the lack of confidence towards higher level education by black parents had more of an effect towards their children’s academic achievement. Because of their opinion towards educational disadvantages and racial achievement disparities in higher level education for black students, black students tended to lagged behind white students in terms of overall academic success (Huguley et al., 2018).
Methodology
I hypothesize that race is the dominant factor of one’s level of confidence in education. Furthermore, because I hypothesize that black Americans are more likely to have less confidence in education in contrast to white Americans, the null hypothesis is that there is no correlation between one’s race and levels of confidence in education. To test the hypothesis, I will have to use the variables race_2 (race: black / white) as my independent variable and coneduc (confidence in education) as the dependent variable in data set GSS 2012 within SPSS. The independent variable race_2 is a dichotomous variable that asks respondents to identify themselves as white or black. The ordinal dependent variable coneduc is categorized by a great deal, only some, and hardly any.
However, I will recode coneduc to coneduc3 so that it reads as hardly any, only some, and a great deal.
To confirm or reject the null hypothesis, we will have to use various forms of measurement. Because this is an ordinal by dichotomous variable, we will use Chisquare and Phi to measure the significance and strength of association between race and confidence in education. I want to be at least 99% confident that I can reject the null hypothesis, therefore, I will need to have an alpha value of .001 or lower. If an alpha value of .001 or lower can be identified, then I am less likely to make a type 1 error and will further calculate strength and direction between the two variables.
While I hypothesize that race is a determinant of one’s level of confidence in education, I have yet to consider other variables that might affect one’s opinion. In previous research, a person’s income has been a recurring variable to measure one’s level of confidence in education. In this case, income will be used as a control variable in SPSS. The ordinal variable rincom06 ranges from under $1,000 to $150,000 and over. Since there are twenty-five categories, I have recoded this to rincom3 and into 3 categories: low income (under $1,000-$12,499), middle income ($12,500-$49,000), and high income ($50,000-over $150,000). Recoding the control variable to have fewer categories as an ordinal variable will allow the variables to compare with one another effectively and efficiently.
As Michael A. Miner (2020) and Johnson and Peifer hypothesized in their research, I also believe that black Americans who have more income are more likely to be confident in education than lower income black Americans (Miner; 2020). Because higher income individuals are able to afford better education and resources, I hypothesize that they will have more confidence in education than those who have lower income and cannot afford better education and resources. To test significance, I will have to use both Chi-square and Phi again to check for significance and strength of association.
Data Analysis
Using Statistical Package for the Social Sciences (SPSS), I have conducted two data analyses. The first analysis will be conducted with the dichotomous independent variable race_2 that asks respondents if they identify as white or black, and the recoded ordinal dependent variable coneduc3 that measures level of confidence in education.
Using SPSS to conduct the first statistical analysis between the independent and dependent variables, the results are as follows. The cross tabulation, bar chart, and statistical tests goes against how I expected the variables to be related. Table 1.1 in the first row shows that white Americans (17.7%) have less confidence in education than black Americans (11.1%). Table 1.1 in the last row, black Americans (36%) shows more confidence in education than white Americans (21.8%). It is noteworthy that in Table 1.1 in the second row, over half of the respondents (58.3%) are in the middle in terms of their confidence in education. There is a statistically significant relationship, as X2(4) is 33.388 with the degrees of freedom being 4, and the p-value being less than .001. The alpha-value allows us to be at least 99% confident to reject the null hypothesis and not make a Type 1 error, as the relationship between the variables does exist in the population. However, although we are at least 99% confident that there is a relationship between race and levels of confidence, it does not support my hypothesis that black Americans have less confidence in education.
In Table 1.1, the degree of the relationship is calculated in Phi. As Phi = .159, this shows a weak relationship between the two variables. This confirms that although there is a statistically significant correlation between race and their levels of confidence in education, it is a rather weak relationship. Overall, my hypothesis that black Americans are more likely to have less confidence in education was not supported by the data analysis. Because Phi calculates the degree of the relationship, there are other potential explanations that I could use to alter the dependent variable, which brings me to my control variable rincom3.
TABLE 2.1: Analysis 2- Introduction to Control Variable
For the second analysis, income was introduced as a control variable and the results are as follows. As seen in Table 2.1, there are many differences between low income Americans and high income Americans in terms of their confidence in education.
For low income Americans, with X2(4) being .786 and the alpha-value (.940) being greater than .1 , it does not nearly reach the 90% confidence, and I have to assume that there is no relationship between low income Americans and how it affects blacks’ and whites’ level of confidence in education. This results in me having to fail to reject the null hypothesis and risk making a Type 2 error, as there is no statistically significant relationship. For Table 2.1, although low-income white Americans (11.2%) have less confidence in education than low-income black Americans (6.7%), compared to how low-income black Americans (40%) have more confidence in education than lowincome white Americans (32.7%), the statistical results show inconsistency and no relationship between the variables. This concludes that race does not have a meaningful relationship in the low income category. One limitation to this category is the small sample size of 133 respondents. If the sample size was increased and results still showed inconsistency and no relationship, I would be more confident that there is no relationship in the population and that I would no longer risk making a Type 2 error.
For high income Americans, with X2(4) being 15.200 and the alpha-value (.004) being less than .01, I can be at least 99% confident that there is a relationship between high income Americans and how it affects one’s level of confidence in education. With the alpha value, I am at least 99% confident that I can reject the null hypothesis and not make a Type 1 error. In Table 2.1, although high income white Americans (21.2%) have less confidence in education than black Americans (11.9%), high income black Americans (26.2%) have more confidence in education than white Americans (15.4%). Because Phi calculates the degree of the relationship and is .210, this indicates a moderate relationship between high income and an individual’s level of confidence in education. When talking about Phi in the middle income category, because Phi = .252 and is slightly higher than the Phi value in the high income category, I can conclude that while there is a relationship between race and both middle and high income Americans, race and the levels of confidence has more of a relationship in the middle income category. Black Americans had higher levels of confidence in all low, middle, and high income categories. Despite the low income category having no statistical relationship between race and levels of confidence, this again goes against Miner’s and my hypothesis that higher income black Americans are more likely to have more confidence in education (Miner; 2020).
In all three categories, white Americans had less confidence in education while black Americans had more confidence in education. My hypothesis that black Americans have less confidence in education was challenged and further not supported. In conclusion, race does determine one’s confidence level in education and is proven to be supported based on my data analysis. However, while it is not in favor of my overall hypothesis, it is in favor of Klugman, and Xu’s hypothesis that white Americans have less confidence in education (Klugman & Xu; 2008).
DISCUSSION/CONCLUSION
The research question and the theories behind it were based on how race can determine one’s level of confidence in education. With higher opportunities within white institutions than black institutions, reduced fundings towards low-income cities that are predominately black, and racial prejudices towards black students in higher institutions, this led me to believe that race was a determinant of one’s level of confidence in education, and how black Americans were more likely to have less confidence in education. The GSS data showed a statistically significant yet weak relationship between race and the level of confidence in education and went against my hypothesis that black Americans have less confidence in education. With using income as an intervening control variable, it also did not support my hypothesis that high income black Americans have more confidence in education than low income black Americans, because they overall had more confidence in all categories. However, because white Americans overall had less confidence in education despite using income as a control variable, the relationship between blacks’ and whites’ confidence in education was not at all in my favor and requires further exploration and perhaps a larger population size.
One limitation that I encountered with the statistical data was the bias that could have been present during the time of the survey. Year 2012 was Democrat President Obama’s second term in office, and because white Americans leaned more conservative and against President Obama, thus making them more critical towards his policies and reforms, this could have led them to be less confident in education at the time of the survey. This was something that was tested in Klugman and Xu’s research and was further explored in my research as well. (Klugman & Xu; 2008). I would like to someday view a 2020 dataset with the same variables, as I predict that with Donald Trump being president at the time, black Americans are more likely to have less confidence in education.
Race cannot stand for itself in terms of one’s level of confidence in education, and this is proven by the different statistical significance and relationship strengths I encountered while using income as a control variable. Although race does appear to be a factor to one’s opinion on education, a much larger sample of respondents, perhaps a mixture of nation-wide and in-person respondents and an updated survey centered around racial biases should be conducted so that I can be more positive in not making a Type 1 error. Confidence in education has fluctuated through the years, and the lack of updated data surveys has made this paper limited in what I can achieve. My hypothesis that black Americans have less confidence in education was not supported, but the relationship between race and levels of confidence has a statistically significant, yet weak relationship. I am determined in the future to revisit this research paper with a newly designed data set and a much larger sample size.