Free Research Paper About Statistical And Computer Methods For Social Research

Introduction

Statistics and probability theory is widely used in social researches. In this paper I will show how the basics of this theory may be applied to a real world problem. My goal is to investigate how education level and race are related. I believe that there is a significant difference between race and level of education. It is well known that the problem of racism is still relevant in present days. The recent researches shown, that racism is inherent to most educated people. But they are carefully hiding this fact. Taking in consideration that white people are more likely to show their negative attitude to other races, I would like to investigate if whites are more likely to have better education than other races.

Hypotheses

Highest the parents education, higher the respondent’s education attainment
Higher the family income, higher the respondents education attainment.
White people are more likely to have better education than other races
Literature Review
It is believed that educated people have a broad view of the world and tolerance, recognize the need for talent equal rights to all people regardless of race, color or creed. American scientist proved that it is not (Anderson, J., 1988).
Racial discrimination in the United States has deep historical roots and, in fact, is the same age as the country. Society based white people always have a negative attitude towards people with a different skin color: Indians - Native Americans and blacks. Racial segregation in the United States has officially existed since 1865, since the adoption of the Thirteenth Amendment to the Constitution, abolishing slavery. Examples of racism can serve as such egregious facts as segregation in public transport and schools. According to the laws of Montgomery (Alabama), the first rows of seats on the bus destined for white passengers, and if they were all busy, negros had to give up their "black" space. This policy resulted in a boycott of "Walking in the name of freedom", which lasted 381 days. During this time, African Americans accounted for 70% of all passengers who do not use public transport. Bus companies have suffered significant losses. December 20, 1956 segregation in public transportation has been canceled. Segregation in schools led to armed conflict. The most famous incident refers to 1957, when racists prevented coeducational white children and nine blacks earned the nickname "Little Rock Nine" (Arkansas). When the situation finally escalated, President Eisenhower was forced to send troops into the city in 1200 Airborne Division. They accompanied the children to school and protected them in the hallways. In 1941, racial segregation was officially abolished, but in fact it continues today. Professor Gary Orfild Project report for Civil Rights at Harvard University (end of 2006) said: "The level of segregation in the country rose to the level of the late 1960s. We lost almost all of the progress made during the abolition of segregation in urban communities." According to sociologists, the basis of racism is a realistic theory of group conflict. Central to this theory occupies a position of conflict of interest, which is the main cause of conflict and contribute to the perception of the threat posed by a foreign group. Threat causes hostility members to its source, strengthens its cohesion and increases penalties for people with behavioral problems. With regard to racial discrimination theory of group conflict will be as follows. Historically, the white population of the United States has the power, control the natural resources. As the dominant group people with white skin color sought to preserve the status quo, change the routine they were unprofitable. However, representatives of the black population would improve their situation; their actions were a threat to the dominant group. Therefore, in this situation, racism is an integral element of ideology, social policy. Currently, there are plenty of studies aimed at understanding the phenomenon of racism. As a consequence of the results of opinion polls among scientists strengthened following opinion.

The better the education that a person has received, the less chance that it will have racial prejudices.

It seems quite logical: an educated person understands the political, social and economic reasons for the origins of discrimination, aware of the dangers and possible consequences of racism. Moreover, people with higher education is tolerant and has a broad view of things, it is objective. He understands that the intellect or character of a person does not depend on the color of his skin. Professor, University of Michigan Geoffrey Whatcom disputes this theory. He is repelled by the theory of group conflict and says white people really consider the action of African Americans seeking to achieve equal rights as a danger. That is why the word of the white man, "I do not want to live next door to a Negro" are not due to a lack of thought or fear, and the support of the state ideology, the stability of society.
Geoffrey Whatcom wondered if his thoughts are correct, not whether higher education is further aggravated the situation. After all, the man who has achieved a certain status and take their place in society, it is much more reason to maintain the existing order of things. In his paper, Professor Whatcom used survey data General Social Survey from 1972 to 2010. General Social Survey - is a case study, which was conducted in the United States each year from 1971 to 1994 (except 1979, 1981 and 1992), and is now held every two years. The survey takes about 90 minutes, and involves US citizens over 18 years old. Questions asked them to relate to all the social life of the state tax policy to faith in God. Geoffrey Whatcom included in his work answers 44,873 people, African-Americans were not among them, 56% were women. The average age of participants - 46.5 years, mean duration of training in secondary and higher educational institutions - 12.9 years. Date of birth ranged from 1883 until 1992. The results of his research, Professor Whatcom presented at the 108th annual meeting of the American Sociological Association (Gill, D., 1992).
All respondents were divided into three groups depending on their level of education. People were asked four types of questions. The first concerned the prejudice against black people: the participants were asked to rate on a seven-point scale, how those or other qualities are inherent in African Americans (for example: 1 - lazy 7 - hardworking). The second group of questions was designed to determine the extent to which people are in favor of equal rights (their attitude towards co-education of children in school, and so on. N.). Also, the study participants were asked to rate the degree of discrimination against black people ("weak", "moderate", "strong") and give your opinion about specific laws against racism. The results were shocking: 38.1% of white Americans believe that blacks are lazy, 18.3% - uneducated, while 52.2% admit to the existence of discrimination and agree that it must be fought. 29.3% do not want to have black neighbors, but 73.6% in favor of the development of anti-discrimination legislation in the field of housing policy (open housing laws). 88.8% of respondents stated that they supported the joint training of black and white children, but only 23.1% actually want their children went to school on the same bus with dark-skinned. 71.5% admit that blacks are discriminated against in the labor market, but only 11.8% agree with the adoption of laws fighting this injustice (Lynn, R., 2008).
Obviously, the word white Americans differed from their actual position. However, after analyzing the survey found another feature: the level of education and the degree of racist statements really related, but not so, as previously thought. Whatcom professor was right among the educated, the percentage of those who acknowledge the existence of racism and affirms the need to deal with it, is much higher than among non-Americans have received a good education. The percentage of those who are willing to go for real action and approves anti-discrimination laws are much lower. Uneducated same people for the most part, recognizing the problem and the actual support equal rights. Geoffrey Whatcom, received confirmation of his theory, and puts forward another explanation for this situation: white Americans do not support anti-racism measures not only to protect and preserve their own privileges, but simply because they are not interested in the restoration of social justice. They do not think that inequality is better, but just do not see any reason to change anything. Moreover, the introduction of new laws intended to increase state intervention in some sectors of society, and that no one wants. In general, the liberal educated white people at heart support racism - and, of course, for the benefit of society.

Variables, Methods and Research

In this case I take the following variables from the data to work with:
Dependent variable:
educ – education attainment

Independent variables:

sex – respondent’s gender
maeduc – highest year of school completed by mother
paeduc – highest year of school completed by father
race – the race of respondent
coninc – family income in constant dollars

Begin with descriptive statistics of the variables participated in analysis:

Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
educ | 4814 13.65538 3.064623 0 20
paeduc | 3623 11.55727 4.2855 0 20
maeduc | 4353 11.61728 3.72499 0 20
sex | 4820 1.557676 .4967138 1 2
race | 4820 1.314938 .6172327 1 3
-------------+--------------------------------------------------------
coninc | 4374 49893.88 46679.36 383 178712.5
We can see that the highest level of education varies from 0 to 20, and the mean value of respondents’ education level is higher than their parents’ level.
Proceed with hypothesis testing. I would like to check the first hypothesis: Highest the parents’ education, higher the respondent’s education attainment. The data is numerical and thus I would like to use Pearson’s r to check the association:
| educ paeduc maeduc
-------------+---------------------------
educ | 1.0000
paeduc | 0.4630 1.0000
maeduc | 0.4504 0.6943 1.0000
The coefficient of correlation between educ and paeduc is 0.4630 which is the evidence of moderate positive linear association between the variables. The coefficient of correlation between educ and maeduc is 0.4504 which is the evidence of moderate positive linear association between the variables.

Thus, the hypothesis of that highest the parents’ education, higher the respondent’s education attainment is approved.

The second hypothesis to check is whether family income and person’s education attainment are associated. Complete the same procedure and obtain the following output:
| educ coninc
-------------+------------------
educ | 1.0000
coninc | 0.4015 1.0000
The coefficient of correlation is 0.4015 which means moderate positive linear relationship between the variables. That’s why I approve the second hypothesis.
The third hypothesis is that white people are more likely to have better education than other races. Look at the histogram of education attainment by race:
We are interested to check if the difference between races in education attainment is significant or not? To check this I firstly can run ANOVA test to see if there is a difference in education level between all races:

Analysis of Variance

Source SS df MS F Prob > F
Between groups 460.414495 2 230.207248 24.75 0.0000
Within groups 44742.8612 4811 9.30011664
Total 45203.2757 4813 9.39191266
Bartlett's test for equal variances: chi2(2) = 86.2162 Prob>chi2 = 0.000
Since p-value of the test is lesser than 0.001 we can conclude that there is a significant difference in education level between the groups (at 1% level of significance). Now we sure that the difference exists and looks at the descriptive statistics of educ by race:
-> race = white
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
educ | 3696 13.8244 3.031144 0 20
-> race = black
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
educ | 721 13.14979 2.599632 0 20
-> race = other
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
educ | 397 13 3.869069 0 20
We can see that the typical value of educ for whites is 13.8244 and it is the highest value that’s why I approve the third hypothesis as well.

The next step of this analysis is to regress educ by race:

Source | SS df MS Number of obs = 4814
-------------+------------------------------ F( 1, 4812) = 45.93
Model | 427.377808 1 427.377808 Prob > F = 0.0000
Residual | 44775.8978 4812 9.30504943 R-squared = 0.0095
-------------+------------------------------ Adj R-squared = 0.0092
Total | 45203.2757 4813 9.39191266 Root MSE = 3.0504
educ | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
race | -.4829669 .0712641 -6.78 0.000 -.622677 -.3432567
_cons | 14.29034 .1034939 138.08 0.000 14.08744 14.49324
The model is significant, however, it is very weak for predictions, because R-squared is only 0.0095. This means that the dependent variable’s variance is explained by this model for only 0.9%. That’s why I add other factors which were notified in the beginning and run multiple regression:
Source | SS df MS Number of obs = 3109
-------------+------------------------------ F( 5, 3103) = 269.24
Model | 8177.0202 5 1635.40404 Prob > F = 0.0000
Residual | 18847.8495 3103 6.07407333 R-squared = 0.3026
-------------+------------------------------ Adj R-squared = 0.3015
Total | 27024.8697 3108 8.69526053 Root MSE = 2.4646
educ | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
paeduc | .1705564 .0144088 11.84 0.000 .1423046 .1988082
maeduc | .1675459 .0167433 10.01 0.000 .1347168 .200375
sex | .2127041 .0890889 2.39 0.017 .0380249 .3873832
race | .0955506 .0760501 1.26 0.209 -.0535631 .2446643
coninc | .0000164 9.46e-07 17.28 0.000 .0000145 .0000182
_cons | 8.83226 .2435619 36.26 0.000 8.354701 9.309819
This model is also significant but R-squared is not that high. Only 30.26% of the educ variance is explained by this model. However, race variable is insignificant in this model (p-value is 0.209). That’s why it should be exluded.
Now I’m interested to develop various regression models of 2, 3 and 4 independent variables and make a summary in a descriptive table.
I regress educ and paeduc, maeduc and call it Model 1, regress educ and paeduc, maeduc and sex and name it Model 2, educ and paeduc, maeduc, sex and race – Model 3. The full model mentioned above is Model 4.

The results are given in the table below:

Conclusion
There are three hypotheses were checked in this study. The hypothesis about the association between parents’ education and person’s education was approved as well as the hypothesis about the association between family income and person’s education level. The hypothesis about that white people is more likely to have better education than other races was also approved. But it is appeared that race is not very significant factor which affects education level. There are also a number of other factors, such as parents’ education level, sex of an individual, family income. Multiple regression analysis shown that there are possible a number of factors which weren’t included in the model (because the goodness-of-fit R-squared indicator is quite low). The model may be reworked after receiving some recommendations from sociologists – which variables to include.
The most accurate model is multiple regression between education of the person and his parent education level (Model 1). The other models are not getting better because the factors included are not significant to predict education attainment.

Works Cited

Anderson, James D. The Education of Blacks in the South, 1860-1935. Chapel Hill: U of North Carolina, 1988. Print.
"Black Men Need More Education Than White Men To Get Jobs 2015 " Web. 17 Mar. 2015.
Gill, Dawn. Racism and Education: Structures and Strategies. Newbury Park, Calif.: Sage Publications in Association with Open U, 1992. Print.
Hacker, Andrew. Two Nations: Black and White, Separate, Hostile, Unequal. New York: Scribner's ;, 1992. Print.
Lynn, Richard. The Global Bell Curve: Race, IQ, and Inequality Worldwide. Augusta, Ga.: Washington Summit, 2008. Print.
Troyna, Barry, and Jenny Williams. Racism, Education, and the State. London: Croom Helm, 1986. Print.
Troyna, Barry, and Bruce Carrington. Education, Racism, and Reform. London: Routledge, 1990. Print.
do.file
*descriptive statistics:
. summarize educ paeduc maeduc sex race coninc
*correlation between person’s education and parents’ education
. correlate educ paeduc maeduc
*correlation between family income and person’s education
. correlate educ coninc
*histogram of education by race
. histogram educ, by(race)
*checking the difference in education level by race
. oneway educ race
*descriptive statistics of education level by race
. by race, sort : summarize educ
*simple linear regression of educ and race
. regress educ race
*Model1, 2, 3:
regress educ paeduc maeduc
regress educ paeduc maeduc sex
regress educ paeduc maeduc sex race
*multiple linear regression between educ and dependent variables
. regress educ paeduc maeduc sex race coninc