Social research Essay

This essay has a total of 2870 words and 12 pages.

social research

Social Research Methods Sahar Thariani
Paper II Section 01


Introduction and Data Source
Attending college is slowly changing from what was once considered a rare opportunity to a
staple part of what constitutes an education today. As the number of colleges has also
inflated, and means of attending college expanded, such as Internet based universities,
the number of people attaining a higher-level education has also increased. This paper
attempts to test and analyze fifty American states and conclude upon factors within states
that may give an individual a better chance of being college educated. The three variables
being tested in this research include median household income, race and Internet access.
In order to do this, statistical data had to be gathered for all the states, these fifty
being my unit of analysis. To ensure accurate results, the statistical data had to be
collected from a reliable source. The numbers used as indicators of educational
achievement and households with Internet access were obtained from the official website of
the U.S Census Bureau. A governmental institution, well known for its detailed statistics
on every state, provided a set of figures that would be most reliable. Data for median
household income for each state and population distribution by gender was acquired by an
organization referenced by Professor Hansell, an acclaimed sociologist. "State Health
Facts online" supplied by the well-reputed Kaiser family Organization is a resource that
contains the latest state-level data on demographics, health, and health policy. The
website also has a section of raw data through which one may verify the statistics.


Hypothesis
The aim of this study is to find issues within states that result in higher education
levels, that is, factors that education is dependant upon. This makes education the
dependant variable in this study. Higher education is usually expensive, and thus often
limited to those that can afford it. In addition to this, individuals growing up in
wealthier households may be more exposed and educated with a stronger motivation to study
and learn. Once having earned a university degree, one may demand a higher salary, and
having been brought up in richer homes, individuals may also feel more pressured by family
to attend an institute of higher education in order to earn more. Hence, my primary
independent variable affecting education levels is median household income. While I
believe that income will have a strong impact on education, as a higher income should
result in higher education, there may be other independent variables that affect education
levels. One of these test variables is race. Through this analysis I want to assess the
role of race where higher education is concerned. As a third variable, this will help
determine if being White-American can actually increase ones chances of attending college.
Lastly, I hypothesize that households that have relatively more access to the Internet
should have higher levels of education. People with Internet are automatically exposed to
boundless information, and may take up virtual classes. Also, people with Internet access
must have a higher median household income than people without Internet access, and the
reasons behind a higher household income affecting education will then apply. In addition
to this, having the Internet may expose people more to the importance of education and its
availability, and ultimately boost education levels.


Univariate Descriptive Statistics
Having gathered all the data for each variable for every state, they had to be arranged in
a data matrix so they could all be viewed in relation to each other. In the data matrix,
all fifty states are listed and to their right is the data for each of the four variables,
starting with the dependant variable, education, followed by the household income, race,
and internet access. For the data for each variable, statistical tests were taken to put
the data into perspective.

Because the purpose of this paper analyzes the factors contributing to higher education,
the dependant variable education was measured in terms of the percent of people in each
state earning a Bachelors degree or more. Once a complete list of percentages of people
receiving a higher education was prepared, the percentages were split up into being either
a high percentage, symbolizing a large number of college degrees earned, coded as a ‘2',
or a low percentage, meaning that the state had a relatively smaller number of highly
educated people, coded as a ‘1'. I determined each case as being a high or low education
state by classifying those below the average, around 25 percent to be low education
states, while those at the average or higher, as being high education states. Statistical
operations concluded that the average percent of people from all fifty states equaled
24.932, or approximately 25 percent. The median gave the figure 24.45, and the mode 24.6,
being the case for three states out of the fifty. These measures of central tendency imply
that the data is not skewed as the mean and median are extremely close to each other. Upon
constructing a frequency table, it could also be determined that 22 states had a college
graduate level lower than the average, while 28 states had a percentage of graduates at or
above the average. Next, the dispersion of the data is examined through the standard
deviation, smallest data value and largest data value, and the range, which is the
difference between the two. West Virginia, with only 15.3% college graduates had the least
of all the states, while Colorado had the highest percentage of 34.6%. None of these
figures were real outliers to the data collection as they both fall within three standard
deviations of the mean. The standard deviation for education levels equal 4.27, implying
that 99% of data should lie within three such standard deviations, implying a data range
of 12-37%- In fact 100% of the data falls within this range in an evenly distributed bell
curve. The range of the data is 19.3%, that is, the difference between the highest and
lowest value.

Next the original independent variable, household income was set up into a frequency
table. Household income for every state was defined as the median household income per
state. Once all the median incomes for each state were listed, they were re-coded as being
a high median income or relatively low median income for every state. The cutoff for being
a high household median income state was annual earnings of $28964, the approximate
average of all the household incomes. States at or above the average were considered high
income states and coded as ‘2', while those states below the average were coded as
‘1', for low income states. A total of 23 states had a high income, while 27 an income
lower than the mean. The median of the median household income figures is $29435, and
there is no mode, as no two states had the same median household income. Once again, the
data lacks any outliers, and the median and mean are close, implying an un-skewed
collection of data. The standard deviation equaled $4030, and most of the data lies within
two standard deviations of the mean. Minnesota boasts the highest median household of
$38,200, while Louisiana reports the lowest of $21030, giving a range of $17180.
Dispersion is closely centered about the mean.

The third variable, race, was measured as the percentage of White Americans in every
state. Once each state was listed, in the frequency table from lowest to highest
percentage of whites, I split the data about the median, considering states with 81%
whites or more to be states with high white populations denoted by ‘2'. The remaining
states with 80% or less white were considered to have a relatively lower white population
and were coded as ‘1'. This split the states evenly, with 25 states in each category of
high or low income. The mean percentage of white Americans is 76%, ranging from a maximum
of 97% in Vermont, to only 23% in Hawaii. This gives the data a range of 74. The 23% white
population in Hawaii presents an outlier from the regular distribution the data. This
skews the data and makes the mean differ from the median. The modes are 88% and 91%, each
appearing 4 times in the data set. The calculated standard deviation of 15% implies that
99% of the data should fall within 45% of the mean. Hawaii's racial distribution lies
outside this range, and thus even further implies that it is skewing the data.

Internet access, the fourth variable in this study was measured as the percentage of
households with Internet access in every state. Arranging these percentages from lowest to
highest, the lowest percentage of Internet access equals 36.1% for Minnesota, and the
highest, 64.1, for Alaska. The range of these percentages is 28. The average percentage of
households with Internet access amounts to 50, the mean, and the median is 50.9. This
closeness between the mean and media shows that the data is evenly distributed and
un-skewed. The standard deviation is 6.3 % from the mean, and all data falls neatly within
three Standard deviations.




Original Relationship Between Independent and Dependant Variable
A cross tabulation of the original dependent and independent variable, education and
income respectively, describe the original relationship between these variables. Using the
recodes of 1 and 2, the table presents a count of the states in combined categories. In
this bivariate table, one can see that 19 out of 23 states that have a low income also
have low education, while only 4 states out of 23 low income states have a high umber of
Continues for 6 more pages >>