This paper investigates the relationship between the unemployment
rates of College graduates and High School graduates. From this
investigation, it appears that the relationship is moderately weak.
Many people pursue a degree to escape the inevitability of
It is obvious that many people feel a college education is
important, and more employment opportunities will arise if one has a
degree. On the other hand, can someone be just as successful with only a
High School degree? Is there an association between the Unemployment Rates Essay
of College and High School graduates? This spring quarter I have become
knowledgeable as to how to use the Storm software. Therefore, I am able to
compare data obtained to determine if a certain relationship exists between
the two variables. As a result of using this information, I was able to
accurately state if there was any kind of relationship between the
unemployment rates of College and High School graduates.
III. Discussion of Variables
It might be thought that the unemployment rates of College graduates
and High School graduates are related in that when the unemployment rates
of High School graduates increases, the unemployment rate of College
graduates might be expected to decline or remain steady.
The reason for
being is because it is assumed that having a college degree means greater
To test this theory, 40 data elements are acquired. Randomness is
sought by selecting the data on the last day of the month for 40
consecutive months starting with January 2001, and ending with April 2004.
This time period includes unemployment rates that are not seasonally
adjusted. The data on the unemployment rates of both College and High
School graduates was found in the U.S.
Department of Labor – Bureau of
IV. Discussion of the Results
The sample is described using a linear regression model. The result is
expressed by the formula: High School (Y) = 2.14 + 1.04 College (X).
R-squared at 0.40 suggests that the relationship is moderately weak due to
the fact that R-squared represents a stronger relationship the closer the
number is to 1.
A study of the residual graphs indicates that the relationship is
poor due to curvilinearity for unemployment rates of College graduates and
poor due to violation of both homoscedasticity and linearity assumption for
the unemployment rates of High School graduates. This impacts on the
results by saying that the graphs show that the model does not describe the
Taken as a whole, this model seems to need more refinement being that
the R-squared is actually fairly moderate at 0.
40. This model might be of
little use in predicting future movements of high school (Y) when college
(X) moves. Particularly interesting is how the unemployment rates for both
College and High School graduates have increased during the years, and that
one if not influenced by the other significantly.
When trying to describe a universe such as the relationship between
unemployment rates of high school graduates versus college graduates, one
might take a random sample and expect that the sample adequately represents
the universe. The sample in this study is the unemployment rates for 40
consecutive months of those with simply a High School diploma versus those
who possess a College degree (Bachelor’s Degree or Higher).
Next, measures are taken of the sample, and a model estimated. If the
model is a good estimator of the sample, it is to be expected that the
model is a good estimator of the universe. In this study, the model is not
a good estimator of the sample, and therefore it is not expected to be a
good estimator of the universe.
The model used in this paper is the linear regression model, which attempts
to model the relationship between two variables by fitting a linear
equation to observed data. One variable is considered to be an explanatory
variable, and the other is considered to be a dependent variable (Poole &
O’Farrell 1). There are several research objectives for which the
regression model may be used, but they may be classified into three groups:
(I) the computation of point estimates, (II) the derivation of interval
estimates, and (III) the testing of hypotheses (Poole & O’Farrell 2).
Care has to be taken to observe the assumptions of the model, which are:
1. The mean of the probability distribution of the random error is 0,
E(e) = 0. that is, the average of the errors over an infinitely long
series of experiments is 0 for each setting of the independent .