HowSocioeconomic Factors Affect Graduation Rate in Michigan
Tableof Contents
ExecutiveSummary…………………………………………………………………….3
Introduction……………………………………………………………………………..3
Modelto Use……………………………………………………………………………..5
Multicollinearityand Economic Model………………………………………………….6
VarianceInflation Factors (VIF)…………………………………………………………7
Heteroscedasticityand Homoscedasticity………………………………………………..8
HypothesisTest……………………………………………………………………………9
Teacher……………………………………………………………………………………10
LocalRevenue…………………………………………………………………………….11
ViolentCrime……………………………………………………………………………11
MEAPMath and Science………………………………………………………………..12
Conclusion………………………………………………………………………………13
Appendix……………………………………………………………………………….. 15
References………………………………………………………………………………..27
ExecutiveSummary
Educationis vital for the productivity future of the workers in Michigan, sodetermining the factors which impact graduation is important to theimminent success of the economy within the state. The researchinvestigates the key socioeconomic aspects in the society affectingthe graduation rates of high school. The research will utilizevarious factors like crime rate, poverty level, class size, studentbody, local taxes on schools as well as standardized test scores. Theresults from multiple regression using the city data, school districtfrom the FBI and CCD propose that the level of poverty and crimerates are negatively impacting the graduation rates of high school.Reforms to the public policy and reducing the rate of crime and thelevels of poverty in the Michigan state could improve the rate ofgraduation of high schools and help minimize the chances of the MotorCity losing its important workforce in the auto industry and helpdiversify the future human capital of the Michigan.
Educationis an essential component to the growth and health of the economy, agauge of technological skills level, and productivity level of thefuture work. It is also a reflection of the societal values andnorms. Every parent is concerned, and the prospective business owneris also concerned because the young adults in the current schoolsystem are the future of the economy. Currently, education was at theforefront of the news as well as politics more so with theintroduction of the No Child Left Behind Act (NCLB) 15 years ago. Asthe concern for our youth’s education increases and public policyfocuses more on ways to improve the quality and outcome of thesecondary and primary education system, it is vital for us as anation, to find out the socioeconomic status factors which influencethe ability of the youth to graduate from the high school.
Understandingthe variables which affect high school graduation will enable us toact in a manner which is beneficial to the society, creating policieswhich will boost education and eliminate the reasons for thestudents’ dropout rate. Moreover, we will be able to evaluate thefactors in our neighborhoods and communities which matter most andfocus all the attention on the local level. Within Michigan, it isvital to examine Detroit Metropolitan area and all thesesocioeconomic factors so that we can easily find out the policies toenact which will not only boost the economic status but also the rateof graduation rate than the previous year in Michigan.
Moreover,it is imperative to note that the state is the best in the automotivesector, and boosting the graduation level will enable the state tohave enough technological skills necessary for the automotiveindustry in the state. It would create the norms and standards whichwill last beyond the current generation. Education is the future forthe whole community, and it will last beyond next generations tocome. Thus, without this perceived value of the education and itsrole in the state, the economy will flounder. Using data from fortydistricts and over seventy schools as shown in Appendix IV, the paperwill determine the socioeconomic factors which are associated withthe rate of graduation in high school within the Metropolitan area.
Focusingon the MetroDetroit area is essential for the wellbeing of theMichigan community. The study will help in developing the economichealth of the future of the economy and absent of the expectedexternal outcomes. Increasing the rate of graduation in the secondaryeducation will help in offering solutions to help to maximize thepotential for the human capital in the Detroit area and lessen thecost of losing the important working base and failing to meet theneeds of the hightech economy. The paper examines the graduationrates in the Metropolitan Detroit and its relationship between thesocioeconomic factors level like poverty degree in the district, thecrime rate of the entire city, teacher ratio and the MichiganEducational Assessment Program (MEAP) investigations technicalsubjects like math and science.
Modelto Use
Accordingto Sledge (2016), despite the commitment of the government to provideequity and opportunity for all the students’ family background, atall stages of development of learning the rate of graduation isdetermined by the socioeconomic status of a young persona. Thepaper will consider social and economic factors like poverty, andviolent crime as they have an adverse impact on the rate ofgraduation for school students. All these social and economic factorspresent a model as shown below:
Grad= β1 + β2Teachers + β3Caucasians – β4Divorce + β5Femalesβ6VCrime – β7Poverty + β8Tests + β9Revenues + Ui
Themodel will include factors which I believe that they have asignificant impact on the graduation rate of the students inMetropolitan are in Michigan. Given the time and the scope, it is notpossible to exhaust all the variables of what can be obtained withinthe area of study. All the schools considered for the study werepicked regarding central location around the Metropolitan area andmore so twentymile radius. A total of forty schools districts andseventy high schools were considered. All these schools are neithercharter nor magnet schools.
Thecore independent socioeconomic variables in this study are violentcrime rate and poverty. However, other factors are added to make thestudy more exhaustive. All these variables considered are continuousand quantitative using the city data and district data. Appendix Ipresent the descriptive statistics and sources for all the variablesin this study while Appendix gives overall data of this study.
Multicollinearityand Economic Model
Theequation below shows the economic model which predicts the teacher,MEAP, Female Local Rev, White, and Teacher impact the public rate ofgraduation. If this model is correct as per many theories, thenVcrime and Free Lunch affect the rate of graduation negatively in thepublic high schools in Michigan Metropolitan area for the school year2016 averagely.
Grad=b_{1}_{}+b_{2}Teacher+ b_{3}White+ b_{4}Female b_{5}FreeLunch+b_{6}LocalRevb_{7}VCrime+ b_{8}MEAP+e_{i}
Thestatistical analysis using the model is as shown below:
^{Y}_{i}_{}^{=}^{1}^{.}^{3}^{0}^{}^{+}^{}^{.}^{00}^{3}^{T}^{e}^{a}^{c}^{h}^{e}^{r}_{i}_{}^{−}^{}^{.}^{06}^{7}^{W}^{h}^{it}^{e}_{i}_{}^{−}^{}^{.}^{64}^{6}^{Fe}^{m}^{a}^{l}^{e}_{i}_{}^{−}^{}^{.}^{39}^{5}^{F}^{r}^{ee}^{L}^{un}^{c}^{h}_{i}_{}^{−}^{}^{3}^{.}^{943}^{E}^{}^{−}^{}^{6}^{L}^{o}^{c}^{a}^{l}^{r}^{ev}_{i}_{}^{−}^{}^{7}^{.}^{00}^{5}^{Vc}^{r}^{i}^{m}^{e}_{i}_{}^{+}^{.}^{08}^{7}^{M}^{EAP}^{m}^{a}^{t}^{h}^{s}^{ci}_{i}
TChr’ 
white 
female 
Free lunch 
Local rev 
Vcrime 
MEAP 

se = (0.322) 
(0.006) 
(0.049) 
(0.471) 
(0.082) 
(0.000) 
(2.545) 
(0.093) 
t = (4.043) 
(0.511) 
(1.374) 
(1.373) 
(4.843) 
(0.648) 
(2.753) 
(0.943) 
sig = (0.000) 
(0.613) 
(0.179) 
(0.179) 
(0.000) 
(0.521) 
(0.010) 
(0.353) 
d.f.= 32r2 = 0.785 Adjusted r2 = 0.738
Accordingto Uriel, when all the necessary conditions are applied, theestimators are BLUE, that is they are “Best, Linear, Unbiased,Estimator” (Uriel, 2013). Best, in this case, shows that theestimator has the lowest variance and it is the most reliableestimator. Unbiased estimators are imperative as they produceaveragely correct results. Estimators is a formula which is used toget the point estimate
However,autocorrelation, multicollinearity, and the heteroscedasticity do notfollow the assumption, which is needed for an estimator to beconsidered BLUE. Appendix II show no direct output of the ClassicalLimit Regression Model assumption violations. However, since thesample is large, we can assume that the disturbance term isdistributed generally based on the Central Limit Theorem. Thus wewill conduct VIF and the hypothesis test basing on the assumptionthat U i ~ N.
VarianceInflation Factors (VIF)
Lookingat the Variance Inflation Factors to ascertain if there is anymulticollinearity between the independent variables, it is evidencedthat there is no significant cause to conclude that the regressionsuffers from multicollinearity. It is because the largest VIF valueof the data is 3.458, that is, for Lunch Rate variable. The LunchRate variable is shown to be less than ten showing that once canconclude that there is some collinearity, but it is not significantstatistically:
COR(X i , X j ) = 0 ∀_{i≠ j}
Autocorrelationviolation is not a big deal in this data because the data iscrosssectional and not time series. But, to ensure an accurateanalysis of the model, it will become vital to test the DurbinWatsonStatistic of 2.007. The null hypothesis is that there is noautocorrelation while alternative hypothesis shows the presence ofautocorrelation.
H0:ρ = 0
H1:ρ ≠ 0
dl= 1.12 < du = 1.924 < 2.007 > 4du = 2.076 > 4dl = 2.88
Sincethe statistics at 2, indicated no autocorrelation, this test showsthat there is autocorrelation. We thus conclude that at 5% level,COR(U i ,U j ) = 0∀i≠ j
Heteroscedasticityand Homoscedasticity
Next,is that I checked if the error of the data is homoscedastic. Thepresence of heteroscedasticity results from such biased estimates ofthe standard error which makes the hypothesis test to be invalid. Iused each of the independent variables and graphed them against theresiduals squared. These charts look suspicious, which shows apossibility of heteroscedasticity in the model as shown in AppendixIII. Also, to determine the homoscedasticity, Park Tests for all thecontinuous independent variables were used. The test was significantto ascertain the statistical significance of the variance as itvaries from observation to observation. The result is also shown inAppendix III
Finally,there was no evidence that the collected the model was violating anyCLRM needed assumption for the estimates to archive BLUE model. Theresult in Appendix III showed that there was no statisticalsignificant collinearity, heteroscedasticity or autocorrelationbetween the independent variables and the disturbance terms. Themeasurement error in this study was limited, and there was no need torerun any further regression. It was vital to undertake Hypothesistesting for the significance level of the model measurements and theindependent variables.
HypothesisTest
AppendixII shows the regression output, and it explains 73.8% of the variancein the rate of graduation in high school. It is imperative to conductthe hypothesis test to ensure that the statistical significance andto ascertain of the regression shows 73.8% of the high schoolvariance. The hypotheses are:
H0: R 2= 0
H1:R 2 > 0
Theresult was that: Fstat = 16.7while the pvalue = 0.00 < 0.01< 0.05 < 0.10
Wewill reject the null hypothesis as a conclusion that there is 0%chance to be observed as an adjusted R2 of the 73.8% if the adjustedpopulation R2 is zero. The regression is statistically significant ata 1% level, and it explains 73.8% of the variance in the rate ofgraduation in high school.
Teacher
Examiningthe “Teacher” the slope is 0.003, which means that for everyadditional student in every teacher, there is 0.003 percentagedecrease in point of graduation rate on an average ceteris paribus. The hypothesis of the linear association between the teacher andgraduation rate is as shown
H0: β2 = 0
H1:β2 ≠ 0
Fromthe result, I got: tstat = 0.511 and thepvalue = 0.613 > 0.10> 0.05 > 0.01
Fromthis, we reject the null hypothesis. It shows that there is asubstantial evidence of the linear association between the percentageof the students who are eligible for a free lunch and the rate ofgraduation in high school at a 1% level of significance. To ascertainthis theory, I conducted a hypothesis test for a statisticallysignificant negative linear relationship between two given variables.
H0: β5 ≥ 0
H1:β5 < 0
Theresults were: tstat = 4.843 while the pvalue = 0.00/2 = 0.00 <0.01
Thus,I rejected the null hypothesis and concluded that there was evidenceof the strong negative association between the percentage of thestudents eligible for the free lunch and the graduation rate at a 1%significance level. It thus supports the theory.
LocalRevenue
Onthe other hand, examining the “Local Rev,’ one can see it’s theslope is 3.943E6, which shows that for every increase in the dollarin the local revenue per student, there is a negative 3.943E6percentage decrease in the rate of high school graduation averagelyceteris paribus. I conducted hypothesis testing of the slope of theLocal Revenue to ascertain its significance of the linear associationbetween it and graduation as follows:
H0: β6 = 0
H1:β6 ≠ 0
Theresults were that the tstat = 0.648 whereas the pvalue = 0.521 >0.10 > 0.05 > 0.01
Thus,we accept the null hypothesis, which there is no evidence of a linearrelationship between the amount of local revenue in real student andhigh graduation rate at a 10% significance level. However, due to thenegative slope of this variable, I suspect that the theory isincorrect and it will be vital to conduct hypothesis testing forstatistically significant negative linear association between twovariables as follows:
H0: β6 ≥ 0
H1:β6 < 0
Theresult is that, tstat = 0.648 while the pvalue = 0.521/2 = 0.2605> 0.10
Thus,we accept the null hypothesis and conclude that no evidence of theadverse relationship between the local revenue per student and rateof graduation in high school at 10% significance level.
ViolentCrime Rate
Theslope of the “VCrime Rate” is equal to 7.005. It shows that forevery percentage point increase in the violent crime rate in thecity, there is a 7.005 decrease in the high school rate of graduationaveragely. Hence, let’s conduct a hypothesis test of the ViolentCrime slope to ascertain statistical significance of the linearassociation between the Violent Crime and the graduation rate.
H0: β7 = 0
H1:β7 ≠ 0
Itresults to: tstat = 2.753 and pvalue = 0.01 = 0.01 < 0.05 <0.10
Theresults show that one should reject the null hypothesis as it showsthat there is a substantial evidence of a linear relationship betweenthe graduation rate of high school and violent crime rate in the cityat a 1% significance level. To test the model theory, it’sessential to conduct a hypothesis test for a statistical significancenegative linear relationship between the two variables.
H0: β7 ≥ 0
H1:β7 < 0
Itresult to a tstat equaling to 2.753 and the pvalue = 0.01/2 =0.005 < 0.01
Wewill reject the null hypothesis as the result shows that there isevidence of a strong negative association between the rate ofgraduation and rate of violent crime at a 1% significance level. Thus, it supports the theoretical model.
MEAPMath and Science
Theperformance of students in math and science classes is oftendetermined by their social status. Thus examining “MEAP Math Sci”with a slope of 0.087 shows that for every percentage point increasein meeting or surpassing the MEAP standards set for math as well asscience there is a 0.087% increase in the rate of high schoolgraduation averagely. Conducting hypothesis of the MEAP math andscience slope result will be important. It is as shown below:
H0: β8 = 0
H1:β8 ≠ 0
Theresult is: tstat = 0.943 while pvalue = 0.353 > 0.10 > 0.05 >0.01
Weaccept the null hypothesis and conclude that there is no evidence ofa linear association between the percent of students meeting orexceeding the MEAP standards for math and science at 10% significancelevel. To test the model theory, we will conduct a hypothesis test,for a statistically significant linear positive relationship betweenthe two independent variables.
H0: β8 ≤ 0
H1:β8 > 0
Thus,tstat = 0.943 and the pvalue = 0.353/2 = 0.1765 > 0.10
Theabove result shows that the null hypothesis should be accepted whilealternative hypothesis should be rejected. There is no evidence thata positive relationship between the percentage of the studentsmeetings or exceeding the standards of MEAP science including math inthe graduation of students in high school at a significance level of10%.
Conclusion
Thepaper has created a theoretical model which illustrates the differentrates of graduation in the high school in Michigan. The theory hasbeen estimated using SPSS as well as Ordinary Least Squaresregression. Although the economic model shows that 73.8% of the highschool graduation rate variance, most of the independent variableshave no significant effect on the high graduation rate in Michigan.However, the percentage of the female students and white students hasweak negative association with the mean graduation rate. Theseresults oppose the model used in this paper. However, there are twosocioeconomic factors which statistically bear a predicting sign.These are violent crime rate and the Free Lunch Rate. These twofactors are mainly proxy of the poverty in surrounding study area.School districts with more poverty and crime rate have lower rates ofgraduation. These results are consistent with the main hypothesisthat poverty and crime pose a serious threat to the attainment ofeducation by children in Michigan.
AppendixI
N 
Minimum 
Maximum 
Mean 
Std. Deviation 

Grad 
40 
.5484 
.9915 
.877603 
.1130667 
Teacher 
40 
16.7 
24.7 
20.298 
1.8955 
White 
40 
.0046 
.9517 
.664658 
.2932748 
Female 
40 
.4275 
.5245 
.489146 
.0222691 
Free_Lunch_Rate 
40 
.0245 
.8932 
.278839 
.2111611 
Local_Rev 
40 
989 
9820 
3621.65 
1868.224 
vcrime_rate 
40 
.0006 
.0242 
.005315 
.0052443 
MEAP_math_sci 
40 
.0400 
.7775 
.416725 
.1614550 
AppendixII

Model
R
R Square
Adjusted R Square
Std. Error of the
Estimate
DurbinWatson
1
.886^{a}
.785
.738
.0578661
2.007

Predictors: (Constant), MEAP_math_sci, Female, Teacher, Local Rev, vcrime rate, White, Free_Lunch_Rate

Dependent variable. Graduation

Model
Sum of Squares
df
Mean Square
F
Sig.
1 Regression Residual Total
.391
7
.056
16.700
.000^{a}
.107
32
.003
.499
39

Predictors: (Constant), MEAP_math_sci, Female, Teacher, Local_Rev, vcrime_rate, White, Free Lunch Rate

Dependent variable. Graduation
Model 
Unstandardized Coefficients 
Standardized Coefficients 
t 
Sig. 
Collinearity Statistics 

B 
Std. Error 
Beta 
Tolerance 
VIF 

1 (Constant) Teacher White Female Free_Lunch_Rate Local_Rev Vcrime_rate MEAP_math_sci 
1.300 
.322 
4.043 
.000 

.003 
.006 
.052 
.511 
.613 
.641 
1.559 

.067 
.049 
.174 
1.374 
.179 
.418 
2.391 

.646 
.471 
.127 
1.373 
.179 
.781 
1.280 

.395 
.082 
.738 
4.843 
.000 
.289 
3.458 

3.943E6 
.000 
.065 
.648 
.521 
.665 
1.504 

7.005 
2.545 
.325 
2.753 
.010 
.482 
2.074 

.087 
.093 
.125 
.943 
.353 
.384 
2.604 
AppendixIII: Park Tests
TeacherPark Test
Modelsummary

Model
R
R Square
Adjusted R Square
Std. Error of the
Estimate
1
.085^{a}
.007
.019
1.84856
a.Predictors: (Constant), lnTeacher

Model
Sum of Squares
df
Mean Square
F
Sig.
1 Regression
Residual
Total
.956
129.853
130.809
1
38
39
.956
3.417
.280
.600^{a}

Predictors: (Constant), lnTeacher

Dependent Variable: lnResSq
Coefficient

Model
Unstandardized Coefficients
Standardized
Coefficients
t
Sig.
B
Std. Error
Beta
1 (Constant)
lnTeacher
1.973
1.689
9.604
3.193
.085
.205
.529
.838
.600

Dependent Variable: lnResSq
FemalePark Test
ModelSummary

Model
R
R Square
Adjusted R Square
Std. Error of the
Estimate
1
.096^{a}
.009
.017
1.84682

Predictors: (Constant), lnFemale
Anova

Model
Sum of Squares
df
Mean Square
F
Sig.
1 Regression
Residual
Total
1.200
129.609
130.809
1
38
39
1.200
3.411
.352
.557^{a}

Predictors: (Constant), lnFemale

Dependent Variable: lnResSq
Coefficient

Model
Unstandardized Coefficients
Standardized
Coefficients
t
Sig.
B
Std. Error
Beta
1 (Constant)
lnFemale
4.353
3.765
4.555
6.347
.096
.956
.593
.345
.557

Dependent Variable: lnResSq
FreeLunch Rate Park Test
ModelSummary

Model
R
R Square
Adjusted R Square
Std. Error of the
Estimate
1
.137^{a}
.019
.007
1.83783

Predictors: (Constant), lnFree_Lunch_Rate
Anova

Model
Sum of Squares
df
Mean Square
F
Sig.
1 Regression
Residual
Total
2.459
128.350
130.809
1
38
39
2.459
3.378
.728
.399^{a}

Predictors: (Constant), lnFree_Lunch_Rate

Dependent Variable: lnResSq
Coefficient

Model
Unstandardized Coefficients
Standardized
Coefficients
t
Sig.
B
Std. Error
Beta
1 (Constant)
lnFree_Lunch_Rate
6.572
.304
.630
.357
.137
10.429
.853
.000
.399

Dependent Variable: lnResSq
LocalRevenue Park Test
ModelSummary

Model
R
R Square
Adjusted R Square
Std. Error of the
Estimate
1
_{.}_{1}_{4}_{8}^{a}
.022
.004
1.83502

Predictors: (Constant), lnLocal_Rev
ANOVA

Model
Sum of Squares
df
Mean Square
F
Sig.
1 Regression
Residual
Total
2.852
127.957
130.809
1
38
39
2.852
3.367
.847
.363^{a}

Predictors: (Constant), lnLocal_Rev

. Dependent Variable: lnResSq
Coefficient

Model
Unstandardized Coefficients
Standardized
Coefficients
t
Sig.
B
Std. Error
Beta
1 (Constant)
lnLocal_Rev
11.421
.541
4.759
.588
.148
2.400
.920
.021
.363

Dependent Variable: lnResSq
ViolentCrime Park Test

Model
R
R Square
Adjusted R Square
Std. Error of the
Estimate
1
.014^{a}
.000
.026
1.85517

Predictors: (Constant), lnvcrime_rate

Model
Sum of Squares
df
Mean Square
F
Sig.
1 Regression
Residual
Total
.025
130.783
130.809
1
38
39
.025
3.442
.007
.932^{a}

Predictors: (Constant), lnvcrime_rate

Dependent Variable: lnResSq
Coefficient

Model
Unstandardized Coefficients
Standardized
Coefficients
t
Sig.
B
Std. Error
Beta
1 (Constant)
lnvcrime_rate
6.870
.032
2.101
.374
.014
3.269
.086
.002
.932

Dependent Variable: lnResSq
MEAPScience & Math Park Test
ModelSummary

Model
R
R Square
Adjusted R Square
Std. Error of the
Estimate
1
.094^{a}
.009
.017
1.84720

Predictors: (Constant), lnMEAP
ANOVA

Model
Sum of Squares
df
Mean Square
F
Sig.
1 Regression
Residual
Total
1.148
129.661
130.809
1
38
39
1.148
3.412
.336
.565^{a}

Predictors: (Constant), lnMEAP

Dependent variable inResq
Coefficient

Model
Unstandardized Coefficients
Standardized
Coefficients
t
Sig.
B
Std. Error
Beta
1 (Constant)
lnMEAP
7.350
.304
.595
.525
.094
12.354
.580
.000
.565

Dependent Variable: lnResSq
AppendixIV: Data Sets
District 
City 
Grad 
Teacher 
White 
Female 
Free Lunch Rate 
Local Rev 
VCrime Rate 
MEAP Math Sci 
Residuals 
Detroit City School District 
Detroit 
0.6516 
22.8 
0.0176 
0.4982 
0.5515 
1970 
0.0242 
0.2115 
0.0200 
Grosse Pointe Public Schools 
Grosse Pointe 
0.9705 
17.6 
0.8910 
0.4938 
0.0292 
4017 
0.0013 
0.7775 
0.0376 
South Lake Schools 
St. Clair Shores 
0.9646 
20.8 
0.7389 
0.4832 
0.1636 
3718 
0.0024 
0.5195 
0.0120 
East Detroit Public Schools 
Eastpointe 
0.8663 
21.3 
0.6732 
0.4887 
0.2563 
1654 
0.0077 
0.3085 
0.0047 
Van Dyke Public Schools 
Warren 
0.7352 
20.6 
0.6980 
0.4790 
0.4930 
3213 
0.0062 
0.3175 
0.0495 
Hamtramck Public Schools 
Hamtramck 
0.7342 
20.9 
0.5275 
0.4275 
0.6044 
1009 
0.0135 
0.3435 
0.0123 
Lakeview Public Schools 
St. Clair Shores 
0.9445 
20.5 
0.9110 
0.5245 
0.0721 
2432 
0.0024 
0.5485 
0.0127 
Roseville Community Schools 
Roseville 
0.9429 
24.1 
0.7639 
0.4472 
0.2262 
2585 
0.0048 
0.3875 
0.0072 
Center Line Public Schools 
Center Line 
0.9489 
17.6 
0.8279 
0.4931 
0.2587 
4957 
0.0028 
0.4945 
0.0660 
Fitzgerald Public Schools 
Warren 
0.8756 
20.6 
0.6408 
0.4485 
0.4014 
4507 
0.0062 
0.2860 
0.0390 
Highland Park City Schools 
Detroit 
0.5484 
19.2 
0.0046 
0.4965 
0.7691 
989 
0.0242 
0.0400 
0.0168 
Warren Woods Public Schools 
Warren 
0.9904 
19.6 
0.8675 
0.5112 
0.1343 
4078 
0.0062 
0.2885 
0.1052 
Lake Shore Public Schools 
St. Clair Shores 
0.9649 
21.4 
0.8848 
0.5200 
0.1152 
2298 
0.0024 
0.4525 
0.0251 
Warren Consolidated Schools 
Warren 
0.9556 
20.2 
0.8631 
0.5042 
0.1966 
4495 
0.0062 
0.4650 
0.0746 
Hazel Park City School District 
Hazel Park 
0.7491 
20.6 
0.8482 
0.5060 
0.3148 
3334 
0.0057 
0.3260 
0.0824 
Fraser Public Schools 
Fraser 
0.9574 
18.1 
0.8884 
0.5091 
0.1335 
2817 
0.0023 
0.5205 
0.0235 
Clintondale Community Schools 
Clinton Township 
0.8970 
21.0 
0.5229 
0.5012 
0.2886 
2756 
0.0034 
0.2000 
0.0214 
Ferndale Public Schools 
Ferndale 
0.9127 
19.7 
0.4441 
0.5165 
0.4754 
2865 
0.0056 
0.4800 
0.1107 
Lamphere Public Schools 
Madison Heights 
0.8883 
20.0 
0.9141 
0.4725 
0.1611 
9820 
0.0023 
0.5710 
0.0390 
Madison Public Schools 
Madison Heights 
0.7778 
21.5 
0.7682 
0.4992 
0.3230 
3299 
0.0023 
0.2725 
0.0823 
School District of the City of Royal Oak 
Royal Oak 
0.9878 
17.3 
0.9078 
0.5005 
0.1081 
5914 
0.0019 
0.6280 
0.0423 
Dearborn City School District 
Dearborn 
0.9254 
19.8 
0.9283 
0.4678 
0.3076 
4861 
0.0049 
0.4450 
0.0639 
Oak Park City School District 
Oak Park 
0.9233 
20.4 
0.0475 
0.4993 
0.2240 
3553 
0.0051 
0.4490 
0.0157 
MT. Clemens Community School District 
Mount Clemens 
0.6429 
17.9 
0.4090 
0.4782 
0.8932 
4393 
0.0072 
0.3795 
0.0112 
MelvindaleNorth Allen Park Schools 
Melvindale 
0.9544 
21.4 
0.6942 
0.5000 
0.3047 
2571 
0.0040 
0.3405 
0.0856 
Utica Community Schools 
Total Utica Area 
0.9583 
19.7 
0.9328 
0.4885 
0.0661 
2896 
0.0021 
0.5935 
0.0250 
Clawson City School District 
Clawson 
0.9457 
18.6 
0.9140 
0.4675 
0.0964 
6022 
0.0006 
0.4010 
0.0184 
Berkley School District 
Berkley 
0.9600 
18.1 
0.8046 
0.5000 
0.0698 
2666 
0.0011 
0.5650 
0.0235 
Lincoln Park Public Schools 
Lincoln Park 
0.8539 
23.0 
0.8337 
0.4873 
0.2786 
3031 
0.0037 
0.2975 
0.0255 
Allen Park Public 
Allen Park 
0.9501 
22.6 
0.9142 
0.5166 
0.0926 
2139 
0.0016 
0.5365 
0.0163 
References
Sledge,C. M. (2016). Socioeconomicstatus and its relationship to educational resources.http://rdw.rowan.edu/cgi/viewcontent.cgi?article=2550&context=etd
Uriel,E. (2013). 3 Multiple linear regression: estimation andproperties. parameters, 1(2),3.http://www.uv.es/~uriel/3%20Multiple%20linear%20regression%20estimation%20and%20properties.pdf