An Analusis of Road Conditions to Students Performance by Spatial Econometrics Techniques

<div>A school is an integral part of human&rsquo;s life. One of the ways to compare&nbsp;</div><div>different schools is a rating. For every student who pass exam well enough some&nbsp;</div><div>points are added on. Points are calculated, ranked and chosen the best schools.&nbsp;</div><div>This paper investigates different factors which influence a rating position&nbsp;</div><div>and a probability to join it. The aim of the research is to evaluate an influence of&nbsp;</div><div>school area on school&rsquo;s rating position. The following steps are made in order to&nbsp;</div><div>achieve that goal:&nbsp;</div><div>1) to find some area factors;&nbsp;</div><div>2) to collect the data for empirical research;&nbsp;</div><div>3) to estimate a probability to join the rating;&nbsp;</div><div>4) to estimate a rating position using different school and school&rsquo;s areas&nbsp;</div><div>characteristics;&nbsp;</div><div>5) to estimate a rating position using different characteristics of space&nbsp;</div><div>around a school.&nbsp;</div><div>Database has been collected with help of Yandex.Maps service, sites of&nbsp;</div><div>Moscow schools and the rating of 300 best schools. All the necessary data is&nbsp;</div><div>gathered automatically, using AutoIt software. GeoDA and R software are used to&nbsp;</div><div>analyze the data.&nbsp;</div><div>During the analysis following type of regression is used:&nbsp;</div><div> logistic regression;&nbsp;</div><div> OLS model;&nbsp;</div><div> spatial regression with lagged variable.&nbsp;</div><div><p>The findings indicate that:</p><p style="margin-left:53.45pt;">1)&nbsp;&nbsp;&nbsp;&nbsp; characteristics of school and school area influence the probability to join the rating;</p><p style="margin-left:53.45pt;">2)&nbsp;&nbsp;&nbsp;&nbsp; for reduced samples: only quantity of 11 grade students and students quantity at all affect the rating position;</p><p style="margin-left:53.45pt;">3)&nbsp;&nbsp;&nbsp;&nbsp; for whole sample: there is no influence of school area, but more variables are significant than in case of reduced sample;</p><p style="margin-left:53.45pt;">4)&nbsp;&nbsp;&nbsp;&nbsp; Moran&rsquo;s test shows that there is spatial dependence in data;</p><p style="margin-left:53.45pt;">5)&nbsp;&nbsp;&nbsp;&nbsp; if spatial regression is used both characteristics of school and its area are significant;</p><p style="margin-left:53.45pt;">6)&nbsp;&nbsp;&nbsp;&nbsp; it is impossible to say, there is a positive or negative effect of neighbors, but it definitely exists.</p></div><p>&nbsp;</p>