Analiza panel podataka

master's thesis

• Author: Matea Pavić
• Mentor: Anamarija Jazbec (mentor)
• Committee member: Anamarija Jazbec (predsjednik povjerenstva)
• Committee member: Siniša Slijepčević (član povjerenstva)
• Committee member: Snježana Lubura Strunjak (član povjerenstva)
• Committee member: Vjekoslav Kovač (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2019-07-15, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

U ovom radu izlažu se osnovne metode i teorijska podloga analize panel podataka kao vrlo važnog aspekta brojnih istraživanja u prirodnim i društveno-humanističkim znanostima. Njima se obuhvaćaju i analiziraju opažanja prikupljena tijekom različitih vremenskih točaka, to jest promatra se promjena varijabli i njihov utjecaj kroz vrijeme. Na početku rad donosi prikaz linearne regresije koja čini osnovu statističkog modeliranja, a kojom se u vezu dovodi zavisna varijabla i funkcija nezavisnih varijabli. U sljedećem poglavlju dan je pregled vrsta podataka te se objašnjava specifična struktura panel podataka koja se svodi na objedinjavanje vremenskog presjeka i vremenskog niza. Nakon toga objašnjavaju se linearni modeli panel podataka: združeni model, model fiksnih efekata i model slučajnih efekata; te testovi za odabir određenog modela: \(F-\) test, Breusch-Paganov test i Hausmanov test. U nastavku rada izloženi teorijski koncept panel podataka nadograđuje se primjerom analize podataka Statističkog ureda Europskih zajednica (Eurostata) u računalnom programu za statističku analizu SAS. Analiza se temelji na broju iseljenih osoba kao zavisnoj varijabli te podacima o mjesečnoj plaći, nezaposlenosti, stupnju obrazovanja i broju rastavljenih kao nezavisnim varijablama. Osim Hrvatske, odabrane su još četiri europske zemlje kao jedinice promatranja u vremenskom razdoblju od 2008. do 2016. godine.

Abstract (english)

This paper presents basic methods and theoretical background of panel data analysis as a very important aspect of numerous research topics in natural and social sciences. Panel data contain and analyze the observations of phenomena obtained over multiple time periods and explain influencing variables. First, the paper presents the linear regression that forms the basis of statistical modeling, which links the dependent variable to the function of independent variables. The next chapter provides an overview of data type and explains specific structure of panel data as combination of cross-section and time series data. Afterwards, text explains various linear models of the panel data: pooled OLS model, fixed effects model and random effects model, as well as tests to evaluate how statistical model corresponds to the data: \(F-\) test, the Breusch-Pagan test and Hausman specification test. The theoretical concept of the panel data analysis presented in this paper is outlined through data aggregated by the European Statistical Office (Eurostat) and computed in the statistical analysis software SAS. Presented panel analysis is based on annual collections of statistics on international migration flows as dependent variable, and statistics on monthly wages, unemployment, educational attainment level and divorce rate as independent variables. Besides Croatia, four other European countries have been selected as observation units over the period of 2008 to 2016.

• Keywords: analiza panel podataka; linearna regresija; statističko modeliranje; linearni modeli panel podataka; SAS
• Keywords (english): panel data analysis; linear regression; statistical modelling; linear models of panel data; SAS
• Language: croatian
• URN:NBN: urn:nbn:hr:217:399336
• Study programme: Title: Mathematical Statistics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra matematike (magistar/magistra matematike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2019-11-29 12:43:46