Looking for a job is influenced by level of economic development and by new information technologies adoption. In this paper, an exploratory and cluster analysis, and a multiple regression analysis was applied when studying variable Y- Percentage of individuals using the Internet for looking for a job or sending a job application (individuals aged 16 to 74; average for 2004-2008) depending on six independent variables which are: X1-Level of Internet access in households (Percentage of households who have Internet access at home), X2- GDP per capita in PPS (GDP per capita in Purchasing Power Standards (PPS) (EU-27 = 100); X3- Total unemployment rate; X4- Public expenditure on education (Percent of GDP); X5- Individuals' level of computer skills (Individuals who have carried out 1 or 2 of the computer related. Percentage of the total number of individuals aged 16 to 74; X6 - A dummy variable for a country being in transition or not.

Data EU-27, Island, Turkey and Croatia were investigated.

Using multivariate analysis of data based on cluster analysis with n=30 countries and k=6 variables, Ward linkage and Euclidean distance, four clusters were created.

Since the multiple linear regression Model I with all six proposed regressors influencing the dependent variable Y showed that only variable X1 is statistically highly significant, but with multicollinearity problem, a smaller dimension final Model II with only K=2 regressors X1 and X2 was developed.

The F-test with p-value=5,75E-07 shows that the overall regression is statistically significant at 1% significance level. Coefficient of determination R2 shows that 65,5% of the total sum of squares is explained by the Model II explaining the variability in the dependent variable Y- Percentage of individuals using the Internet for looking for a job or sending a job application (percentage of individuals aged 16 to 74; average for period 2004-2008). Further, the regression diagnostics was conducted using the t test for testing the significance of each of the independent variables. The variable X1-Level of Internet access in households (Percentage of households who have Internet access at home) is highly statistically significant, and X2- GDP per capita in PPS (GDP per capita in Purchasing Power Standards (PPS) (EU-27 = 100) is significant at 7% significance level.

References:

Dougherty, C.(2006). Introduction to econometrics, Oxford University Press, Oxford

Kuhn, P. and Skuterud, M. (2000) Job search methods: Internet versus traditional, Monthly labour review, Vol. 123, No. 10, pp. 3-11.

Suvankulov, F. (2010). Job Search on the Internet, E-Recruitment, and Labor Market Outcomes, 2010-01-02, http://www.rand.org/pubs/rgs_dissertations/RGSD271.html.

Fountain, C. (2003). Finding a Job in the Information Age: Job Searching, Labor Market Outcomes, and the Internet, Paper presented at the annual meeting of the American Sociological Association, Atlanta Hilton Hotel, Atlanta, 2010-01-02, http://www.allacademic.com/meta/p106817_index.html.

Keywords: Usage of Internet when looking for a job; Multiple regression; Regression diagnostics; Cluster analysis

Biography: Dr Ksenija Dumicic is a Professor at Department of Statistics, Faculty of Economics and Business, University of Zagreb, Croatia. In 2007 she founded the postgraduate studies Statistical Methods for Economic Analysis and Forecasting, which is, by now, unique in the SEE region. She is the member of editor boards in several journals in Croatia, Serbia and Bosnia and Herzegovina and the reviewer of 8 textbooks in statistics in several countries. She is the member of Editorial Board of International Encyclopaedia of Statistical Sciences, published by Springer in 2010. She has authored and co-authored more than 80 papers and 2 books in statistical research methodology, specialized in statistical sample survey methods and statistical quality control with special interest in statistics education. She has supervised more than 10 postgraduate students. Member of IASS, and ASQ Statistical Division.