Page 59 - Contributed Paper Session (CPS) - Volume 1

CPS877 Paula J.G. et al.
several data tables based on the STATIS and PCA (Principal Components
Analysis) methods with a focus on the various individual countries.
The tables related to quantitative data collected for the same countries
(34) and variables (15) in different years (7), and permitted to perform the
simultaneous analysis and exploration of the entire set of data tables. The
study individuals were the countries (Australia, Austria, Belgium, Canada,
Switzerland, Czech Republic, Germany, Denmark, Spain, Estonia, Finland,
France, United Kingdom, Greece, Hungary, Ireland, Iceland, Israel, Italy, Japan,
Korea, Luxembourg, Latvia, Mexico, Netherlands, Norway, New Zealand,
Poland, Portugal, Slovak Republic, Slovenia, Sweden, Turkey, and United
States) while the variables involved several of the indicators measured by the
initiative in accordance with the datasets of the “How’s Life” report of 2017.
The study variables are: Household net-adjusted disposable income (USD
at PPP, per capita, 2015); Employment rate (age 15 to 64, as % population with
same age); Average annual gross earnings per full-time employee (USD at
2016 PPP); Labor market insecurity (monetary loss from unemployment, share
previous earnings); Long-term unemployment rate (% labor force unemployed
more than one year); Rooms per person (average number); Household
expenditure on housing (% household gross adjusted disposable income);
Dwellings without basic sanitary facilities (% people w/o dedicated flushing
toilet); Employees working very long hours (% employees working more than
50h/week); Life expectancy at birth (years); Perceived health status (% adults
self-reporting above “good”); Upper secondary education attainment per
adults (% people 25-64); Social support (% people that can rely on friends or
relatives); Satisfaction with water quality (% people in the population); and
Feelings of safety when walking alone at night (% people).
The analysis produced at a global level permitted to obtain a view on the
general evolution and trends with regard to the conditions of life in the OECD
countries during the period from 2009 to 2015 (i.e., after the 2008 global
financial crisis). For this purpose, each of the years in the analysis period was
treated as an observation (center of gravity), and the study variables were the
selected indicators (15) of the “How’s Life” program. The statistical effect of
the outlier observations related to Mexico and Turkey (on four variables each),
Korea (on three variable), and Spain and Greece (on two variables each) was
attenuated due to the standardization of data given the different units of the
study variables.
In this context, the PCA conducted to eigenvalues (and associated
eigenvectors) for the correlation matrix indicating that the first two axes
largely explained the results given their combined variability (85.6% of the
total inertia). The representation on the first principal plan (Figure 1) indicated
that the first axis related to the evolution over time of the dimensions
associated with the quality and material conditions of life. In the period 2009

48 | I S I W S C 2 0 1 9