Page 140 - IJPS-11-6

International Journal of
Population Studies Social inclusion for refugees

becoming refugees. Most of these refugees have fled to The analysis was conducted using two models designed
neighboring countries, creating significant challenges for to assess the impact of social support systems on refugee
social integration. It is essential to note that host countries integration. These models explored the relationships
are making efforts to provide necessary assistance and between various indicators of integration, such as access
support for the successful adaptation of refugees in their to employment, social services, and participation in
new environments. integration programs, based on the data from the cited
reports. Statistical methods, including K-means cluster
3. Data and methods analysis and dendrogram, were used to assess these

3.1. Data relationships. Detailed results of the analysis are presented
in the subsequent sections.
The selection of indicators to assess refugee integration
was based on secondary data sources, highlighting the Table 3 provides a description of the main indicators of
need for an in-depth analysis tailored to the context integration and social support for refugees, while Table 4
of the present migration crisis. Instead of collecting presents the statistical sources of information.
primary data through direct surveys and interviews, 3.2. Methods
this study relied on secondary data from official reports
and databases, employing an integrated approach using The proposed methodology consists of the following steps:
quantitative methods. The datasets, which included a i. Step 1: Identifying the key integration indicators using
representative sample of refugees segmented by factors hierarchical clustering (Joining Tree Clustering)
– such as age, gender, education level, and employment
status – were gathered from well-established sources such Hierarchical clustering using the Joining Tree method
as EUR-Lex, IOM, and the OECD. These sources provided was employed to identify the most significant indicators
detailed information on social support measures and influencing refugee integration. This approach organizes
integration processes for Ukrainian refugees in various the indicators into clusters based on their similarities,
European countries. This integrated approach ensures a allowing the determination of their relative importance
comprehensive understanding of the migration situation in shaping integration outcomes (Toronen, 2004). The
across multiple countries and demographic groups. initial set of indicators used to identify the most influential
factors for refugee integration is presented in Table 3.
Table 1. Number of Ukrainian refugees in various European Normalization of indicators was carried out using the
countries (as of January 2024) Z-score method to ensure comparability across different
scales (Glänzel et al., 2008):
Country Number of refugees
Poland 1,640,510 X − µ (I)
Russia 1,212,585 Z = σ
Germany 1,125,950
where X is the raw score, μ is the mean, and σ is the
The Czech Republic 547,670 standard deviation of the dataset.
Great Britain 210,800
The Ward hierarchical clustering algorithm was
Spain 186,045 employed to construct dendrograms, representing
Bulgaria 168,570 the grouping of indicators based on their similarities
Italy 163,570 (Murtagh and Legendre, 2014). This method minimizes
Moldova 116,615 within-cluster variance while maximizing between-cluster
Romania 106,786 variance. The formula for calculating distances is as follows:
Source: Ukrainian Refugee Crisis (2024). D = |X -X| 2 (II)
ij
j
i
Table 2. Main demographic characteristics of Ukrainian Where D is the distance between indicators X and X,
ij
i
refugees while |X-X| is the Euclidean distance (Silbergleit et al., 2015). j
i
j
Category Percentage of total refugees ii. Step 2: Classification of host countries using K-means
Women and children 90% Clustering
Men (18 – 60 years old) 10% Based on the significant indicators identified in the
Returnees 4 – 5 million people dendrogram, the K-means clustering algorithm – an
Source: Ukrainian Refugee Crisis (2024). unsupervised machine learning method – was employed

Volume 11 Issue 6 (2025) 134 https://doi.org/10.36922/ijps.4502