Through the Eyes of the Beholder

Unpacking intentions to migrate externally through these analyses means different things for different people, and we believe no story should be left behind. Thus, using both of the prior methods (and encouraging others to use even more) allows us to advocate for a wide range of actions, both temporarily and disciplinarily. We also believe that, when working with data, it is important not to lose sight that it represents real people and their stories. Our datasets are closely linked to such stories.

Note: To understand the single linear regression (SLR) results and their relationship to the stories, we highlighted the factors that showed statistically significant assoaciations with the model. The factors discussed in the story are filled in and the ones that were significant but are not shown in this story are highlighted but not filled in.

WHAT WE FOUND

The diagrams show varied results. In the simple linear regressions,the strongest associations with an intention to migrate externally include having received governmental aid, lacking access to gas or electricity for cooking, lacking housing tenure, living in communities where violence increased in the last year, having a decreased income due to the COVID-19 pandemic considering that most children in the country do not have the opportunity to learn and grow daily, having confidence in the country’s election, and considering that migration brings good consequences to families. In other words, statistically significant associations span across categories, disciplines and temporalities.


The multiple linear regression highlights new variables and downplays others. In this model, an intention to migrate externally correlated most strongly with being young, male, unhappy, and believing that migration is good. Other new associations include coming from a larger household, having the need to reduce food portions in the past 7 days to feed other household members, having insufficient income to cover food and expenses in the last 30 days (economic security) and/or living in a female owned household. In this last model, the R2 value equals 0.70, which means that the model predicts 70% of the variability in the outcome.


Our main diagrams show the strength of the prior associations. A bar with a value between 1 and 3 shows a statistically significant association, with 3 having the strongest statistical significance (1: p < 0.05; 2: p < 0.01; 3: p < 0.001). The color of each bar indicates the group to which the variable belongs (food security, economic security, housing, disasters, individual perceptions, demographics, geographic location, and infrastructure).


To further our analysis, we also unpacked the overarching simple linear regressions to identify significant correlations with intention to migrate specific to Honduras, Guatemala, and El Salvador. The simple linear regressions highlighted between countries, such as a strong relationship with youth in Honduras and a strong relationship with non-governmental assistance programs in El Salvador.