Premise performs ongoing sentiment analysis utilizing our strong user base and their contributions to our Global Index. The Premise Global Index strategically comprises over 35 general topics: from strength and effectiveness of civic organizations, community cohesion and quality of life, economic health, to the availability/effectiveness of basic services (health care, water & sanitation, transportation, housing, etc.) and many more.
Income Levels Matter
In a recent analysis, we looked at sentiment surrounding whether or not the government provides adequate food and nutrition for citizens who need it. We compared two opposite groups and their responses. Group one are those in our lowest income bracket (“I can’t afford enough food for my family”) and group two are those in our highest income bracket (“I can comfortably afford food, clothes, furniture, and I have savings”). Presumably, those who cannot afford enough food are most likely to rely on government assistance for food provisions. Looking at these two opposite income levels will allow us to compare a wider public perception of government assistance versus its implementation.
Countries with a large positive difference indicate a significant disparity between income levels. High-income contributors feel that the government provides enough food and nutrition for citizens who need it and low-income contributors feel like their government does not. This type of analysis gives insight into both the effectiveness of food assistance programs and the degree to which the wealthy (and therefore presumably more politically engaged) are in touch with the realities of food assistance programs in their country.
Our data shows that the countries with the greatest disparities in responses between high and low-income groups were Mozambique, Cambodia, Bangladesh, and Somalia. These countries had an over 20% difference in responses, with more low-income users agreeing that their government provided enough nutrition and food,
while more high-income users disagreed.
In countries where more low-income users felt a lack of assistance (with the largest disparities in Ireland, Germany, Canada, and Lithuania), the gap between high and low-income respondents was much smaller. The plot above displays the countries with the most significant disparities in each arrangement: top disparities for more high-income users agreeing and top disparities for more low-income users agreeing.
To look into this trend, we explored some measurements that could give insight into each country’s economic situation. These included the median household income, percent of GDP spent on welfare programs, the Gini coefficient, and GDP per capita. Initially, we expected countries where more low-income users felt they had enough nutrition to be spending a higher percentage of their GDP on social welfare programs (such as food assistance). But, there wasn’t a clear trend here.
The Gini coefficient measures the degree of income inequality in a given group. Looking at the country wide Gini coefficients for these examples, we expected a larger coefficient for countries with a wider gap between high and low-income user responses. However, there was not a clear relationship between the two. On average, the lower bracket of countries had a higher Gini coefficient, but this was because Mozambique has such a high coefficient. Excluding this outlier, the countries in this bracket had similar coefficients to those in the upper bracket.
Then we moved on to less detailed aspects such as median household income and GDP per capita. While the trend could have existed for both, data for median household income in Mozambique and Somalia was unavailable. Thus, we were left with GDP per capita.
Whether more low-income people agreed or more high-income people agreed with what did correlate with the overall GDP per capita. In the lower-income countries, the low-income users felt that the government provided enough food/nutrition more than the high-income users did. This observation is reversed in higher-income countries.
Interestingly, there was no significant difference between genders regarding how they perceived the food and nutrition assistance programs. While slightly more men strongly agreed that the government provides adequate food/nutrition, the rest of the response categories had little to no difference between genders. I hypothesized that, since women are usually tasked with preparing food for the family, they may have a different opinion on whether or not there is sufficient food and nutrition. This was not the case, however, as we saw similar rates of agreement across genders.
Filling in the Gaps
Overall, it is evident that countries with the highest disparities between high and low-income user responses tended to be lower-income countries. Our lowest-income users felt that the government provided enough nutrition more than the higher-income users in these countries. In wealthier countries, the gap between agreement extents shrunk. This was in addition to having the income level opinions swap, with lower-income users feeling that they do not receive enough nutrition from the government.
While there were no immediate answers behind some of the disparities highlighted here, the power of Premise’s data is its ability to help users surface initial insights like these. From there, we can generate targeted, custom collections to dive deeper into areas of interest. What else is behind contributor’s thoughts regarding the effectiveness of these food assistance programs? How do users/families supplement their diet if they aren’t receiving enough from the government? Are there other non-governmental food assistance programs available? What makes those options more appealing? Premise helps customers monitor the broader threads and then generates more refined collection questions to narrow the root causes. With this detailed and up to date information, they can take corrective action.
Premise has the potential to fill in these gaps in understanding for clients from a variety of backgrounds. With targeted questionnaires and detailed, up-to-date demographic information, there is plenty of room to explore the potential causes of such insights from the data. For more information, email us at firstname.lastname@example.org.