Interpreting the Results
How can we explain these results? In particular, how can we make sense of the failure of direct volunteering to narrow the gaps between the relative scale of volunteer work measured along the same dimensions in better-off and worse-off countries? At least three lines of possible explanation can be examined.
Limitations of the Data
An appealing first line of explanation could be to blame the messenger—in this case the data used to generate the estimates. To be sure, there is blame to go around here, for neither the CNP data nor the TUS data are perfect, which is why the present authors have been making such vigorous efforts to improve the systems and methodologies for measuring volunteer work comparatively on a global level.
In the case of the CNP data, many were collected during an earlier era with limited resources, which typically meant relatively small samples that could be accessed only by purchasing time on existing omnibus survey platforms. Inevitably, it was also necessary to be parsimonious in the identification of questions to include.
In the case of time use surveys, other problems arise. For one thing, unlike eating, sleeping, or working, a fairly small proportion of respondents are likely to engage in volunteering of any type in any particular week—probably less than 10 % and sometimes as low as 1 %. As a result, a country’s score may be unduly determined by the non-random behavior of a very few individuals, creating potentially large deviations in a country’s reported volunteer rate from year to year that are at base statistical artifacts.
A second problem arises from the use of the “household” as the unit of observation in time use surveys. A household is a group of people living together in one place of residence. Statisticians use this as a convention due to the ambiguity of other concepts such as “family.” But this can have ramifications for the measurement of volunteer work since such work is defined as help without pay for someone living outside of one’s household. In other words, the TUS uses a definition of volunteer work different from the ILO Manual. In countries where extended families do not live in the same household, help provided to family members living outside one’s own household would count as direct volunteering . But in countries where extended family members live in the same household, that same help to such a family member would not count as direct volunteering. This could thus artificially lower the relative amount of direct volunteering reported in the countries with extended family members living in the volunteer’s household.
Finally, concepts used in self-reported diaries are subject to interpretation by respondents and do not always correspond with official definitions. For example, while helping neighbors is considered a form of direct volunteering, respondents may instead report it as other types of household activities, such as cleaning, preparing meals, or socializing, depending on the nature of the task.
While these limitations are important and need to be addressed in future data gathering, we do not believe they challenge the basic conclusions reached here. In the case of the CNP data, great care was taken to create a carefully structured common survey protocol for use in all sites that avoided unclear buzz words. Steps were also taken to include both direct and organization-based volunteering and to record the specific auspices of the latter. Finally, at the end of the day, these data have stood the test of time and remain the only widely accepted systematically comparative data on the scope of both types of volunteer work in some 43 countries widely dispersed around the world.
In the case of the time use data, despite its limitations—of which the household unit of analysis may be the most serious—this remains the most accurate method available for recording the time individuals spend on various daily activities, including volunteering. It is far superior to ordinary opinion surveys because it collects information about direct as well as organization-based volunteering (something most other surveys do not do) and reconciles the results with the 24-h framework. What is more, the use of written diaries that record activities as they happen significantly reduces errors associated with recall, which are inherent in general surveys; and TUS results are carefully weighted to the population, which greatly simplifies making national estimates. Finally, if low-income countries with large numbers of extended families living in the households of TUS respondents were devoting anywhere close to enough time to direct volunteering for extended family members living in their households to alter the results reported here substantially, this would logically show up in higher hours doing “work for own household” in the time use surveys for these countries. In fact however, the time reported on “work for own household” for the countries in the South where such family patterns could be expected turn out to reveal fewer house devoted to work for own household than is the case for countries where extended families within the same household are far less common.
In short, while the data used here are certainly not ideal, problems with the data may still not be sufficient to explain the results that have emerged. How, then, can they be explained?