Education as a dimension of human development: A Provincial-level Education Index for Ecuador

the policies implemented in the countries of the region has forced surveys to be representative at increasingly higher levels of disaggregation [73]. Among these surveys, LSMSs are house- hold consumption and expenditure surveys which contain sections on a wide number of social topics, such as education, nutrition, health, or fertility (for a classification of household sur- veys, see e.g. [74]). Although LSMSs are best suited for inequality analysis [75], the fact that these surveys are more complex and costly than other household surveys means that LSMSs are only occasionally conducted in LAC countries. Initially, LSMSs formed part of the Living Standards Measurement Study, funded by the World Bank [76]. This study tool offered a holistic vision of well-being, focusing attention on its conditioning factors and including those derived from the implementation of government policies [77]. After 1996, LSMSs came to form part of the Program for the Improvement of Surveys and Measurement of Living Conditions in LAC, as a joint initiative of the Inter-Amer- ican Development Bank, the World Bank, and the Economic Commission for Latin America and the Caribbean (ECLAC) (for details of countries that participated in this program and their surveys, see, e.g. [75]). This program continues to support countries in conducting household surveys, focusing on improving the knowledge and application of household survey methodologies [47]. In Ecuador, the first round of the LSMS was conducted by the Servicio de Capacitación Pro-

fesional (Professional Training Service), whilst the following rounds were implemented by the Instituto Nacional de Estadística y Censos (National Department of Statistics and Census, or INEC). Since its first round in 1994, the LSMS was designed to gain an insight into the distri- bution of the population’s well-being and to glean information on the impact of the measures of macroeconomic adjustment and government programs aimed at reducing poverty. In fact, the general objectives of the Ecuadorian LSMS are aligned with the logic of Buen Vivir, provid- ing the necessary tools to implement policies that contribute to the elimination of population inequalities. Furthermore, due to its methodological design, the LSMS is a fundamental statis- tical instrument in the study of the living conditions of the Ecuadorian population [78]. As the aim was to also obtain an updated baseline to measure compliance with the United Nations Millennium Development Goals [79], the LSMS provide information concerning the different aspects and dimensions of well-being in Ecuadorian households, which constitute the units of information (the methodology used in this survey may be seen in [78]). The multidi- mensional concept of individual well-being offered by the Ecuadorian LSMS in line with the human development concept of the HDI and its intermediate indexes makes it the most suit- able household survey for the study of educational development in the Ecuadorian provinces. To measure Ecuadorians’ quality of life, the variables considered in the LSMS are classified

into sections. In particular, the section devoted to the educational dimension of human devel- opment aims to ascertain the level of literacy among the population, the level of non-atten- dance or household spending on education, among other aspects. The micro-databases of the fifth round (2005–2006) and the sixth round (2013–2014) of the Ecuadorian LSMS used in this paper contain information on 55,666 and 109,694 individuals, respectively [80, 81]. The two rounds of the LSMS adopt a similar methodology in terms of the structure of the questionnaire, the variables used, and the treatment given to the information by the executing entity, which endows the results obtained with a high degree of reliability. Both surveys offer national, regional, and provincial representativeness [78, 82]. Their method- ology is based on the expansion factor of each individual surveyed, defined as the inverse of the probability of choosing a household and its members in each province [78, 82]. This method allows the indicators derived from their use to be generalized to the population repre- sented in the survey [78] and avoids the problem of missing values in certain household sur- veys (see, for example, [83]). PLOS ONE Education as a dimension of Human development PLOS ONE | https://doi.org/10.1371/journal.pone.0270932 July 8, 2022 6 / 27