Multiple Linear Regression - Effectiveness of Door-to-Door Bio-Waste Collection in Driving Sorting of Dry Recyclables
Description
We hypothesise that implementing door-to-door bio-waste collection system is positively associated with increasing the amounts of dry recyclables collected, and that this effect is more significant compared to alternative waste collection systems. Our research shows that this hypothesis holds when running Multiple Linear Regression (MLR) analysis using R programming language, when controlling for six waste management system related variables (PAYT system, glass bring points, metal bring points, door-to-door paper collection, door-to-door plastic collection and number of other collection systems) as well as seven additional socio-economic and political factors (such as population, population density, ratio of well-informed citizens, governing party’s position on the environment, trust in local government, material and social deprivation ratio and GDP per capita). The dataset is assembled of data from various sources which are referenced in the excel file with extended data. The R code and R data file used to construct MLR models in Table 1 and Table 2 are also provided.
Files
Steps to reproduce
A dataset for an innovative approach using Multiple Linear Regression models requires various types of data from multiple sources. In addition to waste statistics, we collected data on appropriate socio-economic and political factors. The resulting complex dataset was used to build Multiple Linear Regression models and establish relationships that are discussed in the methods section. It is important to note, that our research design is shaped by (a lack of) data availability. Put differently, we choose a research design that makes the most of the little data available. A dataset for an innovative approach using Multiple Linear Regression models requires various types of data from multiple sources. In addition to waste statistics, we collected data on appropriate socio-economic and political factors. The resulting complex dataset was used to build Multiple Linear Regression models and establish relationships that are discussed in the methods section. It is important to note, that our research design is shaped by (a lack of) data availability. Put differently, we choose a research design that makes the most of the little data available. The most recent reliable city level data with the necessary level of detail on waste collection systems and captured waste fractions was from a 2015 study on separate collection in EU capitals commissioned by the European Commission (BiPRO/CRI, 2015). The amount of collected dry recyclables was derived from data in this report, which comprised of six main categories of waste. Five of them are – bio-waste and four types of dry recyclables – glass, metals, paper and plastics. In the absence of more recent data sources and the above discussed problem with data availability, this waste dataset was and remains the best – and only – option to date. Other control variables were found based on the relevance to our research and were obtained from EUROSTAT and other sources, which are referenced in the excel file “Extended-data-manuscript-2022” under tabs called “Waste Data + Controls T1” and “Controls T2”.