======================================================================= Livelihoods and 'Working for Wetlands' in Seekoeivlei, South Africa Data and R code for analysis A. Zabala. aiora.zabala@gmail.com Collected in 2007. Uploaded in 2018 ======================================================================= ======================================================================= ABSTRACT This cross-sectional dataset was collected to assess the local impacts of a wetland restoration project of the Working for Wetlands national programme for poverty alleviation and wetland rehabilitation in South Africa. This programme is managed by the South African National Biodiversity Institute. The data was collected to assess the impacts on livelihoods and on perceptions about the environment and development among workers within the programme for the project implemented in the Ramsar wetland Seekoeivlei, in Free State. Workers were from Zamani, a township of the town nearest to the wetland, Memel. A a proxy for the impacts of the project on its workers, data were collected from a sample of workers in the programme and a sample of people in the township unrelated to the programme (thereafter 'non-workers'). The analysis of these data has been published in an MSc thesis (1), a conference (2), and a peer-reviewed publication (3). ======================================================================= COLLECTION INFORMATION 22 workers were selected randomly from the total 46 employees in the project. Non-workers were sampled using multi-cluster random sampling, based on the aerial photograph of the township. For this latter sampling, the township was divided in blocks of similar size, and a number of houses were selected from within each block. The sample of non-workers maintains a similar gender and age balance to that of workers. The survey instrument collected demographic and socioeconomic data, as well as Likert-scale and open-ended responses about perceptions on the environment and development, both at a local level in a concrete manner (about local wetlands and community needs), as in a more abstract manner (about what environment and development mean for them). The data were collected with the help of a local research assistant who translated te responses from Zulu or Sesotho as necessary. ======================================================================= RELATED RESOURCES (1) Zabala, A., Sullivan, C.A. 2018. Multilevel assessment of a large-scale programme for poverty alleviation and wetland conservation: lessons from South Africa, Journal of Environmental Planning and Management, 61(3):493–514 https://doi.org/10.1080/09640568.2017.1319344 (2) Zabala, A. 2007. From macro to micro: environmental policy and 'the poorest of the poor' in Seekoeivlei, South Africa. MSc dissertation. Oxford: Oxford University Centre for the Environment (3) Zabala, A., Sullivan, C., 2013. Analysis of a national programme for wetland rehabilitation in South Africa: lessons for environment and development policies. 10th Biennial Conference of the European Society for Ecological Economics, Lille, France (4) Zabala, A., Sullivan, C., Accepted. Multilevel assessment of a large-scale programme for poverty alleviation and wetland conservation: lessons from South Africa. Journal of Environmental Policy and Management ======================================================================= FILE DESCRIPTION ======================================================================= DATA FILES ----------------------------------------------------------------------- wtlnd.csv Main dataset of livelihood variables. Described in wtlnd_vardesc.csv -----------------------------------------------------------------------wtlnd_vardesc.csv Describes each column in dataset wtlnd.csv -----------------------------------------------------------------------wtlnd_oe.csv Dataset derived from the open ended questions. For each open ended questions, the concepts mentioned were coded as shown in this dataset. -----------------------------------------------------------------------wtlnd_oe_vardesc.csv Describes each column in dataset wtlnd_oe.csv ----------------------------------------------------------------------- questionnaire.csv Additional questionnaire conducted at the end of 2007, asking solely the year of birth of each member of the household, to complement the data reported in wtlnd.csv. ======================================================================= CODE FILES IMPORTANT: in all the code files, the working directory has to be set to the source of the file, in command "setwd("...")" ----------------------------------------------------------------------- wfw_preparedb.R DATABASE PREPARATION (coded in 2013 by AZ) It prepares data from 3 CSV files exported from the original data spreadsheet & database: - Recodes factor variables: replace numbers by the actual category - Transforms variables to the correct class - Eliminates unnecessary variables - Computes new synthetising variables - Adds NA where necessary - Simplifies column names - Reports ----------------------------------------------------------------------- wfw_analysis.R ANALYSIS OF LIVELIHOOD DATA (coded in 2013 by AZ) Outputs of this code: - results$wd: table of descriptive statistics - ttests: table with t.test() results between 'status' and numerical variables - wilks: table with Mann-Whitney test results between 'status' and numerical variables - fishers: table with Fisher test results between 'status' and categorical variables - "wfw_categoricalVars_byGroup.pdf": plots of categorical variables by 'status' - "wfw_numericalVars_byGroup_kernelDensity.pdf": kernel plots of categorical variables by 'status' ----------------------------------------------------------------------- wfw_analysis_envdev.R ANALYSIS OF OPINION DATA (coded in 2013 by AZ) Outputs of this code: - 'results', a list of contingency tables for all OE variables, exported to OE_tables folder - 'fishers', a table of the p-values of Fisher's exact test to compare workers and non workers - devmelt & envmelt: variables M & N reshaped into long tables (one row for each concept mention) - envtable & devtable, written into envtable.csv & devtable.csv, contingency tables of N & M by work group and by gender - Plots - Barplots for env & dev - 3D mosaicplot for env & dev