Methods
For this project, we used real data obtained from the Government of Canada website and published by Health Canada on boil water advisories (BWAs) in Canadian Indigenous communities from 1997-2017. The data indicated the region, Indigenous community, water system, infrastructure type, advisory type, and start and end date of each advisory. Using the start and end dates, we calculated the length of each advisory in terms of effect days.
We also looked at which federal government party was in power over time and how many advisories fell under each region of the country. The regions included all of the provinces with the Atlantic provinces grouped into one category. We used federal governing parties because the federal government is the one responsible for addressing issue in Indigenous communities such as boil water advisories, and not the provincial government. We did not have data for the Territories. Data was then manipulated using Excel and imported into R, a statistical computing and graphical software, to create our figures, and for statistical testing. |
|
Data
The data table we obtained from Health Canada consisted of information on the Region, District, First Nation, Indigenous community, Water system, Infrastructure type, advisory type, date set, and date revoked. The information we used from this for our project included: region, date set, date revoked. We added effect days (calculated from date set and revoked), federal government party and the year. Predictor variables included, government which is a categorical variable. We acknowledge that there would be other predictor variables not included in this project which would have affected number and length of advisories, but we limited to the one given the scope of this project, and the data we have available. Response variables included number of advisories, and effect days of advisories.
Table 1: Sample of the simplified dataset we used in this analysis. In this table we removed columns from the original data set which we did not use for simplicity (Indigenous community, advisory type, infrastructure type etc.). The last three columns were not part of the government dataset and were created for the purposes of this project. There were a total of 2484 total records between 1997 and 2017.