Data Table
Our studies sampling units consist of the number of wildlife individuals that are detected from the trail cameras in response to our four different deterrence types. The predictor variables are categorical and consist of our deterrents. Our response variable is continuous and includes the before and after measurements of wildlife activity. In our study, the predictor variable was unable to be directly manipulated as we are dealing with the wildlife in their natural setting. Therefore, we simply just observed our response variable as we cannot control the reaction of passing wildlife to any of the deterrents. There is a simplified version of 16 rows out of our total of 128 rows of data that was collected (Table 1). The ID column represents the order in which each sample was taken in accordance to our experimental design (there are a total of 128 ID’s). Our site column included rows containing both roads (R1) and railways (R1). The replication column pertains to the eight different months in which we collected both our before and after data from in 2018 and 2019. Therefore the first replication would be from the month of March, the second from April and so on until the eighth replication being October. The before and after column just defines whether or not the row of data was collected before, or after the deterrence mechanisms were enabled. Then lastly, the final column represents the monthly measurements of wildlife activity that was recorded in relation to each type of deterrent from either R1 or R2.
Table 1. Simplified project data table depicting recorded wildlife activity (WA) measurements from the before (B; 2018) and after (A; 2019) measurements for roads (R1) from the audio deterrent (A). Also includes the eight different replications which coincide to the eight months (March - October) in which recording was conducted.
Diagnostic Plots of Raw Data
There were two graphs generated from the raw data that was collected. The first graph (Figure 1) compares the cumulative before and after wildlife activity measurements across 28 days in relation to the four deterrent types. The purpose of this graph is to initially conceptualize the raw difference between the impacts of all four deterrents to see which deterrence type was most effective. As you can see in Figure 1, the before measurements had higher wildlife activity than the after measurements with the most significant drop being seen between the before and after from the combination deterrent that had both the audio and visual.
Figure 1. Cumulative 28 day wildlife activity based on before and after measurements
The second graph generated from our raw data (Figure 2), compares the cumulative differences in the before and after wildlife activity between the three different road locations and three different railway locations. The intent of creating this graph from the raw data was to gain an initial insight into which location was more effective in deterring wildlife in terms of the four different deterrent types. Overall, you can see that there is a higher amount of wildlife activity occurring at the railway as opposed to the road. One thing to note is the six different outliers that was recorded to have occurred at the road, potentially skewing the data.
Figure 2. Cumulative 28 day wildlife activity based on roads versus railways.