Embedded Files
Conclusions
Results
I concluded my data collection period at the end of February, and immediately began data analysis. Data analysis was slow going, as I had to first tally up the total length of stay for over one thousand dogs.
Example of Data Collection and Color Coding
After compiling all the data, I started sorting each dog into its respective categories of sex, breed, coat color, and age. I color coded each characteristic to make it easier to visualize the data. I then found the average length of stay for each shelter, and for each phenotypic characteristic within the shelters.
I then normalized all the data to make sure that I could compare data across the shelters in spite of their varying average length of stays.
After finding these numbers, I then ran them through a repeated measures ANOVA-test comparing each phenotypic characteristic to the location in order to determine if there was any statistically significant difference between locations.
Example of Averaged and Normalized Data
Example of ANOVA-test Results: p-value under 0.05 for columns meant location had a significant impact on LOS for that characteristic
The rows of my table were the different characteristics and the columns where the different locations. If the p-values were below 0.05 for the different locations then the difference in length of stays of the dogs was statistically significant, and I moved on to the next part of my data analysis.
In order to maintain the anonymity of the shelters, I decided to replaced their names with the letters A-J for data analysis in order to keep their anonymity. I arranged all of my data into graphs, with Shelters A-J being in order from most rural to most urban.
Graph illustrating LOS for dogs of different ages across shelters
However, from most of these graphs, I found that while the difference in length of stay between shelters was in fact significant, it did not seem to be correlated with the population density of the area surrounding the shelter. As can be seen to the left, there was no clear trend of which age of dog was most likely to be adopted from a certain location.
Overall, I found that sex did not have a statistically significant difference in length of stay across the 10 shelters, but age, color, and breed all had a statistically significant impact. However, similar to the results shown for age above, I could not find a correlation/trend between age or color and the length of stay of dogs across the shelters. I did however find that some breed exihibited a different length of stay in urban or rural shelters. Those breeds being hounds, lap, and spitz dogs.
I found that hound dogs were more likely to be adopted from an urban shelter, as can be seen by the graph to the left, with a shorter length of stay meaning the dogs were adopted faster, and thus more likely to be adopted.
Conversely, lap and spitz breeds were more likely to be adopted from a rural shelter.
Discussion
From my results I was able to reject the null hypothesis, so location did indeed have an affect on the phenotypic characteristics of dogs adopted, however, I failed to prove my own hypothesis that certain breeds and ages of dog were more adoptable in an urban or rural environment.
I concluded from my results that while location does have an impact on what dogs are more likely to be adopted, this was simply a result of these being different locations that had different dogs available for adoption at different times. The difference between shelters was not a result of a difference in the population density of the city surrounding the shelter, but just in the fact that they were different shelters.
The implications of my results are that when people adopt dogs, each person has different criteria, and the characteristics they are looking for in a dog is not as highly dependent on their location as I had initially hypothesized. My results also implicate that the dog adopted from a shelter is more dependent on what dogs are available than on the people adopting.
However, there were several limitations in my study that could be addressed if I were to revisit this topic. First of all, my study period was only 3 months, while all the studies I gained inspiration from analyzed several years of data. Additionally, due to the nature of my data collection process and the fact that only two of the shelters included in my study were no-kill, I had no way of knowing the actual outcome of the dogs, and simply assumed that all the dogs got adopted. Additionally, I had no way of knowing how far people were willing to travel to adopt their dogs, so there could've been a confounding factor of people from urban areas adopting a dog from a more rural shelter or vise versa.
My results strength many of the result found in the other studies already existing that found about age breed and color of dogs that are more adoptable overall.
Reflection
Thinking back to my initial curiosity, I just wanted to look at what kinds of dogs were more adoptable, and this process, while it has answered one question, has led me to have even more questions. If I were to continue to do more research similar to this, I would want to explore whether the availability of dogs in a shelter has an impact on their length of stay and if the size of a shelter impacts a dog's likelihood of being adopted. This process has given me the tools to both collect that data and it has taught me how to analyze the data I collect.
There were a lot of points in this part of my journey where I was quite unsure of what to do next, like when I first began my data analysis I wasn't sure how I was going to compare across the shelters. When I was faced with this uncertainty, I turned back and looked at the papers my project was inspired by and looked at their methodology again, and tried to apply it to my own project.
If I were to go back and redo parts of my research, I think I would try to contact the shelters again, and see if I could work with them to collect the data, since doing all the data collection by myself by hand was quite tedious. I would also want to set up all my spreadsheets ahead of time so that the process of tallying up the days dogs were in the shelter would be more streamlined, so then I could spend more time on my actual data analysis. Overall, I just wish I could make minor revisions to my data collection methods so that I could spend more time on data analysis and figuring out what data was significant and where the variation between shelters came from.
Page updated
Report abuse