Rockies and High Plains Vector-borne Diseases Center Summer Internship

 Hi everyone,     I thought it would be worthwhile to give an update about our results with the tick and understory project. We have come up with a final model: a zero-inflated Generalized Linear Mixed Model with a Conway-Maxwell Poisson distribution. This type of model helps with count data, where there are many zero values and some underdispersion. Underdispersion occurs when our observed data's variability is less than the mean or theoretical variability. The results tell us how ticks increase/decrease in abundance with given parameters, or how likely they are to be present/absent given similar parameters. This model ended up fitting the best with our data and gave some interesting results.     Concerning understory data, our models tell us that ticks are more abundant in areas with more flowering plants and more litter relative to other types of plants. The other important parameters were the time during the summer, altitude, and the distance from the trail....

Hello,     We have been working hard on analyzing the results for our tick study from last summer. As mentioned before, we are studying the correlations of tick presence and numbers with different vegetation types in Colorado wild areas. I have been modeling our data with some different parameters to find the greatest fit and to see where these correlations are significant/strongest. Nothing is final, but some interesting patterns and outcomes are worth noting.       We have decided to use a Zero-inflated Generalized Linear Mixed Model with our data, which splits the results into two categories, one for the correlations with the presence of ticks alone, and one for the correlations with the number of ticks. No matter what I have done so far in the modeling, one variable has a very strong correlation with tick numbers: time of year. This was to be expected; we saw a drastic drop in abundance as the summer progressed last year, and it's nice to see that this ...

      At this point in the year, ticks aren't out nearly as much, so it's hard to get great numbers for our mark-recapture study this year. We have stopped fieldwork for the study and are now working on the data we collected from last year and this year. My colleague, Jon Wegryn, and I have been looking into different factors of the understory that may be correlated with tick presence or abundance. I have been researching some statistical models to capture the data as well. So far, it seems that a zero-inflated, mixed-effect model is going to be the best option for our data, but I'm curious to see what parameters will have the largest effects on our data and what correlations are present.      In my free time, I have also been looking into Bayesian Statistics. As I learn more, it seems that there could be some application to this study as well, but I'll have to do more research and keep learning to see if that's true. Overall, it has been really interesting...