I'm working on a data set that examines the probability of observing animals during a 13 day period. The first question that is being asked is as follows: "Does observability of animals decrease following 3 days (first 3 days of the 13 day period) of exposure to predators?" To answer this question, I marked a sample of animals with colored ear tags/numbers and required people to record what animals they observed while afield. Originally, I analyzed this data with a Chi-Square test in SAS but after further examination of the data, I found that hours spent afield by the people declined after 3 days, which may bias my observational results. I may have significant results; however, is it driven by variation in the decline in hunter effort or is it a true reduction in observations. Please see a sample below (Treatment refers to the area of the property I conducted this study, which had a high-density of people; Hunter Effort = amount of time that day spent in the high-density treatment). My first thought was to analyze this data using a logistic regression model and include human effort as a random effect because I'm only interested in accounting for variation in hunter effort. I'm not interested in predicting the probability of observing an animal by hunter effort. However, today I was informed by a fellow colleague that I should run an ANCOVA rather than a logistic regression model and use human effort as a continiuous covariate in the model, Does anyone else have an opinion on how to analyze this type of data? If an ANCOVA is the best route then what is the difference between this and logistic regression? Additionally, I'm interested in accounting for the decline in human effort so I can hopefully determine if observations truly decrease because of the presence of people rather than a reduction in hunter effort, why wouldn't it be approriate to use hunter effort as a random effect? Thank you very much for all of your help!
| ID | OBSERVED_UNOBSERVED (1= OBSERVED; 2= UNOBSERVED) | DATE | YEAR | DAY | EXPOSURE | TREATMENT | Hunter Effort |
| 58 | 1 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 107:27:00 |
| 28 | 1 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 107:27:00 |
| 29 | 1 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 107:27:00 |
| 7 | 0 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 107:27:00 |
| 28 | 0 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 107:27:00 |
| 29 | 0 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 107:27:00 |
| 30 | 0 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 107:27:00 |
| 31 | 0 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 107:27:00 |
| 32 | 0 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 107:27:00 |
| 33 | 0 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 107:27:00 |
| 58 | 0 | 11/22/2008 | 1 | 1 | INITIAL | HIGH | 101:20:00 |
