ABI 150 Section A - R You Ready for Some Stats
Overview
As in life, the class is very much focused on present pursuits (field data collection), but we can also never ignore the future (upcoming final papers)! Today we dealt with this eternal, existential, and dread-inducing problem by splitting our class time up.
We spent some more time polishing some methodology questions that have arisen in our last two field sessions, then turned to our faithful leader Marshall for a crash course on a range of statistical analyses that may be useful to folks moving forward.
A rough schedule report:
10:00 - 10:10 - quick check-ins. How did everything go for our two groups (morning vs later morning)?
10:10 - 10:40 - further methods polishing. How far away is too far for a bird? Marsh Wrens - do they exist? What is nesting?
10:40 - 11:35 - Marshall academy statistics. What are our predictors vs response variables? What type of data are they? When do I use which analysis?
11:35 - 11:45 - plan for next time by co-MCs, oh wait, that's us (Marshall and Ryan).
Checking In
Everyone seemed to have settled into data collection on Thursday despite the warming temperatures and we had > 40 bird observations, the majority of which were the full 7 minutes.We all agreed that the area near Lot H (Stilt City) is the best place to observe our stilts.
Because of people some shuffling around, we ended up having 6 people in the early group and 9 in the later group. They always say hardship breeds innovation (this is ecology so maybe the better syaing would be 'hardship breeds evolution') and it did here too; we found out that three people are perfect for a group!
Methods Polishing
Group math:
And then there were three (people): Though it was de facto decided last Thursday, we as a class officially decided to move to three-person groups as the basic unit of stilt-watching. This means two observers and one note taker per group.
And then there were five (groups): We could now comfortably work in 5 groups instead of 4, as we have enough range finders and people for that (thanks Harvey for the extra range finder!) We do have 16 people though, so we are now aiming for 7 people in the earlier shift and 9 people in the later shift.
And then there was one (sheet): Previously, our physical data was split among three different recording sheets. Now, with one person doing all the note-taking, those sheets can/will be reduced all the way down to one. The trees (this is a wetland study so maybe the reeds?) will thank us for the reduced paper, and the notetakers will have an easier job of it. Shoutout to Riley for volunteering to format and print those new combined sheets.
Method questions:
How far away is too far away? There were questions about when a bird nearest neighbor was too far to measure accurately. Determining whether the rangefinder is bouncing off the bird or something near the bird is difficult at 100s of meters, so we had to make a decision whether we wanted to accept that possible source of error or go with another data recording option. We had a vote and decided that any bird > 250 meters from the observer would just be marked as 250 meters for the data entry sheet. A hard, but important question!
When do we count Marsh Wrens? They're small! They hide in reeds! You can always hear them! We decided that you have to be absolutely sure that a Marsh Wren is in the exact group of reeds you are using as the nearest neighbor, or else you have to pick a different bird or no bird (see above).
How do we categorize nesting? Most people seemed to have characterized nesting behavior as resting, unless they were actively looking around, in which case the behavior would be vigilance.
Marshall McMunn (Eminem?) Stats Crash Course
We have data now, which means we can start doing the really fun part: Stats! Figuring out which statistical test to do, formatting your data, and then actually doing the test, can all be difficult. Marshall broke it down for us.
Types of variables
Predictor = variable in a statistical model that is used to forecast the outcome. AKA independent variable. Example for us: time of day
Response = variable that is presumed to be the outcome of the predictor variable effects. AKA the dependent variable. Example for us: beak dips
Types of data
Continuous = data that can take any value within a given range. Think; distance from an observer to a bird.
Categorical = data that fits into a finite, discrete set of groups. Think; eating, sleeping or walking.
Binary = a data type that is either one thing or another. Think; yes or no, 1 or 0, alive or dead.
Based on this info you can draw a 3x3 matrix that can help you decide which analyses and stats to do. Here is a rough and incomplete outline of that matrix plus a photo of the actual one Marshall drew.
.JPG)
Feel free to refer back to this as we start analyzing data. There will be in-class data analysis discussions coming up and stay tuned for a multivariate analysis (statistical techniques for data with 3+ variables) from me in a few weeks!
Next Class:
Before next class:
- I will print data sheets
- Drink lots of water
- Keep thinking about interesting analysess you want to do
May 21 Info
- 6:30 AM group = Hannah, Lianna, Tanya, Marshall, Riley, Samantha, Sam
- 10:00 AM group = Ryan, Harvey, Josie, Carissa, Taylor, Lily, Olivia, Heuijae, Isabella
- Marshall and Ryan will be Co-MCs (Marshall at 6:30 AM, Ryan at 10 AM)
- Meet at the parking lots
IT WILL BE HOT!
Bring clothes appropriate for the heat, lots of water, and whatever else you might need. Communicate with people around you. Let an MC know if you are not feeling well.
Comments
Post a Comment