Simple research questions that can be answered using iNat data

Hi All,

I’m giving a talk for high school science teachers next year on using iNaturalist – not for posting observations but instead to answer student-led research questions.

During the talk I’m planning on introducing teachers to the web application and show them how to filter, then sift through observations to collect data that can be used to answer (high-school level) research questions. I also wanted to give them a list of sample questions they can use.

Here are questions I’ve brainstormed so far. I would love any additional suggestions people have for research questions, especially for vertebrates, morphological variation, and animal behavior!

Ideal questions a) can be answered via a bar chart or line graph, b) don’t require the data to be downloaded, are c) relatively simple to filter for the relevant species, and d) do not require the user to know how to identify species (i.e., what color flower is the animal on :white_check_mark: ; what species of flower is the animal on :cross_mark:)

Thanks!

1. What is the frequency of the number of rows of Prairie Blazing Star in bloom?

2. What is the frequency of different counts of squash bees in the flower?

3. What human structures do we see raccoons interacting with?

4. What is the frequency of perching structures for red-tailed hawks?

5. How many ducklings do mallards have?

6. How many oak galls do we see per leaf per gall species?

7. What color flowers do bumblebees most visit?

8. How is the timing of flowering of X species correlated to human population density?

9. Does the time of day song birds are photographed vary across the months?

10. For “sleeping” bees in the Tribe Eucerini in the project “Sleepy Bee Slumber Parties,” what is the variation in the number of bees per aggregation?

11. Which bird species are most frequently recorded in bird window strike collision projects?

12. What states / countries are migratory species in each month of the year? (Painted Ladies, Monarchs, Migratory Birds)?

13. What states/countries are widespread plant species in flower during each month of the year (dandelion, clover)?

14. What species of birds are found on each feeder type in the Project “Bird Feeders”?

4 Likes

It might be wise to avoid questions involving dandelions unless you with to get into discussions about apomictic reproduction and the taxonomic questions that go along with this (e.g. what is a species). Also because this means dandelions are likely to be less consistently curated than many other common, distinctive taxa (with some users calling all of them Taraxacum officinale, others going no further than genus).

It also may be helpful to make sure that flower species are selected where a good portion of observations in the target region have been annotated, otherwise it may be difficult to get meaningful results. (Or would the students be going through the observations manually and determining which ones are flowering?)

3 Likes

Student would be going through the species manually to determine which ones are flowering. And this is high school (ages 14-18) level. It’s supposed to be the process of doing science and having ownership over a project they are working on, not getting into the weeds (pun intended) of taxonomy.

5 Likes

sounds like a great project…
with this question below - is there another way to phrase this ? Are you interested in total number of perch-like structures? Or percentages for kinds of perching structures? {e.g., natural perches vs. human made structures - I have reviewed observations Hawks on top of cars…}

and this one

will be interesting…range, median, mode, etc.

1 Like

When I read your list of questions, it feels very “schooly” to me, meaning, unpleasant tasks that are imposed upon the student. To me, one of the greatest benefits of iNat is learning about nature via doing IDs for others. But that often requires a pre-existing love for nature . . .

1 Like

I don’t know how long you have with the teachers, or what level of interaction you can have with them, but how about allowing them to brainstorm their own questions with one another in small groups after you’ve shown them the application (which should include annotations like phenology and sex etc.). Ideally, the teachers could come up with some of their own.

One question I had a couple of college students do for their honors credit for a course was what effect latitude has on timing of flower. The annotated a species of cactus as to whether they were flowering and then compare peak bloom time (date) at the northern vs. southern ends of their distribution (different latitudes).

5 Likes

A few ideas from Ornithology:

-pick a common widespread species like a Song Sparrow and compare observations from desert environments and temperate rainforest environments (or Pacific Northwest vs eastern North America, or something else). Are there differences in:

colour? (Gloger’s Rule)

Bill size? (thermoregulation, diet)

something else?

(to answer those questions students would have to find some standardized way to measure traits from the photos, like scoring colour on a scale from light to dark, comparing bill depth to eye diameter in the photo, etc)

Lots of hypotheses that could be tested using trait data pulled out of photos about how colour, size, or shape varies with latitude, habitat, urbanization, etc.

5 Likes

Thanks! I have an hour so there should be enough time for them to brainstorm some on their own as well.

2 Likes

yah, i’ll standardize the phrasing. But also, ideally, that would be something for the student to figure out. Maybe I’ll rephrase all of them to just be the subject and then they have to figure out what to calculate – mean, median, mode, etc.

So Perching Structures for Red Tailed Hawks, Number of squash bees in squash, etc.

1 Like

This is a good idea! See this story about a paper that resulted from someone checking out wing coloration in dragonflies.

Just a heads-up, you can reply to multiple posts in one reply. Select the text you want to reply to and click “Quote”. It will appear in your post and you can write a response underneath it. Then you can quote another piece of text and reply below that one. Everyone you quote will be notified about it.

1 Like

If they have access to GIS tools, or some other way of creating dashboards, they could create interactive dashboards to display their research/other students’ research. While ArcGIS can be too expensive for a standalone project, if they already have access to it, it’s a great way to create relatively useful websites that can have some amount of lasting impact.

1 Like

What about: watch out for plant parasites and their hosts?

Or interactions in general?

Put your observations in a project like

https://www.inaturalist.org/projects/interactions-between-members-of-different-or-the-same-species

I would be a little cautious about any questions on the time of day (e.g. question 9 on time of day songbirds are photographed). Recorded times can be quite unreliable on iNaturalist.

Another question which might be easy for students to answer- figure out the breeding range of a migratory species by finding and mapping observations of juveniles (chicks, caterpillars, etc). Butterfly / moth / caterpillar records are pretty well annotated in North America, so it should be easy to isolate records of just caterpillars. I’m not sure how much work it would entail to map breeding ranges of other species.

Edit: just tried this for a few species. The common buckeye shows some trends where adults are found further north and west than caterpillars, but it seems most butterfly / moth species don’t have distinct breeding ranges, perhaps because they can have multiple generations per year and may breed in both their summer and winter ranges.

1 Like

Yup! I worked with Mike. Here’s another paper by yours truly looking at wing coloration in carpenter bees. Unfortunately behind a paywall, but PDF here.

2 Likes

I agree that if these are high school students it might be desirable to give them some ideas for possibilities to get started but then also allow them to brainstorm projects of their own – one of the reasons I never found science particularly interesting in school was because most lab work was based around pre-set experiments or tasks rather than finding questions to investigate that interested me.

One thing that might need to be considered is how to ensure fairness, meaning that answering the questions requires a similar amount of work. There are observation fields, projects, and annotation fields which could greatly reduce the amount of work in some cases, if their question involves a set of observations where one or more users has already been active filling out this information. This is obviously going to be less work than manually examining each observation for sets where this data has not been filled out consistently.

Depending on the maturity of the users, it might also be worth considering whether they might use iNat to record their data curation process. E.g., if they are examining observations to look at flowering periods, why not annotate those observations so that this information becomes available to future users? If they are recording variations in wing color, why not create an observation field to track this? One advantage to this is that iNat’s infrastructure then makes it fairly easy to sort this data. (Of course, if the point is that they learn to do this manually using excel or some other method, this might rather defeat the purpose.)

1 Like

If this is supposed to be for a science class, rather than a data management class, I’d suggest looking at ideas that can start at the stage of a hypothesis, have the students develop their own simple methodology and end with an explanation of whether/why the hypothesis was dis/proved. Most of the suggestions in the OP seems like they are just asking a pre-packaged statistical questions that demand a single statistical evaluation.

With what I am suggesting, the students don’t need to come up with their own hypothesis, (though they should be encouraged to do that IMO). They can be easily presented with a question or observation, and then asked to formulate a hypothesis, a basis/explanation for that hypothesis and then a simple test to attempt to falsify the hypothesis. Depending on the results they can then either say that the hypothesis is confirmed, or attempt to explain why it may have failed.

Example:

Obs - In areas with lots of racoons, I never see any coyotes. Q - Do areas with coyotes have fewer racoons?

Hypothesis Coyotes kill/compete with racoons, so the more coyotes, the less racoons you’d see. Alternative, areas with lots of food and dens for racoons will also have lots of food and dens for coyotes, so the more coyotes, the more racoons.

Test – Look at the number of coyote and racoon observations in several similar sized areas over a period of time. This could potentially be done at a suburb/city level. Graph them and see if there is any obvious correlation. Calculate the average number of coyotes and racoons in each observation area, classify into 4 groups – racoons above average/coyotes above average, racoons above average/coyotes below average, racoons below average/coyotes above average, racoons below average/coyotes below average.

Explanation – Hypothesis confirmed – other possible explanations/hypothesis falsified, explanation of observations.

Reconsidering the suggested questions:

Obs - I always see a lot more large squash bees than small ones. Q? Would the frequency of squash bees observed be a normal distribution or does it skew to the top end?

Obs -In my neighbourhood, I only ever see racoons around houses or on fences, never on the treet or in parks or laws. Q. Do raccoons in urban areas stay close to buildings?

Obs – I think that hawks only land in trees, even when there are buildings around. My friend thinks they prefer to land on buildings if there are any around. Q - Do red-tailed hawks prefer to perch in trees or on buildings?

Obs - The first mallard clutches that you see every year are only one or two ducklings with their mother, after a few weeks you start seeing mothers with lots of ducklings. Q- Do mallards have larger clutches early or late in the season?

Q - Are bumblebees more likely to visit light coloured flowers in the shade of a forest?

Q -Do bees form larger aggregations in wet/cold weather?

Q - Are larger, less manoeuvrable bird species more likely to collide with windows?

Q - Does a migratory species start to move out of their winter/summer range earlier in more northerly/southerly areas, or do individuals all start to move at once?

Obs – Trees always come into leaf earlier when late winter/early spring is warm. Q - Does deciduous tree species x reshoot in spring because air temperature increases? (there are dozens of variations on this – day length, precipitation etc. causing leaf fall/reshooting).

Q - Is there a trend in when widespread plant species X starts flowering – eg is it flowering later each year?

Q - What species of birds are found on each feeder type in the Project “Bird Feeders”?

Q - Do migratory birds with access to feeders stay later into autumn/winter?

1 Like

I agree that having students create their own questions is the ideal process that will get them most invested. However, OP said that

So this seems to just be a presentation to the teachers about some types of questions that they can answer to help them understand the scope of what iNat data can address. The teachers themselves will best know what their particular classes can handle in terms of self-direction and how much scaffolding they need. If this is just sample questions for the teachers, it’s less important what the specific questions are and more important that they show the range of possibilities in iNat data so that they can help their students assess whether a question is viable to answer for their activity or not.

In my use of iNat in the classroom (college level, but with students of varying experience/ability), the most important thing I have been able to do is help students refine their brainstormed questions into something of appropriate scope. That involves helping add some complexity if their self-generated questions are super simple to answer or help simplify if their questions are technically too difficult or will take too much time for the activity.

I like many of the ideas above and would add that one of the “quickest” types of questions that users can answer with graphs are using iNat’s own graphs on taxon pages which show phenology and annotations. Students can look at graphs for different taxa and compare them to draw conclusions or filter graphs for different taxa and/or different locations (ie, does phenology differ in two locations?). Of course, there are all sorts of caveats that wouldn’t make this quick analysis suitable for a scientific publication, but the main value of these types of activities isn’t getting perfect data - it’s having students learn the process.

5 Likes

I disagree. Not all types of scientific research require formulating and testing a hypothesis; I think this is particularly true for biodiversity studies that do not include an experimental component. How many of the hundreds of papers using iNaturalist data collected in the wikis in the forum follow the method you describe? Very few, I suspect.

I also disagree that the questions presented here are merely a matter of statistical analysis that will provide a single, straightforward answer. Rather, analyzing the data is only the starting point of the investigation, not the conclusion.

There are still important methodological questions that need to be considered – e.g., identifying possible sources of bias in data that is opportunistic and not systematically collected (For example: Are people more likely to post observations of unusual color varieties than common ones? Does a peak in observations in late April mean that the species is most common then or are there other factors at play? If so, how does one control for such biases when using the data?). All of these are important skills for students who are learning how science is conducted in practice and all the messiness that it often involves.

And then there is are questions of how to interpret the data (starting with: are differences statistically meaningful or randomly distributed; if there are meaningful differences, say, in group sizes or color preferences, what are possible reasons for this, etc.). Asking the question “What does this data tell us?” and looking for explanations for it is where the interesting work begins.

1 Like

Yes cherry red, with a comment from our protea specialist.