Ahi Pepe | MothNet is going to run an experiment

There are five things that all experiments need. you can use our experiment as an example to make an experiment to test almost anything you want to know the answer to. You don't need to be a real scientist you don't need to wear a white coat, or make an explosion, you don't need crazy hair, or glasses (unless you are blowing something up then you need safety glasses), or a y chromosome (but it's ok if you happen to have one), or chemicals. But there are a few things you do need.....

  1. A QUESTION - e.g. Does vegetation restoration restore ecosystem connections?
  2. A TREATMENT designed to test the question (e.g. a vegetation restoration project)
  3. A CONTROL for anything that might affect the results - aside from the one thing we want to test. The treatment and control should be as similar as possible in every way except for the one thing we are testing. If we want to test how vegetation restoration affects moth diversity ideally we would find two locations that were as similar as possible and then randomly assign one to the treatment (vegetation restoration) and the other to the control (no vegetation restoration). The natural world is complex and somethings take a long time or happen over large areas this means Ecology is more difficult than most science fields so sometimes we need to be creative. Sometimes we just need to do a lot more work. In our case we have a difficult question because vegetation restoration takes longer than we have. One way of getting around this limitation is to pick two sites that are as similar as possible in every way except that one has a vegetation restoration project and then think of all the other things that might be important and measure these. We call these other things covariates.
  4. A RESPONSE - Something that we can MEASURE or COUNT that represents the answer. Our question is a bit difficult - But there are some things we can do to make it easier. First, if we can make the question a bit more specific this is easier e.g. instead of ecosystem connections we could say moth diversity; Instead of "Does vegetation restoration restore moth diversity?". The next tricky part is 'restore'. What do we mean by restore? what are we aiming for? What would restored moth diversity look like? The easiest way to decide this is to find somewhere that has the vegetation we are aiming for with our restoration project - A target or reference site. We still need to control for all the other things, the covariates, that might be different between the two sites. We can save ourselves a bit of work by sticking to the ones that are most likely to affect moth diversity. Ok so now we can say "Is the moth diversity in our vegetation restoration site (the treatment) more similar to the moth diversity in our target or reference site than the moth diversity in a similar site without vegetation restoration (our control site), taking into account all the other, hopefully minor, things (the covariates) that might affect moth diversity?"- Ok that's not as snappy but its a bit easier to work with.
  5. REPLICATION allows us to show that the effect of the treatment is real and not due to chance differences between the treatment and the control. Small effects are harder to find. More complex systems (like the nature world) have more other things going on so they make it more difficult to show the difference between the treatment and the control (the signal). You can think of it like looking for something you dropped. If the thing you dropped is large, it's easier to find, nearly every time you reach for it you'll get it. If the thing you dropped always falls in exactly the same way and place its easier to find.

Clutha Valley Primary students setting up an Ahi Pepe MothNet experiment