Here's an example of a field experiment that took place in actual fields. The empirical article appeared in Nature Communications, and the authors of the research wrote their own summary for a popular audience in The Conversation. The quotes below come from their popular audience summary.
In this study, the authors worked with cattle ranchers in Botswana who were trying to reduce the number of lions and leopards who attacked their herds.
The introduction outlines the theory the researchers were testing:
Many big cats – including lions, leopards, and tigers – are ambush predators. This means that they rely on stalking their prey and retaining the element of surprise. In some cases, being seen by their prey can lead them to abandon the hunt. We tested whether we could hack into this response to reduce livestock losses to lions and leopards in Botswana’s Okavango delta region.
Below are some details of the method of their study. It sounds like an experiment: Can you see why?
Working with Botswana Predator Conservation and local herders, we painted cattle from 14 herds that had recently suffered lion attacks. Over four years, a total of 2,061 cattle were involved in the study.
Before release from their overnight enclosure, we painted about one-third of each herd with an artificial eye-spot design on the rump, one-third with simple cross-marks, and left the remaining third of the herd unmarked. We carried out 49 painting sessions and each of these lasted for 24 days.
The cattle were also collared and all foraged in the same area and moved similarly, suggesting they were exposed to similar risk. However the individuals painted with artificial eye-spots were significantly more likely to survive than unpainted or cross-painted control cattle within the same herd.
In fact, none of the 683 painted “eye-cows” were killed by ambush predators during the four-year study, while 15 (of 835) unpainted, and 4 (of 543) cross-painted cattle were killed.
Questions
a) What were the independent and dependent variables in this study? What were the levels of each variable?
b) What kind of experiment was this: Posttest-only? Prettest-posttest? Repeated measures? Or concurrent measures? (Challenge question: How might your answer change if you take the predator's perspective, as compared to the cattle's perspective?)
c) Sketch a bar graph of the results of the study. (Later you can compare your sketch to the figures in the Nature article).
d) The authors don't seem to specify here whether or not they used random assignment to the three painting conditions. Why would random assignment be important?
e) Ask a question about the external validity of this study. (Which term will your answer involve: random assignment or random sampling?)
At the end of their popular press summary, the authors raise two important limitations:
First, it is important to realise that, in our experimental design, there were always unmarked cattle in the herd. Consequently, it is unclear whether painting would still be effective if these proverbial “sacrificial lambs” were not still on the menu. Further research could uncover this, but in the meantime applying artificial marks to the highest-value individuals within the herd may be most pragmatic.
Second, it is important to consider habituation, meaning that predators may get used to and eventually ignore the deterrent. This is a fundamental issue for nearly all non-lethal approaches. Whether the technique remains effective in the longer term is not yet known in this case.
f) How might you design a study to test the first limitation? That is, how could you manipulate the eye-paintings variable in order to test if this intervention still works if there are no "sacrificial lambs" in the herd? What would your different conditions look like?
g) How might the authors test the second limitation? What can they change about, or add to, their study to test whether "the technique remains effective in the longer term"?