I'm new at the StackExchange but got a lot of useful info by reading the posts in the past. Now I'm struggling to find the best way to statistically address my research question and would love to get some advice.

I have very cool data on chimpanzees playing prisoner dilemma 150 times. I additionally collected data with human adults in two different conditions: (1) FULL INFO: participants got all the information about their partner's choice; (2) NO INFO: participants got NO info at all about their partner's choice. My dependent measure being the proportion of cooperation. To sum up, I have three different time-series (Chimps, Full info humans, no info humans) with everything kept equal (same time points, equally distributed, etc.) (see below a plot with the tendencies of our three groups)

y-axis represents proportion of cooperation -smoothed-

My research question is: Which time-series is more similar than Chimpanzees': The Full info or the No-info human groups? I know that "similar" is an open-to-interpretation word (trends? values?...) If it helps, my ultimate goal is to see how chimpanzees are behaving over time: either as a human with information or a human without information.

I've been reading about the Granger causality test to see if either human group time-series is more useful (than the other one) in forecasting the chimps', but have doubts about (1) the theoretical implications of this test (and its applicability to my research question); (2) the procedure itself (how to do run the test, assumptions, etc.). I would very much appreciate any feedback in this respect!!

I already read the following posts thoughtfully but didn't get to fully comprehend them: How to compare two time series? How to statistically compare two time series?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.