“…This is confirmed by conducting a log-likelihood ratio test (LLR) to analyse whether the improvement in explanatory power observed by incorporating the Temp variable in Model II is statistically significant. In particular, we calculated LLR = 2[LL Model II − LL Model I ], where LLModel II and LLModel I are the log-likelihoods of Model II and Model I, respectively (see Green et al, 2019). We found that LLR = 5.6 (χ 1 2 [0.05] = 3.841).…”