I fit a model with brms in the following way:
e <- rnorm(100, 0, 0.1) v <- rnorm(100, 0, 1) data <- tibble(x=v, y=v+e) form <- y ~ s(x) fit <- brm(form, data=data, family=gaussian(), control = list(max_treedepth = 15, adapt_delta=0.999), iter = 9000, warmup=1000, cores = 4)
Then summary of the fit includes smooth term sds(sx_1) and population-level effects Intercept and sx_1. I have understood that the smooth term describes the wigglines of the spline, and the other terms describe the perfectly smooth term (linear component?). So, if sds(sx_1) is close to zero, then the spline is actually a line defined by the Intercept and sx_1 as the slope. But this does not seem to be the case as the value of sx_1 is 5.88, where 1 was expected. How should the coefficients be interpreted? And is it possible to decide based on the fit, whether the spline term should be included or be replaced by a linear term?