For the mice data, instead of modeling weight linearly over time, model weight as a quadratic function of time by including a term for days as well as for days2 . Which model (random intercept, random...

For the mice data, instead of modeling weight linearly over time, model weight as a quadratic function of time by including a term for days as well as for days2 . Which model (random intercept, random slope) and correlation structure(s) seem most appropriate for these data? Describe your reasoning for the model you choose