For the information of those downloading this article, I OLP briefly summarise the reviewers' comments from our first submission here (conclusion: revise and resubmit):
1. More acknowledgement of latent beta distribution possibilities for modelling plant cover data required.
2. Distinct power (freq. models) vs 'mean of means' (Bayesian models) approaches not very satisfying.
3. Consideration of Type 1 errors would be useful.
1. Paper has split goals (advance modelling of interval-censored data vs monitoring-specific power analysis). Reviewer suggested latter focus comes through more.
2. Authors overlook existing hierarchical approaches to ordinal regression.
2. Mid-point regression model could be dispersed with considering frequent existing criticism in the literature.
3. Acknowledge/include latent zero-inflated beta options.
Overall, the conclusion (similar between reviewers) was to focus in on a smaller number of models within a single framework, perhaps with an emphasis on power within the context of a specific monitoring goal. [Note that the modelling/simulation results in this preprint are not incorrect per se, but that the reviewers criticised the framing and selection of models considered, and on the appropriateness of the comparisons between models].