Table 1 Summary of the proposed techniques (LSE: least squares estimation, MLE: maximum likelihood estimation, LogTLM: log-transformed linear model using maximum likelihood estimation, and Bayesian) for the estimation of k, regarding their assumption level, availability of uncertainty estimates (e.g., standard errors, prediction intervals), tools for statistical inference (e.g., p-values, confidence intervals) and their estimate’s asymptotic properties
Technique | Response variable | Level of assumptions | Uncertainty estimates | Statistical inference | Unbiased estimates | Requires priors | Uses information from previous studies | Can be used to propagate uncertainty |
|---|---|---|---|---|---|---|---|---|
LSE | fPARi | Minimum | No | No | Not applicable | No | No | No |
MLE | fPARi | Intermediate | Yes | Yes | Yes | No | No | Yes |
LogTLM | log(1-fPARi) | Intermediate | Yes | Yes | Yes | No | No | No |
Bayesian | fPARi | Maximum | Yes | Yes | Depends | Yes | Depends | Yes |