What inspired this wish list request?
Current issue:
In the current implementation of JMP’s Fit Curve and Nonlinear platforms, the confidence intervals (CIs) for parameter estimates are fixed at a 95% confidence level. This limitation restricts users from aligning their statistical outputs with regulatory or methodological requirements that call for different confidence levels, such as 90%.
Example use case:
In the paper “Parallelism in Practice: Approaches to Parallelism in Bioassays” (Fleetwood et al., 2015), an equivalence test is used to assess parallelism by comparing the confidence interval of the slope difference to a pre-specified equivalence interval. The authors explicitly use a 90% confidence interval for this purpose. Since JMP only provides 95% CIs by default, users cannot directly replicate or validate this approach within JMP, limiting its utility for equivalence-based parallelism testing.
The approach is also detailed HERE.
What is the improvement you would like to see?
Proposed enhancement:
Add in option to adjust the alpha level within the platforms.
Why is this idea important?
Value to users:
- Regulatory alignment: Many bioassay and equivalence testing guidelines (e.g., USP <1034>) recommend or require 90% CIs.