Develop statistical models or simulations

I need a single workbook that runs the Monte Carlo simulation the project team asked for and outputs the 95% confidence interval for expected yield. Use the input sheet where I paste the distribution parameters from our experimental runs, run 10,000 iterations, and produce a one-page summary with histogram, mean, median and the 95% CI that I can paste into the grant update.

Before I share the model with the research lead, make it easy for them to test assumptions and avoid accidental changes. Surface the sensitivity results on the summary page, label each input with its source experiment and date, highlight which inputs would move the median by more than 10%, and protect the calculation cells so they can only toggle assumptions from the input table.

I built a simulation using a normal distribution for parameter X based on limited runs and the PI now doubts that choice. I don’t have enough samples to prove normality and swapping distributions changes the tail risk. I’m worried about publishing misleading uncertainty. What diagnostic checks should I run quickly, and what is the most defensible interim choice to show the PI while we gather more data?

Each time a project changes, I rebuild the same sensitivity tables and charts for different parameters, which wastes days and produces small inconsistent variations across papers. Help me pick one modeling habit to adopt that reduces rework and gives reviewers consistent presentation of uncertainty across studies.