03 idk

Experiment Pointers¶

• Always perform experiment with both trial & control samples

• Always get the raw data; processing should be done by analyst, not data providers

• Every data point should have central tendency & uncertainty associated

• Incorporate all potential uncertainty associated with collecting the data & use Uncertainty Propagation
• Preferably use robust summary statistics such as median and IQR, rather than mean & variance
• Every data point fed to model should be iid observation

Data Template¶

For Collection¶

For Modelling¶

We cannot use the collection data directly for modelling as each row is not iid observation. Hence aggregation is required to obtain the central tendency & uncertainty for each iid observation.

Comments