Be aware of MICE (Multiple Imputation by Chained Equations)!

Last week we talked about missing data and easy tools to tackle this issue. But are these simple methods really enough?

Challenge

⚠️The estimation of a single value causes insecurity

Statistical Method

📊Instead of estimating the missing value only once, we do this a lot of times (e.g. 100 times) using a randomized strategy. We get slightly different results, which we then pool to get one total result, e.g. with Rubin's Rule. This is called multiple imputation. The most common method is Multiple Imputation by Chained Equations (MICE).

Take Home Message

🛠️ Multiple imputation is more complex than single value imputation

💪 But you are rewarded with more meaningful results

 

 📢 Stay tuned and continue Building Statistical Confidence For Market Access