ACTL1101 Lecture Notes - Lecture 2: W. Edwards Deming, Summary Statistics, Quartile
Document Summary
In god we trust; all others must bring data. Summary statistics (e. g. summary() in r) typically include (all in $): min, max, mean, median (2nd quartile, 25%-ile (1st quartile, 75%-ile (3rd quartile) One also wants indicators of the spread (central moments): standard deviation ($ so comparable, comes from , skewness (no unit but comes from , kurtosis (no unit but comes from ) 2 estimation: assign a value to an unknown quantity. Suppose you want t a model to loss data: you have to choose a model, say, a parametric model, for that model, you need parameter values. You can (will!) do all sorts of errors: not have the right model model error", not have the right parameter values (even if the model is right) parameter error", not have a representative sample . Sampling frame error" in the book: all of the above Here we focus only on parameter error (only access to sample of population)
