- categorical variable: each observation lands in a category instead of getting a number, nominal or ordinal ( Descriptive Statistics )
- summarize with a table of counts: frequency and relative frequency
- picture with a bar chart, bar height = proportion; test with chi squared tests
- example: 1000 ball bearings, three categories that are mutually exclusive and exhaustive, classified not measured
| classification | frequency | relative frequency |
|---|---|---|
| conforming | 910 | .910 |
| too thick | 53 | .053 |
| too thin | 37 | .037 |
Python
Output
contrast: chromium vs nickel
- chromium and nickel concentrations are measured, so quantitative: summarize with mean, median, standard deviation, quartiles ( the TI-84’s 1-Var Stats ) and a boxplot
- chromium: 31, 1, 511, 2, 574, 496, 322, 424, 269, 140, 244, 252, 76, 108, 24, 38, 18, 43, 30, 191
- nickel: 23, 22, 55, 39, 283, 34, 159, 37, 61, 34, 163, 140, 32, 23, 54, 837, 64, 354, 376, 471
- both right skewed, mean well above median: chromium , median 124, ; nickel , median 58, ( the 837 drags it )
Python
Output