Term | Definition |
---|---|

Non-Normal Distributions | A statistical distribution that reflects non-normal or atypical randomness in the spread of data points around the central tendency. This distribution doesn't look like the typical "bell curve" of a normal distribution, but may reflect an asymmetrical spread of more than 50% of the data points on just one side of the mean (i.e., the average or center of the distribution). Very often this uneven spread will appear like a tall peak of data points on one side and a long tail of remaining data points stretching across the other side of the bottom of the scale. A long tail forming to the right is a positive skew and a long tail of data points forming to the left is a negative skew. This non-normality implies there is bias or skewness (i.e., non-normality) in the data. The non-normality of a distribution may appear visually obvious but it should be statistically validated such as in using an Anderson-Darling (AD) test of a Normality Test or Probability Plot. |

Non-Value-Added (NVA) | A step in a process that does not add value (as defined by the customer's requirements). It is also referred to as "waste" or "Muda" (the Japanese term for "waste"). Some process steps may be defined as non-value-added but may be required, such as to confirm with safety or government regulations, to preserve financial integrity, or prevent future waste. |

Nonparametric Tests | Statistical tests that don't require the distribution to have certain parameters such as a normal distribution with parameters μ and σ. Since they don't require these parameters typical for normal distributions, these tests can be very powerful when testing non-normal distributions. |

Normal Distributions | A statistical distribution that reflects a normal or typical randomness in the spread of data points around the central tendency. This distribution is often referred to as a "bell curve" because the shape looks like a bell due to the symmetrical spread of 50% of the data points on either side of the mean (i.e., the average or center of the distribution). The normality implies there is little or no bias nor skewness (i.e., non-normality) in the data. The normality of a distribution should be statistically validated such as in using an Anderson-Darling (AD) test of a Normality Test or Probability Plot. |

Normality Test | A statistical test that plots the continuous (a.k.a. variable or categorical) values of a dataset along a logarithmic scale to assess if the data conforms to a normal distribution. The Anderson-Darling (AD) test is often used with this to measure probability (reflected as a P-value). |