Statistics Posts for Improvement Insights

Improvement Insights Blog

Latest "Statistics" Posts

Cp Cpk Formulas and the Mistakes in Homemade Templates

Creating homemade Cp and Cpk templates often results in incorrect values. There are many, many mistakes you might make without realizing. Here are a few examples.

A customer sent me their home grown template for calculating Cp and Cpk and wondered why the QI Macros got such radically different values. It was easy to see from their data that they were using standard deviation, not Sigma estimator (Rbar/d2) to calculate Cp and Cpk. Use Stdev to calculate Pp and Ppk, not Cp and Cpk:

cp-cpk-formula-mistake1

They had run the QI Macros histogram on two columns of data, one measured at 0 degrees and one measured at 90 degrees.

Posted by Jay Arthur in Manufacturing, QI Macros, Six Sigma, Statistics.

Correlation not Causation

As John Johnson and Mike Gluck point out in their book, EVERYDATA, ice cream consumption and murder rates both go up in the summer, but that does not mean that eating ice cream causes murder. Rising summer temperatures seem to be involved.

They also provide a link to Tyler Vigen’s website, Spurious Correlations.

Does U.S. spending on space, science and technology cause suicides? No, but they are correlated at a 99.79%. These and other crazy correlations are available.

So don’t confuse correlation with causation.

Posted by Jay Arthur in Jay Arthur Blog, Six Sigma, Statistics.

What’s Wrong with the New Excel 2016’s Box and Whisker Plot?

Microsoft added a¬†box and whisker plot to Excel 2016, but it’s not everything you might hope for. Here’s an Excel 2016 box and whisker plot:

box whisker chart in excel 2016

You might notice that the whiskers have a crossbar on the end. It seems to have a spare “x” in the middle of each box and it’s a little hard to see where the median is. And there are unnecessary gridlines that are¬†considered chartjunk. It does, however, show the outlier below the first box.

Here’s what the QI Macros Box and Whisker Plot looks like:

box whisker plot in excel

The whiskers are whiskers. The median is easily visible.

Posted by Jay Arthur in Excel, QI Macros, Six Sigma, Statistics.