# Reverse Process Capability

Process capability is a well-established technique for evaluating the degree to which a process is capable of delivering a product within specification. But what if the specifications are unknown or at best tentative?

The calculations of process capability analysis can be reversed so that for a given set of target capability values the associated specification limits can be generated. The calculation is straight-forward for a normal distribution but needs a bit more thought when it comes to asymmetric distributions.

(more…)

# The Lognormal Distribution

The lognormal distribution is a commonly used distribution for modelling asymmetric data.  It’s just the log of a Normal distribution right?  Well no, it’s actually the other way around.  You take the log of a lognormal distribution to arrive at a normal distribution.  Is it just me, but I always have a bit of a mental block about this, it always feels a bit back to front.

In this post I will explore the relationship between a lognormal distribution and a normal distribution.

(more…)

# Capability Indices and Process Shift

In this post I will derive a simple relationship between process shift and the capability indices Cp and Cpk.

(more…)

# Visualising the Null Hypothesis

In this post I describe an approach that I use to teach statistics.  The goal is to use JMP to help develop an intuitive understanding of some common concepts: hypothesis testing, p-values and reference distributions.

(more…)

# The Range Statistic

Quality control techniques such are control charts and Gage R&R studies sometimes feel as if they are stuck in the pre-computer age.  For example they typically use range as a measure of dispersion since it is far easier to calculate than a standard deviation.(more…)

# Statistical Thinking and DOE

An understanding of how variance propagates through a system can help us design experiments that maximise the precision of our results.

Traditionally scientists and engineers are taught how to conduct scientific experiments, then analyse the data in the context of their domain-specific knowledge (physics, chemistry, electronics, etc), and finally to apply statistical methods to calculate the precision of their results.

Statistical design of experiments reverses the sequence of data analysis, by first anticipating the precision of the results and how the precision is influenced by the experimental configuration.  (more…)

# Assessing Curvature

Statistics is not about conforming to some strange protocol that we neither agree with nor like.

(more…)