The Fundamental Thing to Know from Statistics for Design Engineering

Episode Transcript

As a product development engineer, what is one of the fundamental things you need to know from statistics? I’ll tell you after the brief introduction.

Hello and welcome to Quality During Design, the place to use quality thinking to create products, others love for less. Each week we talk about ways to use quality during design, engineering, and product development. My name is Dianna Deeney. I’m a senior level quality professional and engineer with over 20 years of experience in manufacturing and design. Listen in and then join us. Visit quality during design.com.

Do you know what 12 things you should have before a design concept makes it to the engineering drawing board where you’re setting specifications. I’ve got a free checklist for you and you can do some assessments of your own. Where do you stack up with the checklist? You can log into a learning portal to access the checklist and an introduction to more information about how to get those 12 things. To get this free information, just sign up@qualityduringdesign.com. On the homepage, there’s a link in the middle of the page. Just click it and say, I want it.

When I was an engineering youngling…really what I mean by that is that it was early on in my career in uh, well process design, process development…my employer worked with one of the local universities to develop a course for their employees to go through to further their education, to implement some new things within the company. The topic was design of experiments. So I want to say I remember about 30 people in the class and our experiences ranged from ‘we’ve been working for a long time’ to maybe ‘newly graduated/one to two years into our engineering career’. And this was a very formal class. We actually went to the university classroom. We had a professor, we had a textbook. On the first day of class, the teacher started teaching us design of experiments, some of the basics. And then he started backtracking a little bit from what he was talking about after he got some blank stares. And he kept having to backtrack little by little until he stopped and he said, wait, you guys know all about hypothesis testing, right?

To which none of us really volunteered a whole lot that we did. Even if we did, I’m sure it was the case that we had all had it in school at some point. We were using it, but not really thinking it was a hypothesis test. So we were using it, we were exposed to it, but we weren’t willing to raise our hand and say, “Yeah! We know all about hypothesis testing!” To which he was shocked and stammered, “I’m going to have to revise our syllabus because this is the baseline of design of experiments. If you don’t know hypothesis testing, then I’m not sure how to teach you the rest of this course.” And so that’s what we did. And this course was a semester long. It was one or two days a week after work, or maybe we were allowed to leave work a little bit early, like at four instead of five o’clock in order to attend this class.

And he taught us all about hypothesis testing and then moved into design of experiments. It was a really awesome class and it was very neat to be able to go with the people that I worked with. We were all learning it at the same time and implementing things as we were learning it in our engineering work. I later moved from process engineering to quality engineering, which involved a lot more statistics and a lot more to do with hypothesis tests. So I kind of took what I learned in those classes and ran with it.

So the fundamental thing to know from statistics for product design engineering is hypothesis testing. It’s used in so much of what we do in engineering.

It’s used in design of experiments like for our class.

It’s also used for sampling. How many do we need to test? Well, that depends on what we wanna know and what we’re testing!

It involves test results analysis. Is what we’re doing, good, bad, or good enough, and with how far with what confidence? It all has to do with hypothesis testing.

Gauge R&R studies and test method validations are based on hypothesis testing.

And even some SPC statistical process control) that can use hypothesis testing for some rules in monitoring changes in a process.

I’m not going to review the basics of hypothesis testing in a podcast. I will however, link to someone else that does an excellent job at presenting this. His name is Justin Zeltzer and he publishes videos on YouTube under Zed Statistics. I’ll link to two of his videos.

In preparing for this podcast, I was searching for a good resource to refer you to, and part of that was actually watching some of the videos. So I will link to two of his videos.

One is a short video that’s an introduction to hypothesis testing. This is his first video in his series about the topic, and it’s a good introduction to the concepts of it.

The second one that I’ll link to is one that’s a bit longer, but covers many concepts in statistics. If you’re new to it or haven’t used it in a while and you want a refresher, this is a good one to watch. The other reason I’m linking to this one is because of its last section, or chapter, which talks about p-hacking (or other ways to say it is data dredging and data snooping – these are all terms for the same thing). This is where instead of planning and conducting a test to learn something, to test out our hypothesis, we’re instead data mining from a huge database to try to learn…something. We want to try to tie a result of one thing to a variable of another. There are some bad things that can happen when we do this wrong or we’re not careful about doing this. In that Zed Statistics episode, Justin does a good job at showing how this relates with statistical theory.

We also talked about this in a previous episode of the Quality During Design podcast in October of 2022. It’s titled “Variable Relationships, Correlation and Causation”, and we talk about the ways that we can verify that we really have a causal relationship and we’re not just looking for correlations.

What’s today’s insight to action? Even if you’re really not into statistics, it’s not your thing, there’s still basics that we should learn and understand, especially in design engineering. One of those basics is hypothesis testing. So much of our decisions about whether a design is good or not good or good enough is based on this concept. Even if we’re working with others who are doing the calculations, we still need to understand the calculations and the assumptions made from it for ourselves. It will fundamentally change how you look at design activities. So much so, it’s a type of responsibility to understand it. Sometimes we need a reminder or a refresher. There are lots of good and free refresher videos out there, or you can do what I did early in my career and sign up for a class to get into it more in-depth.

If you like this topic or the content in this episode, there’s much more on our website, including information about how to join our signature coaching program, the Quality during Design Journey. Consistency is important, so subscribe to the Weekly newsletter. This has been a production of Deeney Enterprises. Thanks for listening!