On Becoming a Packaging Engineer / Statistics I
Dear Packolic Community
I recently conducted a survey on the use of statistics as a Packaging Engineering tool
The result was as follows:
So, I decided to write some articles about it to share with you as a reflection of this survey and emphasize the importance of statistics in the work of Packaging Engineering
Why is statistical analysis matters?
When I studied high school a few years ago and college, there was a lot of talk about statistical process control and there were specific subjects that addressed it. To my surprise, when I started working, it was practically an unknown subject and it wasn't until I got to Corn Products, where I started, not only to hear from EPC, but to see it in action, with a lot of tools, in the hands of real gurus, like engineer Rafael Ivory and Engineer Benjamin Castillo, and I realized that even though statistics is an extremely valuable tool , few used it in their professional lives.
Statistics are for me, when used rationally and accurately, the most useful tool in packaging engineering work, because it yields data that gives room to make decisions, to obtain strong conclusions of a process, establish specifications and tolerances and approve molds and/or tooling.
What is not statistical analysis?
For pragmatic reasons, many have a habit of jumping to conclusions with little data, to make decisions, using average or a value obtained with very little amount of data or using dimensional and esoteric units, such as "much", "little", "much" that reflect only the lack of discipline in collecting data and carries the risk of making wrong decisions
Here are 3 examples of the misuse of statistics:
- A packaging supplier tells you that the material is within specifications because the average is within specifications. Is the answer accurate? No, the average does not reflect the dispersion of data and depending on this, some values may be out of specs
- Between Bill Gates and I we have an average of 9 Bn dollars. We already know that the answer is not required, while Mr Gates is at the maximum on the data scattering curve, I am at the lowest part of the curve
- I run a test of a new container, collect 5 point data, with an average critical value of 2.3 and with that I conclude that its behavior is different from the current packaging, which average critical value is 2.0. Is the conclusion correct? 2.3>2.0 statistically speaking? No, again, dispersion plays an important role in establishing through ANOVA or hypothesis tests that there is or is not significant difference
There is a phrase I use a lot when trying to explain the effect of variability:
MACHINES LIKE PEOPLE DON'T PERCEIVE AVERAGES BUT ONE-OFF FACTS
Where does statistics apply in Packaging Engineering?
Statistics have countless applications in Packaging Engineering and are worth exploring and applying them to perform more technical and accurate work that will reflect on the performance of a company
- Defining specifications in target values and tolerances
- Qualification of molds and tooling
- Qualification and validation of packaging in the conversion process
- Qualification and validation of packaging tests
- Comparison of conversion processes between suppliers
- Troubleshooting packaging material in inspection and packaging
- Define whether or not there is improvement with a packaging design change in waste reduction, defect reduction, increased productivity, etc.
- Design experiments that tell us which operation or specification factors influence the change of a technical parameter.
- Qualification and technical certification of suppliers
- A supplier's technical ability to meet a specification
- Modification of specifications based on packaging processes or the conversion process.
I conclude here and will continue further with the topic in the following articles explaining in more detail the statistical concepts behind each of the applications
If you have questions or questions or give your point of view please let me know in the comments section and/or email me
And remember this phrase:
"WITHOUT DATA, YOU ARE ONLY A PERSON WITH AN OPINION"
W. Edwards Deming
Jorge Maquita / Packinnovation Consultants
Mail: jorge.maquita@packinnovation.com.mx
LI: www.linkedin.com/in/packinnovation
FB: @Packoholicos