How to Conduct a Kappa Attribute Gage R&R
What is a Kappa Test? Hint – it is not a test given to a college fraternity (i.e. Kappa Pi). A Kappa Test is an awesome statistical tool, and you can use it to get more consistent product assessments from your manufacturing team.
What do I mean by assessments? I mean the subjective decisions made on your products’ workmanship quality level. This includes the grading or assessment of the visual aspects of your product or the taste or even the smell. If your inspectors and operators are assessing products by using their perception (rather than a measurement tool), then a Kappa Test is the tool for you!
Kappa studies can drive improvements in a variety of applications to align your inspectors, operators, engineers, and managers to achieve consistent product assessment results. Examples where this tool applies includes visual inspections associated with print quality, plastic inclusions, packaging seals, internal metal burrs, potato chip defects and basically any application whereas a quality attribute can’t be physically measured. The Kappa test is also a good fit for aligning taste and smell criteria such as seasoning taste or the sweetness of bourbon or the dryness of wine.
Another nice part about the Kappa test is that the grading can be binary (go or no-go, yes or no, pass or fail) or the data can be scaled (i.e. scale of 1 – 10 attribute scoring). So, the Kappa test is very flexible and can be used in most applications where product attributes get graded by human perception and not physically measured.
Who invented this? Jacob Cohen was an American Psychologist and statistician. He developed the Kappa Test in 1960 because he wanted a more accurate method for assessment ratings. At the time, “Percent Agreement” was the primary tool used to do this, and Cohen’s assumption was that graders were guessing to some degree and so he wanted a statistical tool to account for grader guessing. Now the Kappa Test is used to statistically compare attribute assessments. The Kappa test is sometimes called the Cohen Kappa Test or an Attribute Gage R&R.
To perform a Kappa test, I recommend using Minitab because MiniTab can be used to complete many statistical studies such as Kappa, Process Capability, DOEs, ANOVA and much more. Ideally your company purchases at least one MiniTab seat but note that a MiniTab licenses is indeed pricey but the payback is good. But there are some good free on-line sites that will work as well. If you want to go the free route then check out The R Project for Statistical Computing. Whatever you decide on, just find and stick with just one software package. It makes it easier to complete all of your of statistical analysis when you learn how to use just one software package.
Now let’s understand workmanship and product variability. Acceptable workmanship is determined not by the company, but by the customer. Since we’re all customers, this is an easy concept to understand. Let’s use our work shoes as an example. You probably don’t care too much if the printing on the shoebox is a little blurry. The truth is that you probably won’t even notice. And if the stitching is straight and not unraveling, you probably won’t notice some minor inconsistencies there either. But the consistency of the color is important. If the shoes look blotchy, you’re probably not buying them.
What about comfort? Well, that depends on the price. If the shoes are designed well, they should be comfortable enough. But materials matter. Softer leather materials cost more, and the additional cost is transferred to the customer. So, you may be willing to pay $30 for a pair of shoes that may not be perfectly comfortable if you are only going to wear them once to a wedding. But you’ll gladly pay $75-$125 or more for a pair of good work shoes that you wear almost every day in a plant because comfort is important for your everyday footwear.
What’s my point? My point is that the customer determines your quality standards. There are some attributes that are non-negotiable regardless of price and some other expectations of product quality that are negotiated by the customer based on the product's price.
Manufacturing companies should know what their market position is on each product. Are you a premier brand or a bargain brand? There’s a cost associated with quality and the workmanship standards should reflect the price strategy. But a few things need to be considered.
First, at a minimum, you need to be meeting all the non-negotiable standards that your customers expect. Otherwise, they won’t purchase even discounted products.
Secondly, the cost to improve quality should be considered. If you can improve the quality of your product for little additional cost, you should improve it. For example, can lower-priced shoes be made comfortable even when made without real leather? If so, the shoe company would have a market advantage by selling comfortable shoes at a discounted price.
To summarize this concept; While we would all like to make 100% perfect quality, the truth is that there is a range of acceptable visual quality and there is a difference between an imperfection and a defect. The goal of the Kappa study is to get everyone on the same page. What exactly is an imperfection and what constitutes a defect? Imperfections ship to customers, defects do not. Everyone should know the difference between them, otherwise it's costly to the company and risky to the customer.
Conducting a Kappa Study provides product consistency which is a win-win for your manufacturing company and for your customers! You can use this tool in a variety of assessment applications, products, and assessment methods. It’s simple to use and when combined with a well-written workmanship standard, it yields long lasting results.
Do you want to explore more continuous improvement tools? Checkout Tools for the Trenches manufacturing training!
Stay connected with news and updates!
If you want some weekly T4T wisdom coming straight to your inbox for your reading pleasure - look no further! Join our mailing list to receive the latest blogs and updates.
Don't worry, your information will not be shared.
We hate SPAM. We will never sell your information, for any reason.