# Converting Attribute Data (Pass/Fail) to a 1-10 Scale

Six Sigma – iSixSigma Forums General Forums Tools & Templates Converting Attribute Data (Pass/Fail) to a 1-10 Scale

Viewing 16 posts - 1 through 16 (of 16 total)
• Author
Posts
• #250012

Distocont
Participant

Hi there,

I am working in a manufacturing environment where we are producing wire coils. Every lot has different number of coils depending on the order of the customer and each coil has hundreds of spirals. Before delivery to the customer an inspection is being performed whether a coil is OK or Not OK. A coil which is defected may have defects such as darker surface on the wire. Let’s say the surface is darker and the coil is rejected. To me it is fairly an attribute data. But one of my colleagues is insisting that we should convert it into a scale of 1 – 10. 1: Least fine and 10: Most fine. Well it does not matter whether a coil is, for example, has rating of 1 or 9 it cannot be delivered to the customer. In order to deliver to customer it should be 100% OK or according to him a rating of 10. I said it is useless to convert a variable which is in its nature an attribute kind into continuous variable. Is it worth to convert such data into continuous data? I know that there are plenty of tools to use on continuous data and it is easier to work with continuous data. It also follows normal distribution. And one does not need more data points for the analysis of continuous data compared to attribute data. Can anybody give arguments in favor of my colleague? I want to convince myself.

0
#250021

Robert Butler
Participant

Based on your post there not only does not seem to be any need for a 1-10 rating but, it doesn’t sound like your colleague has any idea of how to build a meaningful 1-10 rating scale that would be of any value to you.

What your post does suggest is you have some understanding of why coils are rejected.  For example, you mentioned dark surface on wire is cause for rejection….and what else?

Assuming you are interested in reducing rejects then the first order of business is to quantify the reasons for rejection over some predetermined period of time and bean count their frequency of occurrence.  Once you have this information you can analyze it and see what you see.

For example:

1. Pareto chart the results to identify the most frequent reasons for rejection.  If all reasons for rejection are created equal then you could use that chart as a driver for examination of the process to find the cause(s) of the top two or three reasons for rejection.

2. Look at the frequency count of reasons for rejection over time – do any of them trend over time – daily, weekly, monthly, quarterly, etc.)  (I once ran an analysis of manufactured component failures and found a huge AM to PM delta – the reason – the plant wasn’t air conditioned and the temperature delta between morning and afternoon resulted in far more failures for PM production because part of the component production process turned out to be very temperature sensitive).

3. Check your raw material receiving – any trending in rejection reasons that appear to coincide with raw material lot changes/supplier changes, etc.

…and so on.

So, short answer – if your post is a reasonable representation of your situation then, at the moment, I don’t see any value in rating scale.

Changing subjects:

You said, “I know that there are plenty of tools to use on continuous data and it is easier to work with continuous data. It also follows normal distribution. And one does not need more data points for the analysis of continuous data compared to attribute data.”

1. There are plenty of tools to use to analyze attribute data too.

2. Continuous data will follow whatever the underlying distribution of that data happens to be – it could be normal but it could also be a myriad of other things. Continuity of data does not guarantee normality.

3. Data quantity can be an issue with respect to analyzing continuous vs attribute data but which will need more data will depend on what it is that you are trying to do and the kind of data you have.

1
#250030

Distocont
Participant

Thanks @rbutler for you time :-)

“What your post does suggest is you have some understanding of why coils are rejected.  For example, you mentioned dark surface on wire is cause for rejection….and what else?”

In fact it is his department’s process where coils are wet-processed such as degreasing and then coils go for the inspection before delivery. What I suspect is that he wants to draw some kind of correlation between the degreasing process and the output. But in my view if degreasing process is not optimized or efficient, it will produce dark surface on the wire’s surface. To me, it does not matter whether a coil is 10% clean or 90% clean, it is rejected i.e. it is not delivered to customer. It will be taken as rejected whatsoever.

I performed earlier a regression analysis: Binary regression analysis using minitab. I took pass/fail as response and predictors were also binary: whether coils are processed on A equipment or on B equipment. And then I draw a conclusion that which equipment is better A or B. In this case A: is an inline degreasing where wire is degreased in-line. While B: is batch processing where a coil is dipped into a degreasing solution.

I found it logical that not to convert the severity of defects on the coil on a scale because in the end it is rejected. I did not find any tool other than binary regression analysis in minitab to be suitable. I also thought that what if I would had convert the response into a scale of 1-10 and use a pareto chart. What then? How I would connect response to the process? I did not find the answer to my question.

0
#250031

Distocont
Participant

Continuation…….

Just recently I have proved with some experiments that it is the degreasing process i.e. ineffective degreasing causing dark patches on the wire’s surface. And in batch degreasing process (B equipment), parameters are: Temperature, degreasing agent, lubricant system and rinsing after degreasing. Before degreasing wire is drawn to reduce dimension with the help of a lubricant system. The lubricant system remains on the surface and the aim of the degreasing is to remove lubricant system completely from the wire’s surface. There will be dark surface if there are remaining of lubricant system on the wire’s surface. And if rinsing is not effective there will be dark surface on the wire. I do not know how this scaling 1 – 10 will help him to understand the cause of rejection. That’s I want to understand. Additionally, there are couple of people who judge the coils after processing. To me, determining how much percent of coil has dark surface is quite subjective. Having said that, different people have different evaluation of same coil due to subjective nature of the evaluation.

0
#250033

Distocont
Participant

“Pareto chart the results to identify the most frequent reasons for rejection.  If all reasons for rejection are created equal then you could use that chart as a driver for examination of the process to find the cause(s) of the top two or three reasons for rejection.”

In this case, the reason of rejection is only one problem which is caused by the ineffective degreasing and I have showed this to everyone with proof that degreasing is a problem. I do not know if a Pareto chart will be of any use in this case rather a p-chart can tell when the error rate is high or higher and one can connect it to the time period.

0
#250044

Robert Butler
Participant

I’m not sure where we’re going with this discussion.  The focus of your initial post was that of coil rejection due to defects.  In the follow up discussion I was left with the impression defect reduction and thus increased acceptance of coils was your main concern.  I was also left with the impression that one factor known be a contributor to defects was improper degreasing.  It was for this reason I suggested looking at defect types and using things like Pareto charts to help and perhaps guide the thinking with respect to identifying defect types and the production practices connected to them.

Your latest post gives the impression this isn’t your goal.  It seems like you are just concerned about your colleagues wanting a defect rating scale while you just want to run a simple defect yes/no count.  Under these circumstances the rating scale amounts to nothing more than colorful jargon and has no value because defect “level” is irrelevant. A defect is a defect and if there is a defect the product will be rejected.

1
#250046

Distocont
Participant

“I’m not sure where we’re going with this discussion.  The focus of your initial post was that of coil rejection due to defects.  In the follow up discussion I was left with the impression defect reduction and thus increased acceptance of coils was your main concern.  I was also left with the impression that one factor known be a contributor to defects was improper degreasing.  It was for this reason I suggested looking at defect types and using things like Pareto charts to help and perhaps guide the thinking with respect to identifying defect types and the production practices connected to them.”

I am sorry Robert if you get different impressions. May aim was not to confuse you rather to explain the problem with different aspects.

There is only one kind of defect which is termed as “dark patch” and it does not matter if  a coil is 10% cleaned or 90% cleaned because it will be delivered to customers.  The defect is caused by improper degreasing which we all are aware of that. Then he came up with the idea that defect (ok/not ok) should be converted into a scale of 1-10.

0
#250047

Distocont
Participant

I am really  thankful to your contribution. I mean it!

0
#250048

Distocont
Participant

“There is only one kind of defect which is termed as “dark patch” and it does not matter if  a coil is 10% cleaned or 90% cleaned because it will be delivered to customers.  The defect is caused by improper degreasing which we all are aware of that. Then he came up with the idea that defect (ok/not ok) should be converted into a scale of 1-10.”

Correction: My girlfriend mentioned my mistake that the sentence should read “and it does not matter if  a coil is 10% cleaned or 90% cleaned because it will not be delivered to customers“.

0
#250084

Chuck White
Participant

@Distocont the value of a rating scale depends on your reason for collecting the data. If you are only trying to understand what your reject rate is, then a rating scale has zero value. However, if you want to use the data to improve the process, then a rating scale in many cases can have a lot of value, provided you can rate the severity of the defect consistently.

It all comes down to getting information out of the data. The more resolution, the more information you can get out if the data. A rating scale has more resolution than pass/fail data. Think of a case where you have a machine that is producing a part with a diameter out of spec, and you are tasked with adjusting the machine to get it back in spec. Given the choice, would you use a caliper or a go/no-go gage to check the parts as you make adjustments? I hope you can see that you will be able to dial it in much faster and better with variable measurements (the caliper).

A rating scale can work the same way. Even without statistical tools, if you make an adjustment to the process, and you go from 50% clean to 80% clean, you may not have fixed it, but you know your adjustment is in the right direction. If you go from 50% to 20%, you know you need to reverse direction.

1
#250086

Fausto Galetto
Participant

There is NO NEED to convert anything to get the decision

1
#250141

Distocont
Participant

Thanks Chuck for your time :-)

You are right that it is the severity which can drive the improvement in the process. But there are some limitations: To run any improvement one must have a reliable output’s data, as I understand. In this case, we only have visual judgement i.e. no reliable method or instrument to measure the severity. Even one can’t count the dirty spirals out of total spirals because each coil has 1300 – 1600 number of spirals. And the place where a coil is inspected is limited in space. It does not tell how are those spirals which are underneath of the top layer of spirals. And there used to be at least 5-6 layers of spirals overlapping each other. Plus each spiral is in approx 1.5 m in diameter. Spirals are partially dirty. There is no way to quantify in what extent they are dirty.

0
#250171

Distocont
Participant

@jazzchuck “It all comes down to getting information out of the data. The more resolution, the more information you can get out if the data. A rating scale has more resolution than pass/fail data. Think of a case where you have a machine that is producing a part with a diameter out of spec, and you are tasked with adjusting the machine to get it back in spec. Given the choice, would you use a caliper or a go/no-go gage to check the parts as you make adjustments? I hope you can see that you will be able to dial it in much faster and better with variable measurements (the caliper).”

Yes you are absolutely right. Simple pass/fail won’t help much to improve the process. A caliper will be a natural choice. But in my case, there is no way or instrument to measure the severity of a coil except a visual inspection which is limited by the space because it is impossible to count the dirty spirals of a coil. One can only see those spirals which are visible on the top layer and have no clue if there are dirty spirals underneath of top layer. Plus inside of a coil is also hidden to visual inspection. And of course, there is human factor as well.

0
#250196

Distocont
Participant

Even if converting severity into a scale will not make it a continuous data. or will it?

0
#250386

Strayer
Participant

Might I suggest that your real problem is the measurement system?   You could install instrumentation to continuously monitor the wire before it’s wound into coils.  Sure, that’d take an investment that’s probably outside your sphere of control.  But there’s this basic cost of quality principle:  The later you discover a defect the more it costs.  Plus, inspecting only the visible part of the coils won’t find hidden defects.

0
#250500

Chris Seider
Participant

Color measurement exists.  Maybe see if you can get a variable data measurement device/panel and not have a scale.

0
Viewing 16 posts - 1 through 16 (of 16 total)

You must be logged in to reply to this topic.