100% Inspection - 80% Accurate - Quote from Juran (?)

Status
Not open for further replies.
U

Unregistered

#81
100% Inspection

Thanks for the information about the origins of effectiveness. The truth is a lot less simple! How difficult the inspection is decides how likely you are to find the problem. The example of "Is the fridge covered in paint?" is a good one. The likelihood of picking up an unpainted fridge when the primer is white is approx 100%. If the requirement is "Check the red paint coverage across the whole surface is a minimum of 0.25 mm thick and we're giving you a tape measure to check it" then the likelihood of finding a non compliance is somewhere near to 0%. Any other combination of checks and capability falls somewhere on or between these two markers.

The next point is that you cannnot add inspectors and bring down the probability. i.e 80% probability followed by 80% probability is not 64% probability of finding the problem. You have to take the opposite view. If the first inspector misses it what is the likelihood of the next inspector missing it 20% * 20% = 4%. Therefore the probability of two inspectors catching a fault becomes 96%. This assumes that the two inspections are not connected. i.e. the first inspection isn't cursory because "The next person will inspect it" or the 2nd inspection is cursory because "Well it's already been inspected."

I hope this has helped.
 
Elsmar Forum Sponsor
M

M Greenaway

#82
Strange Statistics

Dont know where that calculation of probability comes from but it sounds very flawed to me.

If the probability of missing a defect is 20%, then it is 20% for everyone. You could have ten inspectors checking the batch, each has a probability of missing the defect of 20% therefore the overall probability is still 20%.

Does one inspector missing a defect make it more likely that a following inspector will catch the defect - I think not, the calculation is flawed.
 
D

D.Scott

#83
I think the poster is going for the probability of the combination rather than the individual probability of the second inspector.

MG is right - the probability of the single inspector is always 20% (simple event), but in the poster's example the second inspector is dependent on the first inspector missing the part (multiplicative law). To determine the probability of both events happening you have to multiply the probabilities (in this case 20% x 20%) which gives you the probability of both inspectors missing the defect (4%). The inverse would be the chance of discovery (96%).

When it is looked at that way, it makes more sense.

Dave
 
M

M Greenaway

#85
Cant get my head around it...

Cant get my head around this one.

I can appreciate that if an inspector picks one piece from a batch and the probability of selecting a defective piece is say 20%, that if he selects another piece the probability of them both being defective is 20% X 20%, and the probability that either is defective is 20% + 20%. But I cant see how this simple theory relates to two inspectors that are unrelated by the process.

Its a bit like saying that the probability of me winning the lottery is 1/14000000, so the probability of someone else winning is 1/14000000 X 1/14000000.

It just dont make sense.
 
D

Dave Strouse

#86
Lotteries

Martin -
We aren't asking what are the chances of winning the lottery. We are asking what are the chances of both winning succesive lotteries.

Make a simple tree diagram. The first inspector has an 80% chance of finding it and a 20% chance of missing it. Make a forked branch with one labeled 0.2 for the chance of missing it and the other branch label 0.8 for the chance of finding it. This is the first inspector. Now, make a second fork ON BOTH BRANCHES and label in the same way. (the second inspection would not be done in the real world on the 0.8 branch because the defect was already found, but we need it to count all the probabilities).

The branch where the defect was found by the first inspector has probabilities of 0.8 times 0.8 = 0.64 (multiplicative rule for independent events) for the case where BOTH inspectors find the defect and 0.8 times 0.2 = 0.16 where the first finds it and the second does not. On the other branch where the first inspector missed it , we have 0.2 times 0.8 = 0.16 and for the case where both miss it we have 0.2 times 0.2 = 0.04. Now adding them all up we get 0.64 + 0.16 +0.16 + 0.04 = 1 which confirms we have not missed any cases. ( that's why we had to "do" the second inspections even when we found it on the first).

The only case we are interested in is the one where both inspectors miss the defect. This is the 0.04 probabilility. So we can say there is 1-0.04 or 96% chance of the defect being found by two inspectors versus 80% chance of one finding it.

Each inspector has a 80% chance of finding it but both together in sequence have a 96% chance.

So your chance of winning the lottery or anyone elses in THAT lottery could be 1/140000000. The chances of you winning the lottery AND any other individual winning the next lottery (both winning in sequence) is 1/140000000 times 1/140000000.

But the chances of ANY body winning within any one drawing is the chances of each summed. So if 1000000 tickets are sold , the chances of one of them being the winner is 1/140. (assuming the numbers played are all different.) That's why people do win, but your individual chances are low.

Hope that helps.
 
E

energy

#88
Re: Lotteries

Originally posted by Dave Strouse
Martin -
We aren't asking what are the chances of winning the lottery. We are asking what are the chances of both winning succesive lotteries.

Make a simple tree diagram. The first inspector has an 80% chance of finding it and a 20% chance of missing it. Make a forked branch with one labeled 0.2 for the chance of missing it and the other branch label 0.8 for the chance of finding it. This is the first inspector. Now, make a second fork ON BOTH BRANCHES and label in the same way. (the second inspection would not be done in the real world on the 0.8 branch because the defect was already found, but we need it to count all the probabilities).

The branch where the defect was found by the first inspector has probabilities of 0.8 times 0.8 = 0.64 (multiplicative rule for independent events) for the case where BOTH inspectors find the defect and 0.8 times 0.2 = 0.16 where the first finds it and the second does not. On the other branch where the first inspector missed it , we have 0.2 times 0.8 = 0.16 and for the case where both miss it we have 0.2 times 0.2 = 0.04. Now adding them all up we get 0.64 + 0.16 +0.16 + 0.04 = 1 which confirms we have not missed any cases. ( that's why we had to "do" the second inspections even when we found it on the first).

The only case we are interested in is the one where both inspectors miss the defect. This is the 0.04 probabilility. So we can say there is 1-0.04 or 96% chance of the defect being found by two inspectors versus 80% chance of one finding it.

Each inspector has a 80% chance of finding it but both together in sequence have a 96% chance.

So your chance of winning the lottery or anyone elses in THAT lottery could be 1/140000000. The chances of you winning the lottery AND any other individual winning the next lottery (both winning in sequence) is 1/140000000 times 1/140000000.

But the chances of ANY body winning within any one drawing is the chances of each summed. So if 1000000 tickets are sold , the chances of one of them being the winner is 1/140. (assuming the numbers played are all different.) That's why people do win, but your individual chances are low.

Hope that helps.
I was just going to say that!:vfunny: :biglaugh: :bonk: :ko: :smokin:
 
R

Rick Goodson

#89
Don't remeber where it cam from, but I always used inspecting 3 and one-half times gets 99% of the defects. May only be an old wives tale.
 

SteelMaiden

Super Moderator
Super Moderator
#90
Originally posted by Rick Goodson
Don't remeber where it cam from, but I always used inspecting 3 and one-half times gets 99% of the defects. May only be an old wives tale.
So to get that 1/2 inspection does that mean someone is doing a half a##ed job, or can we hire one inspector that is half-blind?

I'm sorry, forgive me, I just couldn't help myself.:confused:
 
Status
Not open for further replies.
Thread starter Similar threads Forum Replies Date
S 100% Final Inspection Requirements of Active Medical Devices (MDD or IEC Standards?) EU Medical Device Regulations 4
E Is there a specific regulatory requirement for 100% inspection of IFUs? Inspection, Prints (Drawings), Testing, Sampling and Related Topics 4
N Switch to 100% inspection when lot fails sampling? Inspection, Prints (Drawings), Testing, Sampling and Related Topics 9
G Is Controlled Shipping requirement Sampling Based or 100% Inspection? Misc. Quality Assurance and Business Systems Related Topics 3
N 100% Inspection in Retail Warehouse which Ships Products to Customers Inspection, Prints (Drawings), Testing, Sampling and Related Topics 3
K 100% Inspection or do the usual IQ, OQ, PQ Validation - Medical Molding Design and Development of Products and Processes 3
G Am I silly to do a 100% inspection? Inspection, Prints (Drawings), Testing, Sampling and Related Topics 15
M 100% Inspection - 85% Effective Inspection, Prints (Drawings), Testing, Sampling and Related Topics 10
M Vision System - Camera vs. Sensor to perform 100% EOL Inspection Quality Tools, Improvement and Analysis 2
S AQL .65 Level III - Take a particular component off 100% Inspection AQL - Acceptable Quality Level 4
A Injection Mold 100% Visual Inspection - Short Shots and Debris not being found Inspection, Prints (Drawings), Testing, Sampling and Related Topics 22
M SPC vs. 100% Inspection - Is SPC still value-added when 100% inspection is in place? FMEA and Control Plans 40
S Applying ANSI/ASQ Z1.4 given known lot size and 100% visual inspection AQL - Acceptable Quality Level 2
S Is there an Accept/Reject Criteria for 100% Inspection? Inspection, Prints (Drawings), Testing, Sampling and Related Topics 9
E 100% continuous inspection question Inspection, Prints (Drawings), Testing, Sampling and Related Topics 14
R Transfering Cpk specification to 100% inspection specifications Capability, Accuracy and Stability - Processes, Machines, etc. 24
S Condition of 100% inspection in organization. Capability, Accuracy and Stability - Processes, Machines, etc. 7
G Is 100% Inspection a Control or does a Process need another type of control? FMEA and Control Plans 9
Q Who is responsible for 100% Inspection (QC) activities? Inspection, Prints (Drawings), Testing, Sampling and Related Topics 16
1 U Charts, Defects and 100% Inspection Statistical Analysis Tools, Techniques and SPC 8
C Acceptance Sampling is Better Than 100% Inspection Inspection, Prints (Drawings), Testing, Sampling and Related Topics 10
J Visual Inspection Not 100% Effective - How to Improve Visual Inspection Quality Tools, Improvement and Analysis 51
J Customer Complaint Corrective Action Response - 100% Vision Inspection Customer Complaints 9
M 100% Measurement Shift to Sample Inspection? PCB Substrates Inspection, Prints (Drawings), Testing, Sampling and Related Topics 3
L Customer Requires Cpk 1.33 with 100% Inspection Inspection, Prints (Drawings), Testing, Sampling and Related Topics 18
R 100% Visual Inspection - Would you Remove or Reduce the Visual Inspection? Inspection, Prints (Drawings), Testing, Sampling and Related Topics 9
G Inspection 100% - Process continuos (Continuous Process) Inspection, Prints (Drawings), Testing, Sampling and Related Topics 7
S 100% Inspection Control Chart By Dr. Feigenbaum - SMT (Surface Mounting Technology) Inspection, Prints (Drawings), Testing, Sampling and Related Topics 2
S Tips for convincing general labour that 100% inspection not required Various Other Specifications, Standards, and related Requirements 18
S P-Chart Question - Need Guidance - 100% inspection of all product parts required? Statistical Analysis Tools, Techniques and SPC 11
T Prototype Inspection - 100% Inspection on All Parts or just a Sample? Inspection, Prints (Drawings), Testing, Sampling and Related Topics 5
A How effective is 100% inspection Inspection, Prints (Drawings), Testing, Sampling and Related Topics 25
R 8.2.3.1 Monitoring and measurement - Containment of product and 100% inspection Inspection, Prints (Drawings), Testing, Sampling and Related Topics 1
J Final Inspections - Does 100% inspection really work Inspection, Prints (Drawings), Testing, Sampling and Related Topics 5
CarolX Accuracy of 100% Visual Inspection - Proofs Inspection, Prints (Drawings), Testing, Sampling and Related Topics 2
Ninja Windows 10 100 % disk use after update Coffee Break and Water Cooler Discussions 33
A 21 CFR part 11 - section 11.100 - Electronic Signature Certification Other US Medical Device Regulations 6
Stoic Manual soldering processes - 100% verifiable, or always requiring validation? 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 15
E Correct way to certify hydrostatic testing when it is not 100% (and Sample Size) Various Other Specifications, Standards, and related Requirements 6
A What to do when 100 % verification is not 100 % Manufacturing and Related Processes 5
Nicole Desouza ISO / AS Certification - Small business with less than 100 employees ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 7
M EU – Chemical Safety Report – CSR – REACH Authorisation decisions – Triton X-100 – Ortho-clinical-Use1 REACH and RoHS Conversations 0
D Measuring fine tubing (< 0.100") "Wall by weight" General Measurement Device and Calibration Topics 11
Marc Anniversary of the First Trans-Atlantic Flight 100 years ago - June 2019 Coffee Break and Water Cooler Discussions 2
Y IEC 62304 Section 4.3(a) - 100% probability of failure IEC 62304 - Medical Device Software Life Cycle Processes 3
BHobbs_Busche All Measurements taken with a Variable Gage have to be Recorded 100% IATF 16949 - Automotive Quality Systems Standard 14
Pmarszal Clarification for 21 CFR Part 11.100 - General Requirements Other US Medical Device Regulations 14
B Is Gage R&R required on a 100% Dimensional Layout? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 6
K IEC 61010-2-100 Labeling question - IVD equipment Other ISO and International Standards and European Regulations 1
D Long length measurement of 18" up to 100" with an accuracy of .005 min General Measurement Device and Calibration Topics 8

Similar threads

Top Bottom