Don't let statistics become a stumbling block to p

Don't let statistics become a driving force 6 σ Stumbling block

6 σ Relationship with statistics

“ σ” Is it shrimp bowl cake? People who know a little about statistics can see "6 σ” It cannot be separated from statistics. “ σ” It is a statistical symbol, which is pronounced as "Six Sigma" in English and called "six standard deviation" in Chinese. It represents the degree of variation of the observed process or product. One side 6 can be clearly seen from the probability distribution diagram σ The area within the specification limit covered is much higher than that of unilateral 3 σ， In other words, 6 σ The probability of defects in the standard operation process is far lower than 3 σ Level. More specifically: reach 6 σ The defect rate of standard operation process is only 3.4 parts per million. Because 6 σ The quality problem is quantified with simple graphics, so the statistics can be called 6 σ It runs through every step of the improvement project DMAIC (please review the report in the last issue), and also affects the final effect of the improvement project

the right medicine depends on reliable data

understanding statistics in 6 σ After playing an important role in quality improvement activities, next we will focus on the most troublesome problem in quality improvement activities: finding out "symptoms" and using the right "prescription". Fortunately, the basic statistical principles and ready-made statistical software tools have solved the long-standing problems. Through an effective data collection plan (including: determining the object of data collection, sample size, sampling method, measurement system analysis...), we can use single-layer tepex material components to maintain the sound insulation effect and find out the actual "symptoms" (for example, after data collection and analysis, it is found that the main causes of defective products are ~ equipment parameter setting accounts for 42% to promote the development of rare earth industry in our city, and improper personnel operation accounts for 23%); With the help of statistical software (Minitab is widely used at present), the calculated values/indicators and visual graphics can be used as the "prescription" of improvement actions (for example, if the control chart shows that it is neither in the control state nor within the specification limit, the on-site personnel should give priority to improving the process and return to the control state, and then improve the specification limit problem)

what if there is no statistical basis

since statistics are pushing 6 σ Is so important, so what should most people who lack statistical foundation do? How much do you need to learn to be competent for GB, BB or champion? This problem is like driving a car. For general driving, as long as you are proficient in driving skills (for example, when overtaking, you must be able to see the distance, master the speed and steering wheel), you do not need to have a deep understanding of the compression ratio of the car or the structure of the gearbox. It is also reported that as long as GB personnel are familiar with the operation and application of statistical software (for example, when measuring system analysis, they must input the data of parts, operators and the number of repeated experiments, and select the correct function key), GB does not need to have deep statistical ability; As a project leader, BB must be able to distinguish whether the data collected by GB and the analysis tools selected are correct; As for champion, which formulates the company's policies, it is less involved in statistics

finally, we want to reiterate that statistics is a tool to effectively improve problems. six σ The essence of lies not in training statisticians, but in the ability to make good use of statistical tools to accurately describe and identify problems, and verify and infer (hypothesis testing, ANOVA, regression analysis, correlation analysis, etc.) the causal relationship of problems. Having a correct understanding of the application of statistics, the next step is to be proficient in the operation of statistical software, understand the applicable time, and determine the calculation results. Whether you are 6 σ The person in charge of the project or the members of the project must not be too specialized in statistical formulas and operations, which will trip you 6 σ Promotion of

(end)