Characteristics of Data – Central Tendency and Dispersion
February 12, 2025
Introduction In the modern global and competitive business world, it is very important to the companies to enjoy financial success. This financial success can be achieved through financial discipline, goal setting and periodic reviews. Companies which have enjoyed financial success are likely to see that trust among employees have increased; profit has improved, and employee […]
Project Management is the art and science of managing projects with emphasis on quality, cost and time. If we take the aspect of quality, this is one of the factors of the “triple constraint” that govern the art of project management. Quality is defined as the degree to which the project meets the requirements (PMBOK, […]
Gone are the days when outsourcing meant a steady stream of jobs being shipped to India. Now the trend is for Indian companies to create jobs in the West. This phenomenon which is gathering pace in recent times is known as reverse outsourcing or the trend of Indian IT and BPO companies opening offices in […]
Six Sigma is a process oriented methodology designed to improve business performance by improving specific areas of strategic business processes. There are 2 different methodologies available for carrying out improvements in processes or operations. Improvements can be of two types: improving the existing process or designing a new process altogether. When we have an existing […]
There are several models of behaviour that the project manager can draw upon in his or her work. These include Maslow’s need hierarchy theory, Herzberg’s Hygiene theory and McGregor’s theory X and theory Y as applied to project management. All these behavioural models point to the ability of the project manager to motivate the people […]
Measurement Systems Analysis is a complicated exercise. However, Six Sigma provides a step by step procedure to conduct it. Also, as usual, the focus of the executives should be to understand the focus of the exercise and interpretation of the results. The complex calculations can be performed by software. Here is the 4 step procedure.
Before any study needs to be conducted, there must be a proper plan in place. This ensures that the study can be conducted with minimal wastage of time, money, efforts and other resources. Measurement System Analysis is no exception to this rule. Here is how to create a basic plan for measurement system analysis:
Confirm the Key Measures: Any process consists of a large number of measures. The sum total of these measures is called the measurement system. Just like there are many inputs, but only few of them are vital and have any meaningful impact on the outputs, similar is the case with measures. The same approach i.e. to conduct a Pareto analysis must be followed and key measures must be developed. Usually, the key measures correspond to the critical inputs or outputs already identified in the Define stage.
Develop Operational Definition: The next step must be to come up with exact operational definitions for these measures. Since an entire analysis is going to be conducted on these measures, they must be standardized. The operational definition helps eliminate any ambiguity.
There are a few decisions which need to be made during the course of the analysis. They are as follows:
Determine Number of Measurement Trials: The optimum number of trials that are enough to decide whether a measure is appropriate or not must be decided. This can be considered analogous to the sample size that is used while conducting experiments.
Determine Organization of Trials: Trials can be collected at the same time or at different times of the day. The frequency, with which the trials are collected, plays a vital role in the analysis. Therefore it must be carefully decided.
Different Operators: The study must be designed to negate the influence of a specific operator. Hence data must be collected from different people.
Different Equipment: The study must also negate the influence of specific machines. For instance, if a machine is newer than the other, it may function better as compared to the other. Such influences must be negated to reach the true measurement variation.
Different Conditions: The study must be planned in such a way that trials must be conducted in as varied circumstances as possible. The whole point of this is to negate the influence of any specific factor and bring to the forefront the true measurement error.
Once the study has been conducted, the next step is to analyse the result to see if they are as per your requirements. Whether the measurement system is good enough is a question that depends upon the intended usage of the measurements. In case the measurements are going to be used for precision engineering, they need to be highly accurate.
However, some thumb rules have been established. They work, if the process does not require special attention. They are as follows:
Less than 10%: if the measurements are less than 10% different from the actual process variation, the Six Sigma methodology specifies that the measurement system is good enough. There is an assumption that the process is not highly critical like in medical sciences or precision engineering where the accuracy needs to be much higher.
Between 10% to 30%: If the measurements are off-base between 10% to 30% of the actual process variation, they may or may not be accepted, depending on the degree of accuracy required by the team.
Greater than 30%: If the measurements are more than 30% off base, then there is a serious problem with the measurement system and steps must be taken to ensure that time and resources are not wasted making decisions based on faulty data.
If the measurement system fits the bill, the process ends here. If not, an additional step is required i.e. the measurement system must be fixed.
It is difficult to state a generalized way to fix measurement system errors. However, since we know that most of the process variation is caused by a handful of factors, we can try fixing them to fix the measurement system.
Change Equipment: Most of the times, wrong measurements are the result of faulty systems. This means that if people are using manual means of collecting data, the organization must consider automating it. It could also mean that the measurement systems are old and faulty and need to be replaced. Depending upon the need of the management and the budget it is willing to expend, the measurement system can be fixed.
Train Operators: In many cases, the measurements are taken manually. This could be either because of the nature of the process. Alternatively it could be because of the management’s reluctance to invest in automation. In this case, the operators must be trained to cut down on the error.
Further Analysis: If results reveal that neither the operators nor the machine are responsible for the dismal performance of the measurement systems, then in that case further analysis must be deployed till the problem is found and fixed.
An iteration of the previous step must then be conducted to close the exercise.
Your email address will not be published. Required fields are marked *