Characteristics of Data – Central Tendency and Dispersion
February 12, 2025
Google is the most dominant search engine in the United States as well as in the world. It controls more than two-thirds of the readership in the United States. This means that if it were an actual newspaper, Google would have been open to antitrust violation lawsuits. Google has a greater stranglehold on the readership […]
The Green New Deal proposal announced by Alexandria Ocasio-Cortez suddenly became a talking point in American politics. The United States was visibly split into two camps, one was of people who are supporting the socialistic ideas being purported by the Democrat leader whereas others who oppose her plan. In a rather surprising development, large swathes […]
HR Metrics and Workforce analysis is a useful strategic tool for HR which shares information and evidences about the functioning of the entire system by relying on facts and figures, as opposed to mere assumptions or personal opinions. Over a period of time, the HR Metrics has evolved with the usage of more quantitative and […]
Inventory is a necessary evil that every organization would have to maintain for various purposes. Optimum inventory management is the goal of every inventory planner. Over inventory or under inventory both cause financial impact and health of the business as well as effect business opportunities. Inventory holding is resorted to by organizations as hedge against […]
What is the Hindu Rate of Growth and is the Indian Economy returning to it? For a long time from Independence to the mid 1980s, the Indian Economy used to grow at 3-4% or at the most, 5% a year. Indeed, in the decades following Independence, economists were fond of saying that India is not […]
Apart from accuracy and precision, there are more factors that determine the soundness of a measurement system. Two of these important factors are as follows:
Resolution: Resolution is the ability to see fine details in a system. This ability also gives the system, the ability to distinguish different readings from one another.
A good example would be that of the system that records late coming employees. If the measurement system designed to monitor employees coming late measures it only in hours i.e. uses the wrong resolution, then any employee who is up to 59 minutes late will appear to be on time. This is because the system will display 0 till the 60th minute is reached. In effect the organization will lose its ability to distinguish between employees coming on time and those that come late.
A good thumb rule is to build an extra decimal place in the system. If you need data up to 2 decimal places, build a system that records data up to 3 decimal places. Wrong resolution can seriously hamper the process and lead to wrong decisions.
Linearity: A system is said to be linear if proportional changes in input measurements produce proportional changes in output measurement. This means that if I know for a fact that an employee coming 10 minutes late would lead to a 2% loss of productivity in his daily output, then the same employee if he comes 30 minutes late should ideally lead to a 6% loss in his daily productivity.
The method used to find out linearity is fairly simple. Any given measurement is considered, the input variables are varied in a controlled manner and the resultant output is recorded. It is essential that at least 10 measurements be taken throughout the possible range of measurement. These points are then charted on a graph and an attempt is made to fit a line through these points. The degree to which all the points lie on this line of best fit is the degree of linearity of a system.
The degree to which the points lie away from the line of best fit is called the bias of the measurement system. For instance, if a machine shows a more accurate measurement at the centre of its operating range than at the upper or lower ends then the equipment is biased.
Common reasons for measurements not being linear are as follows:
Your email address will not be published. Required fields are marked *