2019

EXPERT ANSWERS

Data duo

Q: I would like to know an easy way to differentiate between continuous and discrete data types.

Priti Bhandari
Pune, India

A: The two basic data types in quality applications are discrete and continuous. Discrete data occur only as whole numbers. Continuous data can take on fractional values.

Examples of discrete data are the number of parts, shipments, defects or improvement projects. These are always reported as 0, 1, 2 or some other whole number. One-half of a project or 2.5 defects are not valid answers in this case.

Continuous data usually come from something measured, such as length, volume, pressure or cycle time. You can talk about lengths of 12.5 inches, volumes of 2.5 cubic feet or cycle times of 5.75 hours, and they will be accepted as valid answers.

In the case of discrete data, fractional values may come up in the course of an explanation, but the use is a simple expression of a more abstract concept. For example, if we say 3.5 projects are completed, we mean three are done, and a fourth one is partly done. If we say the average household in the county has 2.5 persons living in it, that may be a short way of saying county households most commonly have two or three people living in them.

Sometimes, due to the resolution of measuring instruments or customary reporting practices, continuous data may be shown as whole numbers only. If my ruler shows inches only as 1, 2, 3 and so on with no markings between numbers for fractions of an inch, I can report the length only to the nearest whole number. The odometer on my car shows tenths of a mile, but for purposes of filling out my expense report, I round the miles to the nearest whole number.

The resolution limits or reporting practices don’t change data from continuous to discrete. As long as it’s possible for the data to have valid fractional values, it’s still continuous.

There is a close relationship in statistics between discrete and continuous data. In some applications, continuous data can be used to approximate discrete data. For more information on that, consult a book on quality and statistics.

Joseph D. Conklin
Statistician
Washington, D.C.

For More Information

  • ReVelle, Jack B., "All About Data," Quality Progress, January 2006, p. 96.
     

Get IT under control

Q: I’m a junior high student, and after reading the article "How to Manage Risk In a Global Economy" (March 2008) my question is: What risks are managed by the IT manager in a global economy?

Angel Quinones

A: An effective IT process is important to the operation of any organization. Multiple software applications are needed to run an organization, including financials, sales, customer service and support, inventory management, enterprise resource planning and marketing. This also includes a subset of transaction systems, such as supply chain management, enterprise resource planning and business continuity planning.

Another reason IT is important is compliance with the Sarbanes-Oxley Act (SOX). This is because the correct and safe storage of data is critical to maintaining and analyzing data, and protecting important corporate records. In addition, SOX makes fragmented business management systems even more costly and problematic.

IT directly supports nine sections in the act. This is especially true of the support required to effectively satisfy Section 404, which requires an effective system of internal control and the associated sets of supporting controls.

Regarding information security and the management of the IT control system, the ISO 27000 series of standards describes the method and management system processes used to ensure an effective system of information security. In addition, the COBIT standard describes an effective tool for managing IT risk and controls.

It may also be helpful to review the control of electronic records and electronic signatures as outlined in the U.S. Food and Drug Administration Title 21 CFR Part 11.

Sandford Liebesman
President
Sandford Quality Consulting
Morristown, NJ

For More Information

You must measure

Q: Recently, we went through our AS9100C audit and received a minor corrective action report (CAR) for a situation in which we had monitors that showed processes in control but no plan for future measurement. We haven’t received any CARs or customer complaints about our design process. In fact, we’ve received many positive comments. Do we need to show measurement, or is continued monitoring enough?

Christopher Hetrick
Portland, CT

A: Your question states that you had monitors that showed processes in control. How can a process be monitored without measuring some aspects of a process? Therefore, it seems the auditee is measuring something, but instead of showing individual measurements, the auditee is simply reviewing those measurements and deciding the process is in control.

The fact that this was treated as a "minor corrective action" indicates to me that if the individual measurement results were shown to the auditors, then perhaps this minor corrective action could have been avoided.

The better practice you should use is to show the measurements, even when a process is in control. Showing individual measurements gives you a better idea about variation in the process, which can occur even if the process appears as though it’s in control.

For example, a process can be in control while most of the measurements are close to one of the control limits. Unless you look at the measurements, you would have no way of knowing this, and eventually the process could drift toward being out of control.

Pradip Mehta
Principal
Mehta Consulting LLC
Coppell, TX

For More Information

  • Rodríguez-Pérez, José, and Manuel E. Peña-Rodríguez, "Fail-Safe FMEA," Quality Progress, January 2012, pp. 30-36.

Average Rating

Rating

Out of 0 Ratings
Rate this article

Add Comments

View comments
Comments FAQ


Featured advertisers