# Free Six Sigma Green Belt Overview

Six Sigma has adopted the various belt levels of Karate; the Six Sigma Green Belt certification is designed to qualify the candidate to work in support of and under the supervision of a Six Sigma Black Belt. The Six Sigma Green Belt holder is involved in the implementation of all aspects of the Six Sigma methodology, but the requirements for certification are more fundamental in outlook. A Green Belt is required to have at least three years of work experience and successful completion of the certification test.

As with the Black Belt certification, Green Belt candidates must be adept in the conceptual approach of Bloom's Taxonomy: remember, understand, apply, analyze, evaluate and create. The first part of the certification test requires knowledge of general Six Sigma goals and organizational components. The student must recognize the value of Six Sigma to the business enterprise, recognize the key profit drivers, and be able to define fundamental manufacturing concepts like value chain, theory of constraints, cycle time reduction, push and pull lines and failure effects measurement.

Twenty-five subsequent questions focus upon process management basics, including the collection and analysis of customer data and the identification of the stakeholders in a process. Failure mode and effects analysis; team functioning and dynamics; and the communication tools used to manage teams are vital to success in this area.

Statistics are important on the Green Belt certification exam; there are 30 questions devoted to data measurement and other statistical collection activities. The test poses questions on probability and the use of reliable statistics in decision-making. Process capability, results, and the degree to which statistical data deviated from the mean are also areas covered in this portion of the exam.

Analysis of studies and hypothesis testing methods comprises another 15 questions. The student must be able to interpret variables in data collection and create studies to interpret the differences between positional, cyclical, and temporary variations. There are 15 questions in the area of process improvement and control. Knowledge of basic terminology (e.g., dependent and independent variable, replication and repetition, and error deviation) is required. The student should be proficient in the techniques of brainstorming, chief effect analysis, multi-variant studies, FMEA (failure mode effects analysis), and the measurement of system capabilities.

## Free Six Sigma Green Belt Test Questions

**
1. Which distribution is appropriate for a continuous set of data with a fixed lower boundary but no upper boundary?
**

a. Johnson

b. exponential

c. normal

d. Logormal

**
2. Which of the following is a disadvantage of using engineering process control devices to prevent deviation?
**

a. The devices must be monitored by human operators.

b. The use of these devices precludes the use of statistical process controls.

c. These devices require constant maintenance.

d. These devices cannot handle multiple inputs.

**
3. Which of the following is a disadvantage of higher-order multiple regression models?
**

a. These models do a poor job of defining the area around a stationary point.

b. Comprehensive and detailed experiments must be performed on the main effects.

c. These models rarely have clear peaks and valleys.

d. Small regions are difficult to perceive.

**
4. In hypothesis testing, why is it better to set a p value than to select a significance level?
**

a. It ensures that a true hypothesis will not be rejected.

b. It is then easier to make adjustments later in the experiment.

c. It enables the collection of more samples.

d. It makes it possible to reject the null hypothesis.

**
5. Which statistical distribution is appropriate for continuous data with neither an upper nor a lower boundary?
**

a. Lognormal

b. Weibull

c. Exponential

d. Normal

## Answers & Explanations

**
1. D:** Lognormal. A lognormal distribution is appropriate for a continuous set of data with a fixed lower boundary but no upper boundary. In most cases, the lower boundary on a lognormal distribution is zero. These distributions can be tested with a goodness-of-fit test. A Johnson distribution is more appropriate for continuous data that for whatever reason is inappropriate for a normal or exponential distribution. An exponential distribution is appropriate for any set of continuous data, though these distributions are most often used for frequency data. A normal distribution is appropriate for a set of continuous data with neither an upper nor a lower boundary. The normal distribution follows the pattern of the classic bell curve.

**
2. B:** The use of these devices precludes the use of statistical process controls. One disadvantage of using engineering process controls to prevent deviation is that the use of these devices precludes the use of statistical process controls. An engineering process control is a mechanism that automatically adjusts inputs when it detects variations in the process. A thermostat is a basic example of an engineering process control. It is not necessary for these devices to be monitored by human operators, and in most cases engineering process controls do not require constant maintenance. The constant adjustments made by these devices, however, mean that any data related to their activities is not independent, and therefore cannot be analyzed with statistical process control charts. However, the engineering process controls used by heavy industry are capable of handling a number of different inputs and outputs simultaneously.

**
3. B:** Comprehensive and detailed experiments must be performed on the main effects. One disadvantage of higher-order multiple regression models is that comprehensive and detailed experiments must be performed on the main effects. Otherwise, it will not be wise to assume that the results of the higher-order multiple regression models are useful or accurate. However, higher-order multiple regression models have a number of advantages. For one thing, they are excellent at clearly defining the area around a stationary point. They typically have well-defined peaks and valleys, which facilitates analysis. Also, they are very effective at mapping small regions in the process, so they are able to achieve a high level of precision and detail.

**
4. B:** It is then easier to make adjustments later in the experiment. In hypothesis testing, it's better to set a p value than to select a significance level because it is then easier to make adjustments later in the experiment. In general, a p value allows for more freedom in the later parts of the experiment. There is always a possibility of rejecting a true hypothesis, in what is known as a Type 1 error. The number of samples collected is not dependent on whether a p value is set or a significance level is selected, and either method maintains the possibility that the null hypothesis will be rejected.

**
5. D:** Normal. A normal distribution is appropriate for continuous data with neither an upper nor a lower boundary. Continuous data is obtained through measurement. A lognormal or Weibull distribution is appropriate for sets of continuous data with a fixed lower boundary but no upper boundary. In most lognormal and Weibull distributions, the lower boundary is zero. An exponential distribution is appropriate for continuous data sets in which the values are relatively consistent.

Last Updated: 04/17/2017