Six Sigma Black Belt- Analyze

Front 1

Describe normal distributions.

Back 1

It is used for continuous data that have neither an upper nor a lower boundary. It looks like the classic bell curve.

Front 2

Describe exponential and lognormal distributions.

Back 2

Exponential is used for continuous data or data obtained by measurement. The most common application for this distribution is in the measurement of event rate, or the frequency with which a particular event occurs. Lognormal distributions are appropriate for continuous data that has a fixed lower boundary, usually zero, and no upper boundary. This distribution is used often for reliability data.

Front 3

Describe distributions, including their applications in the measure, analyze, improve, and control stages of DMAIC.

Back 3

Statistical distributions allow Six Sigma teams to make performance assumptions with a minimum of supporting data. The primary utility of these tools involves determining the characteristics, most notably the sigma level, of processes and resources.

Front 4

Describe the process cycle efficiency metric, including its applications in the analyze stage of DMAIC.

Back 4

The measurement of this is used to determine the most useful and positive ways to improve cycle time. The general method for calculating this is to divide value added time by process lead time.

Front 5

Describe binomial and Poisson distributions.

Back 5

Binomial distributions are useful when the units in a population exist in only two states. For instance, if the only possible characteristics of a population are "off" and "on", a binomial distribution can be used to estimate the total number of "off" and "ons' in a population. A Poisson distribution, on the other hand, can guess the number of times a particular condition will occur for a given process or population. It accurately estimates the number of events in each sample unit.

Front 6

Describe process decision program charts, including their applications in the analyze and improve stages of DMAIC.

Back 6

It can break a process down into its component tasks, with special emphasis on potential problems and solutions. As such, these charts are a useful tool for brainstorming protocols for handling crises that emerge.

Front 7

Describe the usage of velocity data.

Back 7

The calculation of velocity indicates the degree to which a process responds to customer demands. If there is less work in progress, lead times are shorter and velocity is greater. If the lead times are longer, then velocity is slower. When velocity is slow, the business cannot respond quickly to new orders from customers. The formula for process velocity (number of value-added steps divided by process lead time).

Front 8

Describe some strategies for reducing time spent on relocation and replacement during setup.

Back 8

One strategy is standardizing setups. A good strategy for reducing replacement tasks is simplifying the setup protocol, altogether. One way to simplify setups is by redesigning fixtures.

Front 9

Discuss the interpretation of hypothesis testing.

Back 9

If the calculated statistic is larger than the critical value of the test statistic for the given level of significance, then the null hypothesis must be rejected. If the critical value of the test statistic is larger, then the null hypothesis is not rejected.

Front 10

Describe Weibull and Johnson distributions.

Back 10

Weibull distributions are appropriate for continuous data with a set lower boundary, usually zero and no upper boundary. Weibull distribution often applies to reliability data, in which the interval between failures is recorded. Johnson distributions often are useful for data obtained after quality improvement campaigns, because the adjustments to process create non-results.

Front 11

Describe the process of hypothesis testing.

Back 11

The first step in hypothesis testing is stating the null hypothesis, Ho. In most cases, the null hypothesis represents the value that the test aspires to prove. The next step in hypothesis testing is defining the alternative hypothesis H1. The third step is either setting a value for p or selecting a significance level. The final step is drawing a conclusion.

Front 12

Describe the process of creating and interpreting an equality of variance test.

Back 12

The most common test for equality of variance is the Bartlett, the appropriate test when the regression residuals are expected to follow a normal distribution. In the Bartlett test, the equality of the treatment variances is compared with the possibility that one variance is unequal to the others.

Front 13

Describe the four primary parameters of a statistical distribution.

Back 13

The central tendency is the general trend indicated by the data. The mean is the best estimate of the central tendency. The skewness of a distribution is essentially the distance between the average and the mode, or the most represented data value. Standard deviation indicates the average variation of data points from the mean. Kurtosis is the sharpness of the distribution's peak.

Front 14

Discuss the interpretation of Pareto Charts.

Back 14

It attempt to isolate the categories that contribute the most to count or cost.

Front 15

Describe the "5 Whys" method of problem solving.

Back 15

This method is used to identify the underlying causes of a defect. The method is simple: the problem is stated, the questioner asks why this problem exists. For example, if the problem is that workers always are late, the questioner will ask, "Why are the workers always late?" An answer is given, and this statement is then questioned.

Front 16

Discuss the identification of NVA activities in the analyze stage of DMAIC.

Back 16

Six Sigma teams use tools such as matrix diagrams and quality function deployment efforts to identify non value added activities. As always, the question of value should be considered from the customer's perspective. If an activity does not create something for which a customer would be willing to pay, it likely does not add value.

Front 17

Describe the Six Sigma definition of setup time.

Back 17

Six Sigma defines setup time as the interval between the completion of the last item in the sequence and the beginning of the next item. Setup time consists of four components: preparation, replacement, location and adjustment.

Front 18

Outline the main criticisms of the "5 Whys" method of problem-solving.

Back 18

One complaint about this method asserts that it encourages investigators to examine and resolve only the symptoms of a problem rather than addressing the true cause. It does not require any further data gathering, so it tends to be limited by the investigator's existing expertise.

Front 19

Discuss value stream analysis.

Back 19

This is the entire series of activities that create value in a product or service. The addition value is considered from the perspective of the customer rather than from the perspective of employees or management within the company. The goal is to eliminate NVA.

Front 20

Describe the value of reducing movement and physical space in processes.

Back 20

This stage of DMAIC often reveals that much time is lost simply from one work site to another. Indeed, the most effective strategy for reducing cycle time often involves consolidating the area in which a task is performed. When possible employees can be cross-trained so they are capable of performing several different tasks within a complete process.

Front 21

Discuss the merits of process simplification.

Back 21

Six sigma involves simplifying needless complex processes. When the same process is used to create a wide variety of products and services, the process probably is too complicated. As much as possible businesses should try to limit the variation in deliverables so processes can be standardized.

Front 22

Describe the disadvantages of batching.

Back 22

Six Sigma has revealed that batching often increases overall cycle time. When process tasks are divided into batches, the time required to perform them clearly is shorter. However, batching creates delays at the beginning and end of the activity.

Front 23

Describe the lean metric velocity, including its applications in the analyze stage of DMAIC.

Back 23

It is a metric that indicates the rate at which value is added during a process phase. The calculation of the lean metric velocity is useful during the analyze stage. The basic formula for the calculation of velocity requires dividing the number of value added steps by the process lead time.

Front 24

Discuss the process of calculating process velocity.

Back 24

The first step is to categorize every task in the process as value added, non-value added but necessary or non-value added and unnecessary. Then measure the physical distance between the sites of each successive task. Next, the team will use a control chart to predict the average time required to complete each process task. Velocity is calculated by dividing the number of value added steps by the process lead time.

Front 25

Describe how control charts are used to identify sources of variation.

Back 25

One of the goals of analysis will be to distinguish special and common causes of variation. Special causes of variation do not occur during every performance of the process. On a control chart, special causes will be indicated by points that lie far outside normal range.

Front 26

Describe some strategies for reducing adjustment time during setup.

Back 26

The best way to limit adjustment is to establish good process controls. When a process can be repeated many times in a row in exactly the same way, it will require less adjustment.

Front 27

Describe some strategies for reducing preparation time.

Back 27

One of these strategies involves keeping all supplies and equipment as close as possible to the workstation, so employees do not waste time in transit. Another strategy is to group employees from different departments in work cells. The last strategy is leave equipment on and ready to go even when not in use.

Front 28

Describe autocorrelation charts and explain their applications in the measure, analyze, and control stages of DMAIC.

Back 28

These functions are used to determine the degree to which current data depends on previously gathered data. This is accomplished by automatically examining multiple observations of a particular characteristic with an eye toward possible correlations.

Front 29

Describe the process of using the autocorrelation function.

Back 29

The first step is testing for autocorrelations between each of the isolated observations. Each step will be considered in relation to the steps immediately before and after it.

Front 30

Discuss the interpretation of autocorrelation charts.

Back 30

When interpreting an autocorrelation chart, one should be aware of phenomena that might produce false correlation. For instance, sometimes autocorrelation will be significant only at adjacent data points, where the lag is very low. Another source of false correlation emerges with sampling from several different streams in a process.

Front 31

Describe cause and effect diagrams, including their applications in the analyze and improve stages of DMAIC.

Back 31

These diagrams, sometimes known as fishbone or Ishikawa diagrams are an easy and effective way to depict the reasons for a particular event.

Front 32

Describe the process of using the partial autocorrelation function.

Back 32

The significance limits for both the autocorrelation function and the partial autocorrelation function (if the true population of the ACF or PACF is zero) is calculated at the stated significance level.

Front 33

Describe the process of creating and interpreting a goodness of fit test.

Back 33

Most experts agree that a distributional fit cannot be tested properly without at least two hundred data points. However, if a process is not in statistical control, no distribution will fit its data points. Goodness of fit tests have a null hypothesis that the data follows the distribution in question. The measure of this fit is known as the p value. Typically a larger p value fits better.

Front 34

Describe interaction plots, including their applications in the analyze and improve stages of DMAIC.

Back 34

These plots illustrate the interrelationships of three parameters. In most cases, these parameters are two factors and one response. If the plot variables exhibit no interaction, then the lines basically will be parallel. That is, both plot variables will produce similar trends when combined with the response variable.

Front 35

Describe mult-vari plots, including their applications in the analyze stage of DMAIC.

Back 35

These plots are effective tools for assessing the variation within samples or within particular parts. They are also used to analyze variation over time between different batches. During analyze they are used to isolate the causes of variation and to obtain more information about the interactions among factors. However, this tool is not a control chart.

Front 36

Describe nominal group technique, including its applications in the define and analyze stages of DMAIC.

Back 36

This is a system for ranking non-objective data and is used primarily to create consensus or agreement in groups. In the define stage, it is used to simplify the project loads by combining redundant projects and eliminating unnecessary projects. In the analyze stage, nominal group technique helps teams agree on which solutions should be pursued.

Front 37

Describe the process of creating a Pareto chart.

Back 37

The first step in creating a Pareto chart involves identifying the correct categories and they should be non-overlapping. The count/cost metric will be along the left axis of the chart. The categories will be placed in descending order from left to right. An indication of the percent demarcations, or the percent of total cost/count represented by each variable, will run along the right vertical axis.

Front 38

Describe the primary objectives of the analyze stage.

Back 38

During this stage the team focuses on analyzing the sources of variation in the target process. This may require use of sophisticated statistical tools. The team also will analyze the value stream which includes those activities necessary for the creation of the product or service.

Front 39

Discuss the creation and interpretation of cause-and-effect diagrams.

Back 39

The first step in creating a cause and effect diagram is making a provisional list of the possible relationships between the process and the outcome. For some problems, beginning with either the four Ps (people, plants, policy, and procedures) or the 5Ms and E (measurement, material, methods, machines, manpower, and environment) will be appropriate. Cause and effect diagrams are primarily brainstorming tools, not final assessments.

Front 40

Describe nonparametric tests on equality of means, including their applications in the analyze and improve stages of DMAIC.

Back 40

These tests occasionally are used in place of traditional hypothesis tests for the equality of two means. In the analyze stage, nonparametric tests are used to compare the means from samples with different conditions. In the improve stage, they are used to assess whether process averages have been improved over baseline estimates after the implementation of changes.

Front 41

Describe equality-of-variance tests, including their applications in the analyze and improve stages of DMAIC.

Back 41

These tests determine whether a similar degree of variation exists in particular subsets of data. This question is important because an effective analysis of variance cannot be performed unless equal variance has been established. The most typical statistical test used to measure equality of variance is the Bartlett, although versions also are offered on Levene and Minitab.

Front 42

Describe hypothesis testing, including its applications in both the analyze and the improve stages of DMAIC.

Back 42

This testing establishes a degree of confidence and then compares a sample statistic against a historical value or another sample statistic. In other words, hypothesis testing allows one to make statistical inferences about the characteristic of a population.

Front 43

Describe the importance of identifying key decision points during the measure stage.

Back 43

Once the process map has been created, it can be used to identify the most important decisions made during the process. Team members will especially be alert to any areas that seem to require excessive decision making, usually a sign of inefficiency.

Front 44

Describe process definition.

Back 44

The first major step in the measure stage is to create a comprehensive process level map of processes as currently performed. In other words, during this stage the group defines and clearly describes all of the activities they aim to improve.

Front 45

Describe the basic objectives of the measure stage of DMAIC.

Back 45

During this stage of DMAIC, the Six Sigma team will focus on gathering the information necessary to complete the project. No metric is complete until accomplished by a measurement analysis system that identifies and quantifies any common errors in the metric.

Front 46

Describe some types of matrix diagram.

Back 46

In a directional system, arrows suggest the nature of the relationship between each pair of items. In a numerical matrix diagram, numbers perform the same functions as weighted arrows. In a plus-minus system, a plus sign indicates a relationship and a minus sign indicates the absence of a relationship. In a symbol system, finally, triangles, circles, and squares indicate characteristics of each relationship.

Front 47

Describe the process of creating and interpreting nonparametric tests on equality of means.

Back 47

For these tests on the equality of means, the null hypothesis Ho will be that population 1's median is equal to population 2's median. The alternative hypothesis H1, then, is that population 1's median does not equal population 2's median. Median is preferable to mean in these tests because it indicates central tendency regardless of distribution. The next step in the test involves declaring a significance level or p value. The null hypothesis is rejected if the calculated statistic is greater than the critical value of the test statistic.

Front 48

Describe the Kruskal-Wallis and Mood's median tests.

Back 48

These median tests are nonparametric estimation methods. In a Kruskal-Wallis test, the null hypothesis asserts that all medians are not all equal.

Front 49

Describe some scenarios in which batching may be preferable.

Back 49

For instance, if the time required to set up the equipment for a particular activity is significant, performing the activity in batches may be more efficient. An example is making soup from scratch.

Front 50

Discuss the elimination of NVA activities in the analyze stage of DMAIC.

Back 50

These activities create nothing for which a customer would be willing to pay.

Front 51

Describe Pareto charts, including their applications in the define and analyze stages of DMAIC.

Back 51

This chart is a form of bar graph in which problems are ranked according to their urgency.

Front 52

Discuss the interpretation of histograms.

Back 52

These effectively pinpoint the process location and variation. It also indicates when data is symmetrical or bounded. A symmetrical distribution is characterized by data spread evenly about it's center in an arrangement similar to a bell curve. It only indicates the performance of the process as it is being measured.

Front 53

Describe histograms, including their applications in the measure and analyze stages of DMAIC.

Back 53

It is a tool for presenting data pictorially. It looks like a standard bar graph, except that each bar represents the total number of observations that lie within a range of ranked values. It is created by first placing all data in order from least to greatest.

Front 54

Describe U charts, including their applications in the measure and improve stages of DMAIC.

Back 54

It is a control chart designed to handle attributes data. It depicts the percentage of samples that have a particular condition in situations where sample sizes may vary and each sample may have more than one occurrence of the condition.

Front 55

Describe scatter diagrams, including their applications in the analyze stage of DMAIC.

Back 55

It is a simple plot on two axes, useful for investigating the correlation between two variables. The x-axis of a scatter diagram measures the independent variable (is the variable manipulated in the experiment) and the y-axis measures the dependent variable (the variable not manipulated in the experiment).

Front 56

Describe spaghetti diagrams, including their applications in the analyze stage of DMAIC.

Back 56

These diagrams depict the movement of resources, materials, or personnel throughout an organization. They are used during the analyze stage of DMAIC. These kinds of diagrams isolate wasteful movement, whether in regard to people or resources.

Front 57

Discuss the interpretation of U charts.

Back 57

On this chart, the upper and lower control limits indicate the boundaries of expected process behavior. The variation of points that lie within the control limits is attributed to common causes, while any points outside the statistical control must be attributed to special causes.

Front 58

Describe goodness of fit tests, including their applications in the measure, analyze, improve, and control stages of DMAIC.

Back 58

The number of useful goodness of fit tests measure the validity of statistical assessment. In other words, these tests indicate whether a chosen statistical test provide an accurate and relevant measure. Some of the most common goodness of fit tests are the chi-square, Anderson-Darling, and Kolmogorov-Smirnov (K-S).