View stunning SlideShares in full-screen with the new iOS app!Introducing SlideShare for AndroidExplore all your favorite topics in the SlideShare appGet the SlideShare app to Save for Later — even offline
View stunning SlideShares in full-screen with the new Android app!View stunning SlideShares in full-screen with the new iOS app!
Benchmarking An improvement process whereby a company measures its performance against that of best-in-class companies, determines how those companies achieved their performance levels, and uses the information to improve its own performance. Black Belt Full-time Six Sigma project leader who is certified following a four-month training and application program and successful completion of two Six Sigma Projects, the first under the guidance of a Master Black Belt, the second more autonomously. “ Breakthrough Strategy” The data driven, Six Sigma process improvement strategy involving four phases: Measure, Analyze, Improve and Control. Cause That which produces an effect or brings about change. Cause-And-Effect Diagram A schematic sketch, usually resembling a fishbone, which illustrates the main causes and subcauses leading to an effect (symptom). Also known as Fishbone Diagram. Champion Member of the senior Aircraft Engines staff who has undergone extensive Six Sigma training. Champions provide direction, resources and support to the Six Sigma effort and approve and review projects. Characteristic A definable or measurable feature of a process, product or variable. Control Chart A graphical rendition of a characteristic’s performance across time in relation to its natural limits and central tendency. Correlation The determination of the effect of one variable upon another in a dependent situation. Cp A widely used capability index for process capability studies. It may range in value from zero to infinity with a larger value indicating a more capable process. Six Sigma represents Cp of 2.0. Page 1
Cpk An index combining Cp and K (Difference between the process mean and the specification mean) to determine whether the process will produce units within tolerance. Cpk is always less than or equal to Cp. When the process is centered at nominal, Cpk is equal to Cp. Critical To Quality (CTQ) An element of a design or a characteristic of a part that is essential to quality in the eyes of the customer, formerly known as a key quality characteristic (KQC). Data Factual information used as a basis for reasoning, discussion or calculation; often refers to quantitative information. Defect A failure to meet an imposed requirement on a single quality characteristic or a single instance of nonconformance to the specification. Defects Per Million Opportunities (DPMO) The number of defects counted, divided by the actual number of opportunities to make a defect, then multiplied by one million. A direct measure of sigma level. Defects Per Unit (DPU) The number of defects counted, divided by the number of products or characteristics produced. A process of counting and reducing defects as an initial step toward Six Sigma quality. Defective A unit of product containing one or more defects. Design For Manufacturability (DFM) A concept in which products are designed within the current manufacturing process capability to ensure that engineering requirements are met during production. Design of Experiments (DOE) Statistical experimental designs to economically improve product and process quality. A major tool used during the “Improve Phase” of Six Sigma methodology. Distributions Tendency of large numbers of observations to group themselves around some central value with a certain amount of variation or “scatter” on either side. Page 2
Effect That which was produced by a cause. Experiment A test under defined conditions to determine an unknown effect; to illustrate or verify a known law; to test or establish a hypothesis. Experimental Error A test under defined conditions to determine an unknown effect; to illustrate or verify a known law; to test or establish a hypothesis. “ Factory” Processes For Six Sigma purposes, defined as design, manufacturing, assembly or test processes which directly impact hardware (see also transaction processes). Fishbone Diagram A schematic sketch, usually resembling a fishbone, which illustrates the main causes and subcauses leading to an effect (symptom). Also known as Cause-And-Effect Diagram. Failure Mode Effects Analysis (FMEA) A process in which each potential failure mode in every sub-item of an item is analyzed to determine its effect on other sub-items and on the required function of the item. “ Five Ms” Major sources of variation: manpower, machine, method, material and measurement. Additionally, “environment” is considered to be a source of variation. Frequency Distribution The pattern or shape formed by the group of measurements in a distribution. Gage Repeatability & Reproducibility (Gage R&R) A measurement system evaluation to determine equipment variation and appraiser variation. This study is critical to ensure that the collected data is accurate. Histogram Vertical display of a population distribution in terms of frequencies; a formal method of plotting a frequency distribution. Independent Variable A controlled variable; a variable whose value is independent of the value of another variable. Page 3
Interaction When the effects of a factor A are not the same at all levels of another factor B. Lower Control Limit A horizontal dotted line plotted on a control chart which represents the lower process limit capabilities of a process. Master Black Belt An expert in quality techniques specially trained to advise leaders, facilitate quality teams and accelerate process improvement. Master Black Belts select, train and mentor Black Belts; develop and implement the Six Sigma deployment plan; and select and ensure completion of Six Sigma projects. Nonconformity A condition within a unit which does not conform to some specification, standard, and/or requirement; often referred to as a defect; any given nonconforming unit can have the potential for more than one nonconformity. Normal Distribution A continuous symmetrical density function characterized by a bell-shaped curve, e.g., distribution of sampling averages. Pareto Diagram A chart which ranks, or places in order, common occurrences. Primary Control Variables The major independent variables used in the experiment. Probability The chance of something happening; the percent or number of occurrences over a large number of trails. Process A particular method of doing something, generally involving a number of steps or operations. Process Capability The relative ability of any process to produce consistent results centered on a desired target value when measured over time. Page 4
Process Control Chart Any of a number of various types of graphs upon which data are plotted against specific control limits. Process Map Flow chart to analyze a process by breaking it down into its component steps, and then gaining a better understanding of the process, step-by-step. Process Spread The range of values which a given process characteristic displays; this particular term most often applies to the range but may also encompass the variance. The spread may be based on a set of data collected at a specific point in time or may reflect the variability across a given amount of time. Quality Functional Deployment (QFD) Structured methodology to identify and translate customer needs and wants into technical requirements and measurable features and characteristic. This tool is used to identify Critical to Quality Characteristics (CTQCs). Random Selecting a sample so each item in the population has an equal chance of being selected; lack of predictability; without pattern. Random Cause A source of variation which is random; a change in the source (“trivial many” variables) will not produce a highly predictable change in the response (dependent variable), e.g., a correlation does not exist; any individual source of variation results in a small amount of variation in the response; cannot be economically eliminated from a process; an inherent natural source of variation. Random Variation Variations in data which result from causes which cannot be pinpointed or controlled. Regression Analysis A statistical technique for determining the relationship between one response and one or more independent variables. Robust The condition or state in which a response parameter exhibits hermetically to external cause of a nonrandom nature; e.g., impervious to perturbing influence. Page 5
Rolled Yield The combined resulting quality level, stated as a percent acceptable, that occurs when several processes known to produce defects at some rate are combined to produce a product. For example, a product that requires 100 steps, each of which produces a yield of 98.78% will produce a rolled yield of 0%, that is, no acceptable products. Scatter Diagram A diagram that displays the relationships between two variables. Sigma Standard deviation; an empirical measure based on the analysis of random variation in a standard distribution of values; a uniform distance from the mean or average value such that 68.26% of all values are within 1 sigma on either side of the mean, 95.44% are within 2 sigma, 99.73% are within 3 sigma, 99.9% are within 4 sigma and so forth. Sigma Level A statistical estimate of the number of defects that any process will produce equivalent to defects per million opportunities for that process. Six Sigma Quality A combination of verified customer requirements reflected in robust designs and matched to the capability of production processes that creates products with fewer then 3.4 defects per million opportunities to make a defect. World-class quality. A collection of tools and techniques for raising quality to worked-class levels. Stable Process A process which i free of assignable causes, e.g., in statistical control. Standard Deviation A statistical index of variability which describes the spread. Statistical Control A quantitative condition which describes a process that is free of assignable/special causes of variation, e.g., variation in the central tendency and variance. Such a condition is most often evidenced on a control chart. Statistical Process Control The application of statistical methods and procedures relative to a process and a given set of standards. Page 6
Transaction Processes For Six Sigma purposes, defined as any business process that contributes to customer satisfaction or impacts operating efficiency and which is designated by a vice president or by GE Corporate as a focus for process improvement. Such efforts will be led by the process owner, with teams being led by specially trained transaction project leaders and/or by certified Black Belts. Transaction Project Leader An individual designated to lead a transaction process improvement project. Transaction project leaders attend a four-day course in specific Six Sigma tools and tactics. Upper Control Limit A horizontal line on a control chart (usually dotted) which represents the upper limits of process capability. Variable A characteristic that may take on different values. Variables Data A numerical measurement made at the interval or ratio level; quantitative data, e.g.., ohms, voltage, diameter; subdivisions, of the measurement scale are conceptually meaningful, e.g.., 1.6478 volts. Variation Any quantifiable difference between individual measurements; such differences can be classified as being due to common causes (random) or special causes (assignable). “ Xs” Designation in Six Sigma terminology for those variables which are independent, root causes; as opposed to “Ys” which are dependent outputs of a process. Six Sigma focuses on measuring and improving Xs, to see subsequent improvement in Ys. X & R Charts A control chart which is a representation of process capability over time; displays the variability in the process average and range across time. “ Ys” Designation in Six Sigma terminology for those variables which are dependent outputs of a process, as opposed to “Xs” which are independent root causes. Page 7
6M’s - Man, Machines, Materials, Methods, Measurement, Mother Nature ANOVA - Analysis of Variance BB - Black Belts C&E Matrix - Cause & Effect Matrix CAP - Change Acceleration Process C&E - Cause & Effect COPQ - Cost of Poor Quality COQ - Cost of Quality Cp - Capability Process Index (Ideal) - Pooled Cpk - Capability Process Index (Real) - Pooled CTQ - Critical to Quality CUSUM - Cumulative Sum DF - Degrees of Freedom DFM - Design for Manufacturing DFSS - Design for Six Sigma DOE - Design of Experiments DPM - Defects per Million DPMO - Defects per Million Opportunities DPO - Defects per Opportunities DPU - Defects per Unit EVOP - Evolutionary Operation EWMA - Exponential Weight Moving Average FMEA - Failure Mode & Effect Analysis GAGEAOV - Gage Analysis of Variance GRR - Gage Repeatability & Reproducibility IDOV - Identify, Design, Optimize, Validate IQR - Inter Quartile Range ISO - International Organization for Standardization KNP - Key Noise Parameters KPI (Factors) - Key Process Inputs KPIV (KCP) - Key Process Input Variable (Key Control Parameter) KPOV or - Key Process Output Variable(Response) LCL - Lower Controls Limits LSL - Lower Specification Limits MAIC - Measurement, Analysis, Improvement, Control MBB - Master Black Belt
MBNQA - Malcolm Baldrich National Quality Award MGF - Minitab Graph File MSA - Measurement System Analysis MTB - Minitab MTW - Minitab Worksheet NPI - New Product Introduction OJT - On the Job Training P(ND) - Probability (Not Defective) PEAR - Process, Engineering, Application, Regulatory CTQ’s Pp - Capability Process Index (Ideal) - Overall Ppk - Capability Process Index (Real) - Overall PPM - Parts per Million QA - Quality Assurance QFD - Quality Functional Deployment P/T Ratio - Precision / Tolerance Ratio ROI - Return of Investment RPN - Risk Priority Number RSM - Response Surface Methodology RTY - Rolled Throughput Yield SOP - Standard Operating Procedure SOV - Source of Variation SPC - Statistical Process Control SQC - Statistical Quality Control T - Target TCS - Total Customer Satisfaction TOP - Total Opportunities TQL - Total Quality Leadership TQM - Total Quality Management UCL - Upper Control Limits USL - Upper Specification Limits WIP - Work in Process XLS - Excel Spreadsheet Zlt - Z-long term ZST - Z-short term
= Summation; i.e., 1 + 2 + 3 + 4 + 5 = 15 ! = Factorial; i.e. 5! = 5 x 4 x 3 x 2 x 1 = 120 e = Natural constant = 2.7183 g = Total number of subgroups. i = The i th element in a string of 1, 2, 3, 4, -- i j = The j th element in a string of 1, 2, 3, 4, -- j n = Subgroup size (for high volume production, the range for n would normally be between 3 and 10. R = Range = difference (subtraction) between the maximum and minimum measurements observed/recorded for a subgroup R = Average of subgroup ranges = R; g S = Standard deviation = X = A variable measurement made on an individual characteristic and on an individual unit (often a process output variable) recorded onto a data log or control chart. Note : X is also used in another sense to denote the variables that cause process variation. X = Average of the X observations associated with a subgroup of size n X = Average of observations over all subgroups = X n g LT = Standard deviation of the total population over a long period of time. / j = 1 g X / n i = 1 i = 1 j = 1 i i / 971162A-bw Glossary #2 n X i
LT = Estimate of long-term standard deviation = = Standard deviation of an individual subgroup = ST = Standard deviation of a population over a short period of time ST = Estimate for short-term standard deviation ST ; ST = W = Pooled standard deviation = 2 = Variance u = Process average or mean = X u = Subgroup average or mean = X Y = A process output variable - may likely be a CTQ Y RT = Rolled thruput yield Cp = Short term process capability assuming no shift. Cp = 3 X Z ST Cpk = Short term process capability including mean shift occurring in the process. Z ST = Number of short-term standard deviations ( ST ) that fit between the specification center and the specification limit (in either direction) j < < < i n i = 1 (X - X) 2 i N - 1 < < (X - X) 2 i j g j = 1 i = 1 n g - 1 < < (X - X ) 2 i j g j = 1 i = 1 g ( n - 1) n < + –+ = ST 2 2 2
Z LT = Number of long-term deviations ( LT ) that fit between the observed process average (X) and the closest specification limit.
Z LT =
2 = CHI square distribution - Used for hypothesis testing as follows:
• Test for independence (used to test for independent relationship between two discrete variables)
• Goodness of fit (used to determine if the data fits an assured distribution) • Establishing the confidence interval for standard deviation
F = F distribution - associated with hypothesis testing of standard deviation between two or more process distributions.
T = T distribution - associated with hypothesis testing of the means (averages) between two distributions (when sample sizes are less than 100).
X (1 SL -X) LT
ABSCISSA The horizontal axis of a graph. ACCEPTANCE REGION The region of values for which the null hypothesis is accepted. ALPHA RISK The probability of accepting the alternate hypothesis when, in reality, the null hypothesis is true. ALTERNATE HYPOTHESIS A tentative explanation which indicates that an event does not follow a chance distribution; a contrast to the null hypothesis. ANALYSIS OF VARIANCE A statistical method for evaluating the effect that factors (ANOVA) have on process mean and for evaluating the differences between the means of two or more normal distributions. ASSIGNABLE CAUSE A process input variable that can be identified and that contributes in an observable manner to non-random shifts in process mean and /or standard deviation. ASSIGNABLE VARIATIONS Variations in data which can be attributed to specific causes. ATTRIBUTE DATA Quality data that typically reflects the number of conforming or non-conforming units or the number of non- conformities per unit on a go/no go or accept/ reject basis. AVERAGE Sum of all measurements divided by the total number of measurements. Statistic which is used to estimate the population mean. Same as MEAN.
BACKGROUND VARIABLES Variables which are of no experimental interest and are not held constant. Their effects are often assumed insignificant or negligible, or they are randomized to ensure that contamination of the primary response does not occur. Also referred to as environmental variables and uncontrolled variables. BENCHMARKING A process for identification of external best-in-class practices and standards for comparison against internal practices. BETA RISK The probability of accepting the null hypothesis when, in reality, the alternate hypothesis is true. BINOMIAL DISTRIBUTION A statistical distribution associated with data that is one of two possible states such as Go-No Go or Pass-Fail. It is also the distribution generated by rolling dice. BLACK BELT A process improvement project team leader who is trained and certified in Six Sigma methodology and tools and who is responsible for successful project execution. BLOCKING VARIABLES A relatively homogenous set of conditions within which different conditions of the primary variables are compared. Used to ensure that background variables do not contaminate the evaluation of primary variables. BRAINSTORMING A team-oriented meeting used in problem solving to develop a list of possible causes that may be linked to an observed effect. CAPABILITY INDICES A mathematical calculation used to compare the process variation to a specification. Examples are Cp, Cpk, Pp, PpK, Zst, and Zlt. GE uses Zst & Zlt as the common communication language on process capability. CAUSALITY The principle that every change implies the operation of a cause. CAUSATIVE Effective as a cause. CAUSE That which produces an effect or brings about a change.
CAUSE AND EFFECT (C&E) One of the seven basic tools for problem solving and is DIAGRAM sometimes referred to as a “fishbone” diagram because of its structure. Spine represents the “effect” and the major legs of the structure are the “cause categories.” The substructure represents the list of potential causes which can induce the “effect.” The 6M’s (man, machine, material, method, measurements and mother nature, are sometimes used as cause categories. C CHARTS Charts which display the number of defects per sample. Used where sample size is constant. CENTER LINE The line on a statistical process control chart which represents the characteristic’s central tendency. CENTRAL TENDENCY Numerical average, e.g., mean, median, and mode; center line on a statistical process control chart. CHAMPION An executive level business leader who facilitates the leadership, implementation, and deployment of Six Sigma philosophies. CHANGE ACCELERATION A process which helps accelerate stakeholder buy-in and PROGRAM PROGRAM (CAP) implementation of new philosophies and processes within a business. CHARACTERISTIC A definable or measurable feature of a process, product, or service. CHI-SQUARE See x (symbol glossary). CLASSIFICATION Differentiation of variables. COMMON CAUSE See RANDOM CAUSE. CONFIDENCE LEVEL The probability that a randomly distributed variable “x” lies within a defined interval of a normal curve. CONFIDENCE LIMITS The two values that define the confidence interval. CONFOUNDING Allowing two or more variables to vary together so that it is impossible to separate their unique effects.
CONSUMERS RISK Probability of accepting a lot when, in fact, the lot should have been rejected (see BETA RISK). CONTINUOUS DATA Data obtained from a measurement system which has an infinite number of possible outcomes. CONTINUOUS RANDOM A random variable which can assume any value VARIABLE continuously within some specified interval. CONTROL CHART A graphical rendition of a characteristic’s performance across time in relation to its natural limits and central tendency. CONTOL LIMITS Apply to both range or standard deviation and subgroup average (X) portions of process control charts and are used to determine the state of statistical control. Control limits are derived statistically and are not related to engineering specification limits in any way. CONTROL PLAN A formal quality document that describes all of the elements required to control variations in a particular process or could apply to a complete product or family of products. CONTROL SPECIFICATIONS Specification requirements for the product being manufactured. CORRELATION The relationship between two sets of data such that when one changes, the other is likely to make a corresponding change. Also, a statistical tool for determining the relationship between two sets of data. COST OF POOR QUALITY Cost associated with providing poor quality products or (COPQ) services. Can be divided into four cost categories: Appraisal, Scrap, Rework, and Field Complaint (warranty costs). CRITICAL TO QUALITY (CTQ) A drawing characteristic determined to be important for CHARACTERISTIC variability reduction based on a requirement from production, engineering, customer application, or regulatory agency. Can also apply to transactional or service delivery processes.
CUTOFF POINT The point which partitions the acceptance region from the reject region. DATA Factual information used as a basis for reasoning, discussion, or calculation; often refers to quantitative information. DATA TRANSFORMATION A mathematical technique used to create a near normally distributed data set out of a non-normal (skewed) data set. DEFECT Any product characteristic that deviates outside of specification limits. DEFECT PER MILLION Quality metric used in the Six Sigma process and is OPPORTUNITIES (DPMO) calculated by the number of defects observed divided by the number of opportunities for defects normalized to 1 million units. DEGREES OF FREEDOM The number of independent measurements available for estimating a population parameter. DENSITY FUNCTION The function which yields the probability that a particular random variable takes on any one of its possible values. DEPENDENT VARIABLE A Response Variable; e.g., y is the dependent or “Response” variable where Y = f(X 1. . . X N ) process input variables. DESIGN OF EXPERIMENT A formal, proactive method for documenting the selected (DOE) controllable factors and their levels, as well as establishing blocks, replications and response variables associated with a planned experiment. It is the plan for conducting the experiment and evaluating the results. DISCRETE DATA Data obtained from a measurement system which has a finite number of possible outcomes. DISCRETE RANDOM VARIABLE A random variable which can assume values only from a definite number of discrete values.
DISTRIBUTIONS Tendency of large numbers of observations to group themselves around some central value with a certain amount of variation or “scatter” on either side. EFFECT That which was produced by a cause. EVOLUTIONARY OPERATIONS A DOE process used to optimize the key process input (EVOPS) variables in a production environment, is usually limited to 2-3 variables, is performed over a long period of time, and is non-disruptive to the process. EXCEL Spreadsheet package within Microsoft Office used for data manipulation & analysis. EXPERIMENT A test under defined conditions to determine an unknown effect, to illustrate or verify a known law, or to establish a hypothesis. See DESIGN OF EXPERIMENT (DOE). EXPERIMENTAL ERROR Variation in observations made under identical test conditions. Also called residual error. The amount of variation which cannot be attributed to the variables included in the experiment. EXPONENTIALLY WEIGHTED A control charting method where the most current data MOVING AVERAGE (EWMA) point is weighted on an exponential basis such that older data points carry less value in calculating average. This charting technique is used to detect small shifts in process average. FACTORS Independent variables. FAILURE MODE & EFFECTS Analytical technique focused at problem prevention thru ANALYSIS (FMEA) identification of potential problems. The FMEA is a proactive tool that is used pragmatically to identify potential failure modes and their effects, to numerically rate the combined risk associated with severity, probability of occurrence and delectability and to document appropriate plans for prevention. FMEA’s can be applied to system, (application) and product design and to manufacturing and non-manufacturing processes (i.e., services & transactional processes).
FIRST TIME YIELD Yield that occurs in any process step prior to any rework that may be required (see Yft Symbology) to overcome process shortcomings. FIXED EFFECTS MODEL An experimental model where treatments are specifically selected by the researcher. Conclusions only apply to the factor levels considered in the analysis. Inferences are restricted to the experimental levels. FLUCTUATIONS Variances in data which are caused by a large number of minute variations or differences. FREQUENCY DISTRIBUTION The pattern or shape formed by the group of measurements in a distribution based on frequency of occurrence. GAGE ACCURACY The average difference observed between a gage under evaluation and a master gage when measuring the same parts over multiple readings. GAGE LINEARITY A measure of gage accuracy variation when evaluated over the expected operating range. GAGE REPEATABILITY A measure of the variation observed when a single operator uses a gage to measure a group of randomly ordered (but identifiable) parts on a repetitive basis. GAGE REPRODUCIBILITY A measure of average variation observed between operations when multiple operators use the same gage to measure a group of randomly ordered (but identifiable) parts on a repetitive basis. GAGE STABILITY A measure of variation observed when a gage is used to measure the same master over an extended period of time. GREEN BELT Six Sigma role similar in function to Black Belt but length of training and project scope are reduced. HISTOGRAM Vertical display of a population distribution in terms of frequencies; a formal method of plotting a frequency distribution.
HOMOGENEITY OF VARIANCE The variances of the data groups being contrasted are equal (as defined by a statistical test of significant difference). HYPOTHESIS When used as a statistical term, it is a theory proposed or postulated for comparing means and standard deviations of two or more data sets. A “null” hypothesis states that the data sets are from the same statistical population, while the “alternate” hypothesis states that the data sets are not from the same statistical population. INDEPENDENT VARIABLE A controlled variable; a variable whose value is independent of the value of another variable. INSTABILITY Unnaturally large fluctuations in a process input or output characteristic. INTERACTION The tendency of two or more variables to produce an effect in combination which neither variable would produce if acting alone. INTERVAL Numeric categories with equal units of measure but no absolute zero point, i.e., quality scale or index. KEY NOISE PARAMETERS Variables which are Hard or Expensive to control. KEY PROCESS INPUT The vital few input variables, called “x’s”, (normally 2-6) VARIABLES (KPIV’S) that drive 80% of the observed variations in the process output characteristic (“y”). a.k.a Key Control Parameters LINE CHARTS Charts used to track the performance without relationship to process capability or control limits. LOWER CONTROL LIMIT A horizontal dotted line plotted on a control chart which represents the lowest process deviation that should occur if the process is in control (free from assignable cause variation). MASTER BLACK BELT A person who is “expert” on Six Sigma techniques and on project implementation. Master Black Belts play a major role in training, coaching and in removing barriers to successful project execution in addition to overall promotion of the Six Sigma philosophy.
MEAN See AVERAGE. MEAN TIME BETWEEN Average time to failure for a statistically significant FAILURES (MTBF) population of product operating in its normal environment. MEASUREMENT SYSTEMS Means of evaluating a continuous or discrete ANALYSIS (MSA) measurement system to quantify the amount of variation contributed by the measurement system. Refer to Automotive Std. (AIAG STD) for details. MEDIAN The mid value in a group of measurements when ordered from low to high. MINITAB Statistical software package that operates on Microsoft Windows with a spreadsheet format and has powerful statistical analysis ability. MISTAKE PROOFING Mistake proofing is a proactive technique used to positively prevent errors from occurring. MIXED EFFECTS MODEL Contains elements of both the fixed and random effects models. MULTI-VARI Method used in the measure/analyze phase of Six Sigma to display in graphical terms the variation within parts, machines, or processes between machines or process parts, and over time. NONCONFORMING UNIT A unit which does not conform to one or more specifications, standards, and/or requirements. NONCONFORMITY A condition within a unit which does not conform to some specific specification, standard, and/or requirement; often referred to as a defect; any given nonconforming unit can have the potential for more than one nonconformity. NORMAL DISTRIBUTION A continuous, symmetrical density function characterized by a bell-shaped curve, e.g., distribution of sampling averages.
NORMALIZED ROLLED The estimate of the average process yield used to THROUGHPUT YIELD (RYTN) determine RTY. It is determined by taking the nth root of the RTY (where “n” is the # process step) included in the RTY calculation. NULL HYPOTHESIS An assertion to be proven by statistical analysis where two or more data sets are stated to be from the same population. ONE-SIDED ALTERNATIVE The value of a parameter which has an upper bound or a lower bound, but not both. ORDINAL Ordered categories (ranking) with no information about distance between each category, i.e., rank ordering of several measurements of an output parameter. ORDINATE The vertical axis of a graph. OUT OF CONTROL Condition which applies to statistical process control chart where plot points fall outside of the control limits or fail an established run or trend criteria, all of which indicate that an assignable cause is present in the process. PARAMETER A constant defining a particular property of the density function of a variable. PARETO DIAGRAM A chart which places common occurrences in rank order. P CHARTS Charts used to plot percent defectives in a sample where sample size is variable. PERTURBATION A nonrandom disturbance. POISSON DISTRIBUTION A statistical distribution associated with attribute data (the number of non-continuities found in a unit) and can be used to predict first pass yield. POPULATION A group of similar items from which a sample is drawn. Often referred to as the universe. POPULATION The entire set of items from which a sample is drawn.
POWER OF AN EXPERIMENT The probability of rejecting the null hypothesis when it is false and accepting the alternate hypothesis when it is true. PRECISION TO TOLERANCE A ratio used to express the portion of engineering RATIO (P/T) specification consumed by the 99% confidence interval of measurement system repeatability and reproducibility error. (5.15 standard deviations of R&R error) PREVENTION The practice of eliminating unwanted variation before the fact, e.g., predicting a future condition from a control chart and then applying corrective action before the predicted event transpires. PRIMARY CONTROL The major independent variables used in the experiment. VARIABLES PROBABILITY The chance of an event happening or condition occurring by pure chance and is stated in numerical form. PROBABILITY OF AN EVENT The number of successful events divided by the total number of trials. PROBLEM A deviation from a specified standard. PROBLEM SOLVING The process of solving problems; the isolation and control of those conditions which generate or facilitate the creation of undesirable symptoms. PROCESS A particular method of doing something, generally involving a number of steps or operations. PROCESS AVERAGE The central tendency of a given process characteristic across a given amount of time or at a specific point in time. PROCESS CONTROL See STATISTICAL PROCESS CONTROL. PROCESS CONTROL CHART Any of a number of various types of graphs upon which data are plotted against specific control limits.
PROCESS MAP A detailed step-by-step pictorial sequence of a process showing process inputs, potential or actual controllable and uncontrollable sources of variation, process outputs, cycle time, rework operations, and inspection points. PROCESS SPREAD The range of values which a given process characteristic displays; this particular term most often applies to the range but may also encompass the variance. The spread may be based on a set of data collected at a specific point in time or may reflect the variability across a given period of time. PRODUCERS RISK Probability of rejecting a lot when, in fact, the lot should have been accepted (see ALPHA RISK). PROJECT A problem, usually calling for planned action. QUALITY FUNCTION QFD is a disciplined matrix methodology used for DEPLOYMENT (QFD) documenting customer wants and needs – “the voice of the customer” – into operational “requirement” terms. It is an effective tool for determining critical-to-quality characteristics for transactional processes, services and products. R CHART Plot of the difference between the highest and lowest in a sample. Normally associated with the range control portion of an X, R chart. RANDOM CAUSE A source of variation which is random, usually associated with the “trivial many” process input variables, and which will not produce a highly predictable change in the process output response (dependent variable), e.g., a correlation does not exist; any individual source of variation results in a small amount of variation in the response; cannot be economically eliminated from a process; an inherent natural source of variation. RANDOMNESS A condition in which any individual event in a set of events has the same mathematical probability of occurrence as all other events within the specified set, i.e., individual events are not predictable even though they may collectively belong to a definable distribution.
RANDOM SAMPLE One or more samples randomly selected from the universe (population). RANDOM SAMPLE Selecting a sample such that each item in the population has an equal chance of being selected; lack of predictability; without pattern. RANDOM VARIABLE A variable which can assume any value from a distribution which represents a set of possible values. RANDOM VARIATIONS Variations in data which result from causes which cannot be pinpointed or controlled. RANGE The difference between the highest and lowest values in a “subgroup” sample. RANK Values assigned to items in a sample to determine their relative occurrence in a population. RATIONAL SUBGROUP A subgroup is usually made up of consecutive pieces chosen from the process stream so that the variation represented within each subgroup is as small as feasible. Any changes, shifts and drifts in the process will appear as differences between the subgroups, selected over time. REGRESSION A statistical technique for determining the best mathematical expression that describes the functional relationship between one response and one or more independent variables. REJECT REGION The region of values for which the alternate hypothesis is accepted. REPLICATION Repeat observations made under identical test conditions. REPRESENTATIVE SAMPLE A sample which accurately reflects a specific condition or set of conditions within the universe.
RESEARCH Critical and exhaustive investigation or experimentation having for its aim the revision of accepted conclusions in the light of newly discovered facts. RESIDUAL ERROR See EXPERIMENTAL ERROR. RESPONSE SURFACE A graphical (pictorial) analysis technique used in METHODOLOGY (RSM) conjunction with DOE for determining optimum process parameter settings. ROBUST The condition or state in which a response parameter exhibits a high degree of resistance to external causes of a nonrandom nature; i.e., impervious to perturbing influence. ROLLED THROUGHPUT YIELD The product (series multiplication) of all of the individual (RTY) first pass yields of each step of the total process. ROOT SUM SQUARED (RSS) Square root of the sum of the squares. Means of combining standard deviations from independent causes. SAMPLE A portion of a population of data chosen to estimate some characteristic about the whole population. One or more observations drawn from a larger collection of observations or universe (population). SCATTER DIAGRAMS Charts which allow the study of correlation, e.g., the relationship between two variables or data sets. SHORT RUN STATISTICAL A statistical control charting technique which applies to PROCESS CONTROL any process situation where there is insufficient frequency of subgroup data to use traditional control charts (typically associated with low-volume manufacturing or where setups occur frequently). Multiple part numbers and multiple process streams can be plotted on a single chart. SIX M’S The major categories that contribute to effects on the fishbone diagram (man, machine, material, method, measurement, and mother nature.
SIX SIGMA A term coined by Motorola to express process capability in parts per million. A Six Sigma process generates a maximum defect probability of 3.4 parts per million (PPM) when the amount of process shifts and drifts are controlled over the long term to less than + 1.5 standard deviations. SKEWED DISTRIBUTION A non-symmetrical distribution having a tail in either a positive or negative direction. SPECIAL CAUSE See ASSIGNABLE CAUSE. STABLE PROCESS A process which is free of assignable causes, e.g., in statistical control. STANDARD DEVIATION A statistical index of variability which describes the process spread or width of distribution. STATISTICAL CONTROL A quantitative condition which describes a process that is free of assignable/special causes of variation (both mean and standard deviation). Such a condition is most often evidenced on a control chart, i.e., a control chart which displays an absence of nonrandom variation. STATISTICAL PROCESS The application of standardized statistical methods and CONTROL (SPC) procedures to a process for control purposes. SUBGROUP A logical grouping of objects or events which displays only random event-to-event variations, e.g., the objects or events are grouped to create homogenous groups free of assignable or special causes. By virtue of minimizing within subgroup variability, any change in the central tendency or variance of the universe will be reflected in the “subgroup-to-subgroup” variability. A predetermined sample of consecutive parts or other data bearing objects removed from the process for the purpose of data collection. SYMPTOM That which serves as evidence of something not fully understood in factual terms.
SYSTEM That which is connected according to a scheme. SYSTEMATIC VARIABLES A pattern which displays predictable tendencies. TEST OF SIGNIFICANCE A statistical procedure used to determine whether or not a process observation (data set) differs from a postulated value by an amount greater than that due to random variation alone. THEORY A plausible or scientifically acceptable general principle offered to explain phenomena. TWO-SIDED ALTERNATIVE The values of a parameter which designate both an upper and lower bound. TYPE I ERROR See ALPHA RISK. TYPE II ERROR See BETA RISK. UNNATURAL PATTERN Any pattern in which a significant number of the measurements do not group themselves around a central tendency. When the pattern is unnatural, it means that non-random disturbances are present and are affecting the process. UPPER CONTROL LIMIT A horizontal line on a control chart (usually dotted) which represents the upper limits of capability for a process operating with only random variation. VARIABLE A characteristic that may take on different values. VARIABLES DATA Data collected from a process input or output where the measurement scale has a significant level of subdivisions or resolution, e.g., ohms, voltage, diameter, etc. VARIATION Any quantifiable difference between individual measurements; such differences can be classified as being due to common causes (random) or special causes (assignable). VARIATION RESEARCH Procedures, techniques, and methods used to isolate one type of variation from another (for example, separating product variation from test variation). X & R CHARTS A control chart which is a representation of process capability over time; displays the variability in the process average and range across time.