- Why the Measure and Analyze Phases Matter on the CSSGB Exam
- Measure Phase Overview: What the BOK Expects
- Process Mapping and Documentation Tools
- Data Collection Planning and Sampling Strategies
- Measurement System Analysis (MSA)
- Basic Statistics You Must Know Cold
- Process Capability and Performance Indices
- Analyze Phase Overview: From Data to Root Causes
- Hypothesis Testing Essentials
- Regression and Correlation Analysis
- Failure Mode and Effects Analysis (FMEA)
- Graphical Analysis Tools
- Exam Strategy: Tackling Statistical Questions
- Frequently Asked Questions
Why the Measure and Analyze Phases Matter on the CSSGB Exam
The Measure and Analyze phases together represent 38% of the ASQ Certified Six Sigma Green Belt exam — the single largest combined weighting across all six DMAIC domains. If you walk into Prometric without a solid grasp of statistical tools, you are essentially gambling with more than a third of the 100 scored questions. These two phases are where Six Sigma separates itself from generic process improvement, relying on data-driven decision making instead of gut instinct.
For context on what the full exam covers across all domains, review our breakdown of the ASQ Six Sigma Green Belt Body of Knowledge 2026: All 6 DMAIC Domains Explained. In this article, we focus specifically on the statistical concepts, formulas, and tools you must understand before sitting for the exam.
Measure Phase Overview: What the BOK Expects
The Measure phase is about quantifying the current state of your process. Before you can improve anything, you need a reliable baseline. The ASQ Body of Knowledge (2022 version, effective August 2022) expects Green Belt candidates to understand process documentation, data collection strategies, measurement system reliability, basic statistics, and process capability. Each of these subtopics carries exam questions that range from conceptual definitions to practical calculations.
Remember that the CSSGB is an open-book exam — you are allowed to bring bound reference materials, and an on-screen scientific calculator is provided. This means the exam tests your ability to apply formulas correctly under time pressure, not simply to memorize them. For strategies on leveraging your reference materials, see our guide on ASQ CSSGB Exam Day Tips: Open-Book Strategies and Reference Material Guide.
Process Mapping and Documentation Tools
Before collecting data, you must understand the process well enough to know what to measure and where to measure it. The BOK includes several process documentation tools within the Measure phase.
Process Flowcharts and SIPOC Diagrams
A SIPOC diagram (Suppliers, Inputs, Process, Outputs, Customers) provides a high-level view of the process. On the exam, expect questions asking you to identify the correct order of SIPOC elements or to select the appropriate tool for documenting a process at a macro level. Detailed process flowcharts use standard symbols — ovals for start/end, rectangles for process steps, diamonds for decision points, and arrows for flow direction.
Value Stream Mapping
Value stream maps extend basic flowcharts by adding time data (cycle time, wait time, lead time) and inventory levels. You should be able to distinguish value-added steps from non-value-added steps. The exam may ask you to calculate process cycle efficiency: the ratio of value-added time to total lead time.
PCE = Value-Added Time / Total Lead Time × 100%
A typical transactional process has a PCE of 1–10%. Manufacturing processes often range from 5–25%. The exam may present a value stream scenario and ask you to compute PCE or identify waste categories.
Data Collection Planning and Sampling Strategies
Collecting the wrong data — or collecting the right data incorrectly — invalidates everything that follows. The CSSGB BOK expects candidates to understand data types, sampling methods, and sample size considerations.
Data Types: Continuous vs. Discrete
Recognizing data types is foundational because it determines which statistical tools are appropriate. Continuous (variable) data can take any value within a range — measurements like temperature, time, weight, and length. Discrete (attribute) data falls into categories — pass/fail, count of defects, number of complaints. The exam frequently tests whether you can identify which type of data a scenario describes and, critically, which statistical test or chart applies to that data type.
Sampling Methods
| Sampling Method | Description | When to Use |
|---|---|---|
| Simple Random | Every item has equal probability of selection | Homogeneous population, no subgroup structure |
| Stratified | Population divided into subgroups; random sample from each | Known subgroup variation (shifts, machines, regions) |
| Systematic | Select every nth item after a random start | Ordered process flow, no periodic patterns |
| Subgroup/Rational | Small consecutive groups at regular intervals | SPC applications, monitoring process stability |
On the exam, you will encounter scenarios describing a production environment and be asked to select the most appropriate sampling method. Pay close attention to whether the scenario mentions shifts, machines, or other stratification factors.
Measurement System Analysis (MSA)
If your measurement system is unreliable, your data is meaningless. MSA evaluates five properties of a measurement system: bias, linearity, stability, repeatability, and reproducibility. The most heavily tested MSA concept on the CSSGB exam is the Gage R&R study.
Gage R&R Studies
Repeatability refers to variation when the same operator measures the same part multiple times with the same gage. Reproducibility refers to variation when different operators measure the same part with the same gage. The combined Gage R&R value tells you how much of total observed variation is attributable to the measurement system itself rather than actual part-to-part differences.
Less than 10% of total variation: Measurement system is acceptable.
10–30%: May be acceptable depending on application, cost, and risk.
Greater than 30%: Measurement system is unacceptable and must be improved before proceeding.
The exam frequently tests these thresholds. If a question provides Gage R&R results and asks whether the measurement system is adequate, apply these rules directly.
Basic Statistics You Must Know Cold
Even though the CSSGB exam is open-book, certain statistical concepts appear so frequently that working through them from scratch each time will consume too much of your 4 hours and 18 minutes. Build fluency with these fundamentals during your study period.
Measures of Central Tendency and Dispersion
You must understand mean, median, and mode, as well as range, variance, and standard deviation. The exam will test when each is appropriate. For example, median is more robust than the mean when data contains outliers — a fact that exam questions exploit regularly. Know that standard deviation uses (n−1) in the denominator for sample data and (N) for population data.
The Normal Distribution and the Empirical Rule
The normal distribution is the backbone of Six Sigma statistical analysis. You should know the empirical rule without looking it up: 68.27% of data falls within ±1σ, 95.45% within ±2σ, and 99.73% within ±3σ. Six Sigma quality targets ±6σ, corresponding to 3.4 defects per million opportunities (DPMO) when accounting for a 1.5σ process shift.
Central Limit Theorem
The Central Limit Theorem (CLT) states that the distribution of sample means approaches a normal distribution as sample size increases, regardless of the underlying population distribution. This is critical because it justifies using normal-based statistical tests even when individual data points are not normally distributed. On the exam, questions about CLT often test whether you understand that it applies to the distribution of sample means, not individual values, and that a sample size of 30 or more is the general rule of thumb.
Process Capability and Performance Indices
Process capability analysis compares process output to customer specifications. This is one of the most calculation-heavy topics on the exam and consistently appears in multiple questions.
| Index | Formula | What It Measures |
|---|---|---|
| Cp | (USL − LSL) / 6σ | Potential capability (spread only, ignores centering) |
| Cpk | min[(USL − μ) / 3σ, (μ − LSL) / 3σ] | Actual capability (accounts for centering) |
| Pp | (USL − LSL) / 6s | Overall performance (uses overall std dev) |
| Ppk | min[(USL − μ) / 3s, (μ − LSL) / 3s] | Overall performance with centering |
A process can have a high Cp (the specification range is wide relative to process spread) but a low Cpk (the process is not centered). Cpk ≤ Cp always. If Cpk equals Cp, the process is perfectly centered. The exam loves presenting scenarios where Cp looks good but Cpk reveals an off-center process. A Cpk of 1.33 or higher is generally considered capable.
For a thorough breakdown of how to structure your study time across all domains including these calculation-heavy topics, refer to our Six Sigma Green Belt Study Plan: How to Prepare for the CSSGB in 8 Weeks.
Analyze Phase Overview: From Data to Root Causes
While the Measure phase answers "What is happening?", the Analyze phase answers "Why is it happening?" At 18% of the exam, this domain tests your ability to use statistical tools to identify root causes and validate relationships between variables. The Analyze phase bridges the quantitative baseline from Measure with the solution-oriented work of the CSSGB Improve Phase: DOE, Lean Tools, and Kaizen Concepts for the Exam.
Hypothesis Testing Essentials
Hypothesis testing is the statistical framework for making decisions from data, and it is heavily represented on the CSSGB exam. You need to understand the structure, the common tests, and how to interpret results.
The Hypothesis Testing Framework
The null hypothesis (H₀) assumes no effect, no difference, or no relationship. The alternative hypothesis (H₁ or Hₐ) is what you are trying to prove. Example: H₀: μ₁ = μ₂ (no difference between means); H₁: μ₁ ≠ μ₂ (means differ).
Alpha (α) is the probability of rejecting H₀ when it is actually true (Type I error). Common values are 0.05 (5%) and 0.01 (1%). On the exam, unless stated otherwise, assume α = 0.05.
The correct test depends on your data type, sample size, number of groups, and whether you know the population standard deviation. This is where the exam focuses — matching the scenario to the right test.
Compute the test statistic and compare it to the critical value, or compare the p-value to α. If the p-value is less than α, reject H₀. The exam provides the calculator — you need to know which formula to use.
State the conclusion in context. "There is sufficient evidence to conclude that the process mean has shifted" — not just "reject H₀." Some exam questions test your ability to correctly interpret the result.
Common Hypothesis Tests on the CSSGB Exam
| Test | Purpose | Data Type | Key Condition |
|---|---|---|---|
| 1-sample t-test | Compare sample mean to a target value | Continuous | σ unknown, small sample |
| 2-sample t-test | Compare means of two independent groups | Continuous | Independent samples, σ unknown |
| Paired t-test | Compare means of two related measurements | Continuous | Before/after on same subjects |
| 1-proportion test | Compare sample proportion to a target | Attribute | np ≥ 5 and n(1−p) ≥ 5 |
| 2-proportion test | Compare proportions from two groups | Attribute | Independent samples, adequate counts |
| Chi-square test | Test association between categorical variables | Attribute | Expected frequencies ≥ 5 |
| ANOVA (F-test) | Compare means of three or more groups | Continuous | Independent groups, equal variances |
Type I Error (α): Rejecting H₀ when it is true (false positive). You conclude there is an effect when there is none.
Type II Error (β): Failing to reject H₀ when it is false (false negative). You miss a real effect.
Power = 1 − β: The probability of correctly detecting a real effect.
The exam will ask you to identify which error type applies in a given scenario. A common trick question: "A team concludes the new process is better, but in reality there is no improvement." This is a Type I error.
Regression and Correlation Analysis
Regression and correlation quantify relationships between variables. The CSSGB exam focuses primarily on simple linear regression and the correlation coefficient.
Correlation Coefficient (r)
The Pearson correlation coefficient r ranges from −1 to +1. A value of +1 indicates a perfect positive linear relationship, −1 indicates a perfect negative linear relationship, and 0 indicates no linear relationship. On the exam, you may be shown a scatter plot and asked to estimate r, or given r and asked to interpret its meaning. Remember: correlation does not imply causation — this is a frequently tested concept.
Coefficient of Determination (R²)
R² is the square of r and represents the proportion of variation in the dependent variable explained by the independent variable. If r = 0.8, then R² = 0.64, meaning 64% of the variation in Y is explained by X. The remaining 36% is unexplained. Exam questions often present a regression output and ask you to interpret R².
Simple Linear Regression
The regression equation Y = a + bX allows you to predict the output (Y) based on an input (X). You should understand the components: a is the y-intercept, b is the slope (change in Y per unit change in X). The exam may ask you to use a given regression equation to make a prediction or interpret the slope in context.
Failure Mode and Effects Analysis (FMEA)
FMEA is a structured risk assessment tool used in the Analyze phase to prioritize potential failure modes. Each potential failure is scored on three criteria:
- Severity (S): How serious is the effect if the failure occurs? (1–10 scale)
- Occurrence (O): How likely is the failure to occur? (1–10 scale)
- Detection (D): How likely is the failure to be detected before reaching the customer? (1–10 scale, where 10 = least likely to detect)
The Risk Priority Number (RPN) = S × O × D, with a maximum possible value of 1,000. Higher RPN values indicate higher priority for corrective action. Pay special attention to the detection scale — it is counterintuitive because a higher detection score means the failure is harder to detect, which increases risk.
When an exam question asks which failure mode to address first, do not automatically choose the highest RPN. Some questions may specify a company policy of prioritizing severity above all else — any failure with a severity of 9 or 10 gets immediate attention regardless of RPN. Read the question carefully for these policy-based qualifications.
Graphical Analysis Tools
The Analyze phase relies heavily on graphical tools for pattern recognition. You should be able to identify and interpret each of these on the exam.
Histograms
Histograms display the distribution shape of continuous data. Look for normality, skewness, bimodal patterns (which suggest two mixed populations), and truncation (which may indicate sorting or inspection). The exam may show a histogram and ask what it reveals about the process.
Pareto Charts
Pareto charts display defect categories in descending order with a cumulative percentage line. The 80/20 rule (roughly 80% of problems come from 20% of causes) is the guiding principle. Exam questions may ask you to identify which categories to address first based on a Pareto chart or to recognize when a Pareto chart is the appropriate tool.
Scatter Diagrams
Scatter diagrams visualize the relationship between two continuous variables. They complement correlation and regression analysis by providing a visual check for linearity, outliers, and clustering patterns that numerical summaries alone might miss.
Box-and-Whisker Plots
Box plots display the median, quartiles, and potential outliers in a compact format. They are especially useful for comparing distributions across multiple groups. The exam may present side-by-side box plots and ask you to compare group medians or identify which group has greater variability.
Exam Strategy: Tackling Statistical Questions
With roughly 42 scored questions across the Measure and Analyze phases, your approach to statistical problems can make or break your result. The CSSGB exam difficulty is often driven by these two domains, and the 77% pass rate in 2024 reflects the challenge these statistical concepts pose.
Since the CSSGB is open-book, create tabbed sections in your reference materials for: (1) hypothesis test selection flowchart, (2) capability index formulas, (3) Gage R&R thresholds, (4) control chart selection guide, and (5) z-tables and t-tables. Being able to flip to the right page in under 10 seconds per question can save you 20+ minutes across the exam.
Practice applying these tools with realistic scenarios. Our CSSGB practice test questions mirror the statistical rigor of the actual exam and let you test your understanding under timed conditions. Working through Six Sigma Green Belt Practice Questions 2026: Free CSSGB Sample Problems is one of the most effective ways to build speed and accuracy.
Decision Framework for Exam Questions
When you encounter a statistical question, follow this sequence: First, identify the data type (continuous or attribute). Second, determine how many groups or samples are involved. Third, identify what is being compared (means, proportions, variances, or relationships). This three-step filter will lead you to the correct tool nearly every time.
Once you have mastered the Measure and Analyze phases, the remaining domains become more manageable. The CSSGB Control Phase: SPC Charts, Control Plans, and Sustaining Improvements builds directly on the statistical concepts covered here, particularly SPC chart selection and interpretation. Connecting these domains in your study plan creates natural reinforcement that strengthens retention.
For candidates weighing the broader value of earning this credential, the investment in mastering these statistical tools extends well beyond the exam room. Understanding how certification impacts compensation is covered in our analysis of Six Sigma Green Belt Salary 2026: How CSSGB Certification Impacts Your Earnings.
Frequently Asked Questions
The Measure phase accounts for 20% and the Analyze phase accounts for 18% of the exam, totaling 38%. With 100 scored questions on the CBT format, you can expect approximately 38–42 questions from these two domains combined. Since there are also 10 unscored pretest items distributed randomly, some additional Measure and Analyze questions may appear that do not count toward your score.
No. The CSSGB is an open-book exam, meaning you can bring bound reference materials that include formulas, tables, and notes. However, you should be familiar enough with key formulas (Cp, Cpk, standard deviation, basic hypothesis test structures) to apply them quickly. Looking up every formula will cost valuable time during the 4 hours and 18 minutes of testing. An on-screen scientific calculator is provided, so you do not need to bring your own.
Cp measures potential capability by comparing the specification width to the process spread (6σ), but it ignores whether the process is centered. Cpk accounts for centering by measuring capability relative to the nearest specification limit. A process can have a high Cp but a low Cpk if it is off-center. The exam frequently presents scenarios where you must calculate both and explain the discrepancy, or choose which index is more appropriate for a given situation.
Use a z-test when the population standard deviation (σ) is known and the sample is large (n ≥ 30). Use a t-test when σ is unknown and you are using the sample standard deviation (s) as an estimate, which is the far more common scenario in practice and on the exam. At the Green Belt level, most exam questions involving hypothesis testing use t-tests because real-world scenarios rarely provide the population standard deviation.
Many candidates find the Analyze phase more challenging because it requires applying statistical tools to draw conclusions rather than just collecting and summarizing data. Hypothesis testing, regression interpretation, and root cause validation require deeper analytical thinking. However, difficulty is subjective — candidates with a strong statistics background may find the Measure phase's MSA and capability concepts equally demanding. Practicing with realistic exam questions across both domains is the best way to identify and address your personal weak areas.
Ready to Start Practicing?
The Measure and Analyze phases make up 38% of your CSSGB exam. Test your mastery of statistical tools, hypothesis testing, process capability, and root cause analysis with practice questions designed to match ASQ exam rigor.
Start Free Practice Test →