- Domain 4 (Analyze Phase) requires both qualitative root cause tools and inferential statistical testing - you must know when to use each.
- Fishbone diagrams, 5 Whys, and regression analysis are all fair game; CSSGB questions test application, not just definitions.
- Hypothesis testing on the CSSGB exam focuses on selecting the correct test for a given data type and sample scenario.
- Analyze Phase outputs feed directly into Domain 5 (Improve Phase), so weak Analyze knowledge compounds into Improve errors.
What the Analyze Phase Actually Tests
The Analyze Phase sits at the pivot point of the DMAIC methodology. By the time a Six Sigma project reaches Domain 4, the team has already defined the problem scope, identified the critical-to-quality characteristics, and collected baseline process data. The Analyze Phase demands something harder: explaining why the problem exists, not merely confirming that it does.
On the Certified Six Sigma Green Belt (CSSGB) exam, Domain 4 is one of the most technically demanding sections. Candidates are expected to move fluidly between qualitative reasoning tools - such as cause-and-effect diagrams and process analysis - and quantitative statistical tools like hypothesis tests and regression. The exam does not ask you to memorize formulas in isolation; it presents realistic process scenarios and asks you to identify the correct analytical approach, interpret outputs, and draw defensible conclusions.
This dual demand - soft analytical thinking paired with hard statistical reasoning - is what catches many candidates off guard. Someone who studied statistics but ignored cause-and-effect mapping, or vice versa, will struggle with the scenario-based format the CSSGB uses throughout its question bank.
Core Root Cause Analysis Tools You Must Master
Cause-and-Effect Diagrams (Fishbone / Ishikawa)
The fishbone diagram is the most recognizable root cause tool in Six Sigma, and the CSSGB exam tests it with more nuance than most candidates expect. You must know not only how to construct one but how to categorize causes using the classic 6M framework: Man, Machine, Material, Method, Measurement, and Mother Nature (Environment). Exam questions often present a partially completed diagram and ask which category a specific cause belongs to, or they describe a manufacturing or transactional scenario and ask which branches would be most relevant.
A second layer of difficulty involves knowing the difference between a cause and a symptom. The CSSGB will frequently present a diagram where one branch contains an effect rather than a true cause, and the question asks candidates to identify the error.
The 5 Whys Technique
The 5 Whys is deceptively simple and deceptively tricky on the exam. The method involves iteratively asking "why" a problem occurs until the root cause - typically a systemic or process failure - is revealed. CSSGB questions test whether candidates can distinguish between a root cause and an intermediate cause. A stopping point that still describes a symptom rather than a correctable process variable is a wrong answer, and the exam exploits this distinction repeatedly.
Failure Mode and Effects Analysis (FMEA)
FMEA sits at the intersection of Analyze and risk management. For the CSSGB, you need to understand how to calculate the Risk Priority Number (RPN) - the product of Severity, Occurrence, and Detection ratings - and how to interpret it. More importantly, you need to understand what changes to each factor accomplish. Reducing Occurrence (fixing the root cause) is more valuable than improving Detection (catching the defect after it forms), and the exam tests this judgment call.
Domain 4: Analyze Phase - Key Competency Areas
CSSGB candidates must demonstrate competency across both process analysis and statistical analysis within this domain.
- Process and value stream mapping to identify non-value-added steps
- Cause-and-effect diagrams, affinity diagrams, and interrelationship diagraphs
- 5 Whys and fault tree analysis
- Hypothesis testing: selecting the right test, stating hypotheses, interpreting p-values
- Regression analysis and correlation interpretation
- Multi-vari studies and sources of variation
- FMEA construction and RPN prioritization
Multi-Vari Studies
Multi-vari analysis is a graphical tool used to identify families of variation: positional (within a unit), cyclical (unit-to-unit), and temporal (over time). On the CSSGB exam, multi-vari questions often describe a production scenario and ask which type of variation is dominant based on a described chart pattern. This tool bridges qualitative process observation and quantitative data collection in a way that CSSGB scenarios are designed to exploit.
Hypothesis Testing: The Statistical Heart of Analyze
If root cause tools are the qualitative backbone of Domain 4, hypothesis testing is the statistical proof mechanism. The CSSGB exam requires candidates to understand not just what hypothesis testing is, but which test to apply in which situation. This selection logic is one of the most tested competencies in the entire certification.
Structuring the Null and Alternative Hypotheses
The CSSGB will present scenarios where candidates must correctly frame H₀ and H₁. A common question type shows a flawed hypothesis pair and asks candidates to identify the error - for example, a null hypothesis that contains an inequality sign, or an alternative hypothesis that is non-directional when the scenario clearly implies a directional test. Understanding the logic of burden of proof (the null is assumed true until evidence rejects it) is foundational here.
Selecting the Right Statistical Test
Test selection is arguably the highest-leverage skill in the Analyze Phase. The CSSGB exam presents scenarios with varying data types, sample sizes, and comparison goals, and candidates must identify the appropriate test without being led by explicit cues.
| Scenario | Data Type | Appropriate Test |
|---|---|---|
| Compare one sample mean to a target value | Continuous | One-sample t-test |
| Compare means of two independent groups | Continuous | Two-sample t-test |
| Compare means before and after within same units | Continuous | Paired t-test |
| Compare means of three or more groups | Continuous | One-way ANOVA |
| Assess relationship between two continuous variables | Continuous | Correlation / Regression |
| Compare proportions or counts across categories | Discrete | Chi-square test |
| Non-normal data, compare two medians | Continuous (non-normal) | Mann-Whitney U test |
Interpreting p-Values and Practical Significance
The exam tests a critical distinction: statistical significance is not the same as practical significance. A very large sample size can produce a statistically significant result (low p-value) even when the actual difference is negligible in process terms. CSSGB questions present both the p-value and the effect size context, and candidates must reason through whether the result is actionable - a judgment that requires understanding both the statistics and the business context of the scenario.
Key Takeaway
When a CSSGB question gives you a p-value below 0.05, your job is not finished. Always check whether the magnitude of the difference or relationship is large enough to matter in the described process context. Statistical significance and business significance require separate evaluations.
Regression and Correlation
Simple linear regression is tested extensively in the Analyze Phase. Candidates must interpret the coefficient of determination (R²), understand what slope and intercept represent in a process context, and identify when a regression model should not be used for prediction (extrapolation beyond the data range, non-linear relationships, outlier influence). The CSSGB also tests correlation coefficient interpretation - including the important caveat that correlation does not establish causation, a logic error the exam specifically probes.
How CSSGB Exam Questions Frame Analyze Concepts
The CSSGB exam is scenario-driven. Questions in Domain 4 rarely ask "what is a fishbone diagram?" - they describe a team meeting, present a partially completed analysis, and ask what the team should do next or where the analysis has gone wrong. This format rewards candidates who have practiced applying tools, not just reading about them.
A recurring question structure presents two plausible root causes identified through qualitative analysis, then gives statistical output showing that only one has a significant relationship with the output variable. The question asks which root cause should be prioritized for improvement. The correct answer requires integrating both the qualitative framing and the quantitative evidence - exactly the kind of reasoning the CSSGB practice tests at our exam prep platform are designed to develop.
Another common trap involves FMEA prioritization. A question may show a process failure with a very high Severity rating but low Occurrence and Detection, resulting in a moderate RPN. A separate failure has a lower Severity but very high Occurrence and Detection scores, yielding a higher RPN. The question asks which failure should be addressed first. While RPN drives prioritization mathematically, the CSSGB also tests whether candidates recognize that failures with catastrophic Severity scores (typically rated 9 or 10) often warrant attention regardless of the composite RPN.
How Analyze Connects to Other DMAIC Domains
Domain 4 does not exist in isolation, and the CSSGB exam tests candidates on the logical handoffs between phases. The Analyze Phase receives its inputs from Domain 3 (Measure Phase) - specifically the baseline process data, measurement system validation results, and capability indices collected there. If the Measure Phase was executed poorly, the Analyze Phase will be working with unreliable data, a scenario the exam explicitly addresses.
On the output side, the verified root causes from Domain 4 become the design targets for Domain 5 (Improve Phase). CSSGB questions test this connection by presenting an improvement solution and asking whether it logically addresses the root cause identified in the Analyze Phase. Candidates who understand this forward linkage will recognize when a proposed solution is a mismatch - a common distractor answer type on the exam.
Before you reach the Analyze Phase in your actual exam preparation, it is worth revisiting the foundational certification requirements. Our article on CSSGB Certification Requirements: Eligibility and Application covers the prerequisites, work experience documentation, and application process that all candidates must navigate before sitting for the exam.
For a deeper dive into Analyze-specific practice scenarios and worked examples, the CSSGB Exam Prep practice test platform offers domain-filtered question sets so you can drill Domain 4 material in isolation before tackling full-length mixed exams.
Sequencing Your Analyze Phase Study
Because Domain 4 depends heavily on Domain 3 foundations (particularly understanding variation, measurement system analysis, and process capability), your study sequence matters. Attempting to learn hypothesis testing before you are confident in basic process statistics - control charts, standard deviation interpretation, normal distribution properties - creates unnecessary confusion.
Qualitative Root Cause Tools
- Build and critique fishbone diagrams across manufacturing and service scenarios
- Practice 5 Whys until you can distinguish root causes from intermediate causes consistently
- Study FMEA construction: complete an RPN table and practice prioritization logic
Statistical Foundations for Hypothesis Testing
- Review sampling distributions, standard error, and the Central Limit Theorem
- Practice framing null and alternative hypotheses from scenario descriptions
- Work through test selection logic using the comparison table above as a decision guide
Applied Statistical Analysis and Integration
- Interpret regression outputs: R², slope, residual plots, and extrapolation limits
- Practice ANOVA interpretation and post-hoc reasoning
- Take Domain 4-specific timed practice sets to identify remaining gaps before moving to Domain 5
The spaced repetition principle applies well here: after completing each week's focus, return to the prior week's material for a brief review session before advancing. For CSSGB candidates, this means revisiting fishbone and FMEA concepts during your Week 2 statistical work, reinforcing the connection between qualitative cause identification and the quantitative confirmation that follows. The full article on the CSSGB Analyze Phase: Root Cause Tools and Hypothesis Testing can serve as a reference anchor throughout this study sequence.
When you feel confident in Domain 4 concepts individually, challenge yourself with full mixed-domain practice exams on the CSSGB Exam Prep platform to ensure you can apply Analyze reasoning even when questions are interleaved with Define, Measure, Improve, and Control material - which is exactly how the actual certification exam presents them.
Frequently Asked Questions
The CSSGB exam covers all six domains across the DMAIC framework plus organizational and Lean/DFSS fundamentals. Domain 4 (Analyze Phase) is one of the most content-dense sections because it spans both qualitative root cause tools and quantitative statistical methods. Candidates should expect a meaningful portion of exam questions to draw on Analyze Phase concepts, either directly or as embedded logic within multi-phase scenario questions.
The CSSGB exam focuses heavily on conceptual application and interpretation rather than manual calculation. You should understand how statistics are derived and what they mean in context, but the exam is primarily testing your ability to select the right tool, interpret outputs, and apply results to process decisions - not to compute test statistics by hand.
A fishbone diagram categorizes potential causes across multiple dimensions (the 6Ms) to generate a comprehensive list of hypotheses. The 5 Whys drills vertically into a single causal chain to reach root depth. The CSSGB exam may ask which tool is appropriate for a given situation: use fishbone when you want broad cause exploration across categories; use 5 Whys when you have a specific failure to trace to its systemic origin.
First, identify what the output is measuring (a test statistic, p-value, R², RPN, etc.). Second, apply the correct interpretation rule for that metric. Third, connect the interpretation to the process scenario described - statistical results must be evaluated in business context, not just in mathematical terms. Practice with scenario-based questions is essential because the exam combines all three steps in a single question.
Neither gap is safe to carry into the exam. Domain 4 explicitly requires competency in both areas, and many CSSGB questions integrate both by presenting a qualitative analysis scenario that leads into a statistical validation step. A candidate who is strong in one area but weak in the other will consistently get these integrated questions wrong. Prioritize closing your weaker area before exam day.