Exam Content Guide
JMP Certified Associate:
Statistical Thinking for Industrial Problem Solving
Section 1: Statistical Thinking and Problem Solving: 10%
Define statistical thinking and problem solving
• Explain the benefits of statistical thinking
• List the principles of statistical thinking
• Describe the importance of following a structured approach to problem solving
• Explain the importance of a well-written problem definition
• Identify and define KPIs (Key Performance Indicators)
o Define what success looks like and determine how to measure for success and
improvement
o Describe the importance of establishing a baseline for your KPIs using
o Establish a baseline for your KPIs using control charts
Define the process for creation of a process map or SIPOC
• Interpret a process map
o Explain the importance of a process map in understanding the process and scoping
the problem
o Define the parts of a process map
• Interpret a SIPOC model (Supplier, Input, Process, Output, Customer)
o Explain why you would do a SIPOC
Identify root causes of a problem, compile and collect data
• Develop and interpret a cause and effect (C&E) diagram
o Identify the components of a simple C&E diagram
o Given a completed C&E diagram, determine action to take based on the completed
diagram.
• Define rules for effective brainstorming
o Explain methods for brainstorming and methods for gaining consensus.
o Explain how Affinity diagrams are used for organizing output from a brainstorming
session
• Define operational definitions and why they are important
o Given a particular KPI, identify the appropriate operational definition for it
• Describe data types and modeling types
1
Exam Content Guide
• Given a problem, be able to create a data collection plan
o Determine the data you need to collect
o Determine if there is historical data available that can be used
▪ Determine locations/sources of existing data
o Given a project, determine what other type of data are needed
▪ Describe the benefits/risks of using different types of data (observational,
retrospective/historical, experimental)
▪ Describe different sampling strategies for collecting data
Section 2: Exploratory Data Analysis: 20%
Describe data using descriptive statistics and graphical summaries
• Describe measures and graphs for summarizing data
o Interpret summary statistics
o Generate and interpret: histograms, box plots, comparative box plots, run charts,
and Pareto plots
▪ Describe when to use each type of plot
• Distinguish between population vs sample, distinguish between parameter vs statistic
• Recognize the importance of probability distributions
o Describe characteristics of the normal distribution.
o Explain the central limit theorem
o Use a normal quantile plot to determine if the underlying distribution is
approximately normal.
Visualize and explore data
• View relationships between variables
o using scatterplots
o using box plots
o using bar charts
• Create graphs to visualize many variables at a time
o using heat maps
o using multivari (variability) charts
o using trellis plots and overlay variables
o using geographic maps
• Describe the characteristics of a good data visualization
• Explore subsets of your data using data filtering
• Communicate the message in your data
• Identify your target audience
• Translate analytic results into practical/actionable results
• Determine the best visualization to tell a story
• Given a particular visualization, describe the story in the data in the simplest terms
2
Exam Content Guide
Save and share results
• Produce high quality graphics
o Edit and customize graphs
• Save work for reproducibility and reuse
o Manage and organize your work (datasets and output)
o Save scripts to data tables
• Select the appropriate method to share your work based on your intended audience
o HTML w/data, PowerPoint, scripts, dashboards, copy and paste images with
selection tool
Prepare data for analysis
• Structure data appropriately for analysis
o Describe a data table (observations and variables)
o Determine if data needs to be restructured for analysis
o Describe restructuring techniques (split, stack, concatenate, sort, and join)
• Identify data quality issues such as missing or messy data, outliers, and redundant variables
o Define the possible data quality issues
o Recognize data quality issues
o Determine and recommend actions to take to address data quality issues
• Group and bin data and create new variables
o Describe need for derived variables
o Demonstrate when recoding variables might be helpful
o Create new columns with formulas
3
Exam Content Guide
Section 3: Quality Methods: 20%
Use Statistical Process Control Charts
• Explain how control charts are used
o Describe structure of control charts
▪ Interpret the elements of a control chart
o Explain how to use control charts for process improvement
o Define special and common cause variation
o Describe tampering
• Create simple charts, I and MR, X-bar and R or S, 3-way charts
o Describe differences between I and MR charts and X-bar R or S charts
o Describe how and why to use rational subgrouping
o Use phase variables for exploratory analysis
• Use and interpret tests for special causes
o Interpret control charts and determine if they are stable
o Add tests for special causes to the charts, determine which tests to add
• Use control charts to determine if improvements have been sustained
Assess process capability
• Explain the difference between voice of the process to voice of the customer
o Describe Cp, Cpk, Pp and Ppk
o Define process capability
o Compare and contrast control limits and spec limits
o Compare and contrast capability to stability
• Interpret capability indices
o Check the assumptions for interpreting capability indices
o Relate capability measures to ppm defective
o Interpret graphics and indices for non-normal data
o Interpret graphics and indices for many variables at one time
▪ Goal plot
▪ Process performance graph
o Use capability indices (including stability index and target index) and plots to
determine if a problem is present and the nature of the problem
▪ Instability
▪ Off target
o Determine actions to take based on capability measures
4
Exam Content Guide
Conduct Measurement System Studies
• Define the goal/objective of (Measurement System Analysis), determine language and
define terminology to be used:
o Explain importance of evaluating your measurement system
o Define terms associated with MSAs
▪ repeatability, reproducibility, stability, bias, linearity
o Describe the process of conducting an MSA
o Design an MSA
• Utilize variability (multivari) charts to graphically evaluate the data from your MSA
o Calculate common metrics for evaluating the measurement system
• Interpret variance components output to determine the next actions to take from your MSA
• Interpret Gauge R&R analysis results
• Examine EMP (Evaluate Measurement Process) output
o Interpret the EMP output to determine if your measurement system is fit for use or
if it needs improvement
▪ Using average and range charts
▪ Using parallelism plots
5
Exam Content Guide
Section 4: Decision Making with Data: 15%
Define/describe statistical intervals
• Contrast the different types of intervals: confidence, tolerance, and prediction
• Explain the need for statistical intervals
o Compare point estimation to statistical intervals
• Explain when to use each type of interval
• Interpret and explain each type of interval
• Describe the effect of sample size on statistical intervals
Perform statistical tests
• Compare confidence intervals with hypothesis tests
• Perform a test for one mean
• Perform a test for two or more means
o Perform, interpret and explain multiple comparison tests
• Interpret and explain output from means tests
• Compare statistical significance with practical significance
Determine sample size
• Define statistical power and alpha risk
• Explain the factors that influence power and sample size calculations
• Determine sample size for a given power (and s, and delta and alpha risk)
o Calculate sample size for one and two sample mean problems
• Determine the power for a given sample size (and s, delta, and alpha risk)
6
Exam Content Guide
Section 5: Correlation and Regression: 20%
Be able to define and use correlation
• Use correlation to measure relationships between continuous variables
o Describe the strength of a linear association using correlation. Provide a definition of
correlation.
o Interpret a correlation matrix and scatterplot matrix
o Interpret a density ellipse
• Describe common errors in interpreting correlations
o Explain the difference between correlation and causation.
▪ Lurking variables
• Describe the impact of outliers on correlation.
• Explain why correlation does not apply when there is curvature.
Define simple linear regression
• Explain the difference between correlation and regression.
• Explain and interpret the coefficients in the regression model.
• Describe the method of least squares.
• Evaluate regression model assumptions.
o List the model assumptions and use a residual analysis to determine if model
assumptions are met.
o Explain what might be causing particular patterns in residual plots (curvature,
heteroscedasticity, outliers).
• Interpret regression analysis results.
o Perform a regression analysis.
o Determine whether the regression model is significant.
o Interpret the slope coefficient.
o Interpret the confidence interval for the slope coefficient.
• Fit regression models with curvature
o Fit a model with a quadratic effect and interpret the results.
Define/Explain Multiple Linear Regression
• Fit multiple linear regression models
o Describe when multiple linear regression is appropriate.
o Explain the difference between SLR and MLR.
• Compare/contrast the goals of explanatory and predictive modeling.
• Interpret the multiple linear regression model
o Interpret the coefficients in a regression model (continuous and categorical)
o In JMP, use the prediction profiler to interpret model coefficients and interactions.
7
Exam Content Guide
• Be able to interpret the output of a multiple regression model
• Verify model assumptions
o Conduct a residual analysis and determine if regression assumptions have been met.
• Identify influential observations
o Describe the impact of influential observations
o Interpret Cook’s D values
• Explain/describe/define collinearity
o Interpret the VIF
o Describe the impact of collinearity
• Perform variable selection
o Define variable selection and explain the need for variable selection
o List methods of model variable selection (forward, backward, mixed)
o Use the Effect Summary table in JMP to reduce the model
Define/explain logistic regression
• Explain the difference between logistic regression and multiple linear regression.
o Describe when logistic regression is appropriate.
o Explain how logistic regression works.
• Fit a logistic regression model and interpret results.
o Interpret the coefficients in a logistic regression model (continuous and categorical)
o Interpret a logistic curve
o Use the prediction profiler to interpret model coefficients and understand
interaction effects.
• Identify potential problems with the logistic regression model such as:
o Separation
o Not enough data
8
Exam Content Guide
Section 6: Design of Experiments: 10%
Define DOE and compare it to OFAT
• Determine the appropriate circumstances for using DOE.
o Explain the difference between DOE and OFAT
• Explain why you need to do a DOE.
• Describe the importance of a well-designed experiment.
Discuss best practices and considerations to take when running an experiment
• Determine what you want to measure.
• Explain the importance of randomization and the consequences of not doing it
• Explain the trade-offs between budget and the complexity of the experiment
• Determine the goals of your DOE
Analyze the results of an experiment
• Analyze the results of a factorial experiment
o Interpret analysis results
o Identify significant model effects
o Use the Prediction Profiler to explore your model
Design a simple experiment
o Define the following terms: factor, level, treatment, replicate
o Describe a full factorial experiment
o Design a 2^k full factorial experiment in JMP
9
Exam Content Guide
Section 7: Predictive Modeling and Text Mining: 5%
Explain the difference between explanatory modeling versus predictive modeling
o Explain overfitting
o Describe model validation
Interpret results from a predictive model with validation
Interpret results from a multiple linear model with validation.
• Interpret RMSE (RASE)
Interpret results from a logistic model with validation.
• Interpret the misclassification rate (error)
• Interpret the confusion matrix
• Define the following terms: False positive, false negative, true positive, true negative
Interpret decision tree models.
• Describe a decision (partition) tree model
• Describe the difference between a classification tree and a regression tree.
Describe neural network models.
• Describe the basic structure of a neural network (input layer, hidden layer, output layer).
10
