Data-Driven Decision Making (C207) Exam Topics and Questions

This topic sets the conceptual foundation the rest of the course builds on. It is less about calculation and more about framing: knowing what kind of question you are asking, what kind of data you have, and what category of error can corrupt your answer before you even start. Get these distinctions sharp early, because exam items will test them in contexts where a fuzzy understanding costs you the point.

– What analysis is and how decisions get structured

The distinction between analytics and quantitative analysis is narrower than it sounds, and the exam tests whether you know where one ends and the other begins. More consequential is the four-type taxonomy: descriptive, diagnostic, predictive, and prescriptive. Each answers a different question, and items will put you in a scenario and ask which type applies. Confusing predictive with prescriptive is the most common slip.

Davenport and Kim's three-stage model (frame, solve, communicate) matters because the exam can place a scenario at any one of the three stages and ask what a manager should do next. The framing stage is where most candidates underestimate the difficulty: it is where you define what problem you are actually solving, and getting it wrong makes every subsequent step irrelevant.

– Data types and measurement scales

Scales of measurement (nominal, ordinal, interval, ratio) determine which statistical operations are legitimate. The exam tests this by giving you a variable and asking what you can do with it. The trap is ordinal data: you can rank it, but the gaps between ranks are not equal, so averaging ordinal scores is technically wrong. Know where interval differs from ratio (a true zero) and why that matters for multiplication and division.

Discrete versus continuous data is a related distinction that affects which probability distribution applies. Discrete data is counted; continuous data is measured. The practical exam consequence is that mixing them up leads you to the wrong distributional assumption and the wrong tool.

Data errors complete this cluster. Random errors are unpredictable and average out; systematic errors do not and introduce bias. Omission errors distort the sample. Outliers can be legitimate data points or the sign of a measurement problem. The exam expects you to diagnose which type of error a described scenario contains.

– Data in organizational context

Business intelligence is not just software: it is the organizational practice of turning data into decisions. The exam connects BI to strategy by asking how data informs direction, not just operations. Know that BI sits at the intersection of data infrastructure and managerial judgment.

Results-Based Management (RBM) gives that judgment a structure. The four stages, inputs, activities, outputs, and outcomes, form a logic chain. Inputs are resources consumed; activities are what is done with them; outputs are the immediate products; outcomes are the changes that result. The exam distinguishes outputs from outcomes, because managers frequently conflate them. An output is what you delivered; an outcome is what changed because of it. A training session is an output. Improved job performance is an outcome. Items will test whether you can tell the difference when both appear in the same scenario.

How The Case for Quantitative Analysis is tested

Items from this topic tend to be scenario-classification tasks. You read a business situation and identify which type of analytics applies, which scale of measurement fits a described variable, or which error type explains a described data problem. The demand is precision: the options are designed so that two of them are plausible if your definitions are approximate.

The Davenport and Kim model appears in items that describe a manager mid-process and ask what stage they are in or what they should do next. The trap is assuming the stages are always sequential in an obvious way; the exam sometimes presents a scenario where a team has skipped framing and asks you to recognize the consequence.

RBM items reward the ability to classify a described element as input, activity, output, or outcome. Candidates lose marks by treating any tangible deliverable as an output when the scenario describes a downstream change, which is an outcome. Read for what changed, not just what was produced.

Where the first topic asked you to classify and frame, this one asks you to apply. The statistical tools here are not tested at a mathematician's depth, but the exam does expect you to know which tool fits which situation, read outputs correctly, and avoid the reasoning errors that show up when managers misuse statistics. The breadth is wide, and the traps are specific.

– Hypothesis testing and significance

The null hypothesis is the default position: no effect, no difference. The alternative hypothesis is what you are trying to support. Items test whether you can identify which is which in a described study and what a given p-value means for the decision. The p-value trap is common: a small p-value means the result is unlikely under the null, not that the effect is large or practically meaningful. Statistical significance and practical significance are different things, and the exam exploits that gap.

The t-test cluster covers one-sample and two-sample versions. The one-sample t-test compares a sample mean to a known value; the two-sample t-test compares two group means. ANOVA extends that logic to three or more groups. The exam will give you a scenario with a specified number of groups and ask which test is appropriate. Choosing a t-test when ANOVA is needed, or vice versa, is the most common error here.

– Probability, distributions, and decision trees

Probability concepts underpin everything in this cluster. The normal distribution and Z-scores let you express where a value sits relative to a mean in standard-deviation units. The exam tests whether you can interpret a Z-score in context, not just compute one. Know that the standard normal distribution has a mean of zero and a standard deviation of one, and understand what that implies about probabilities in the tails.

Decision tree analysis is tested at the managerial level: constructing a tree from a described scenario, assigning probabilities and payoffs to branches, and computing expected values. The exam expects you to read a tree and identify the preferred branch, which requires multiplying probabilities by outcomes and summing across branches. The trap is ignoring a branch or misreading which node is a decision node versus a chance node.

– Regression, visualization, and the causation trap

Linear regression in this course is conducted in Excel. The exam tests interpretation more than computation: what does R-squared tell you, what does the slope coefficient mean, and what does a scatterplot with a trendline reveal about the relationship? R-squared is the proportion of variance in the dependent variable explained by the independent variable. A high R-squared does not mean the relationship is causal.

That last point connects to the correlation-versus-causation distinction, which the exam tests directly. Two variables can move together for reasons that have nothing to do with one causing the other. Items will describe a correlation and ask what conclusion is warranted; the strongest response limits the claim to association.

Scatter plots and histograms are tested as interpretation tasks. A histogram shows the distribution of a single variable; a scatter plot shows the relationship between two. Misidentifying which visualization fits which analytical purpose is a reliable way to lose a point.

How Statistics as a Managerial Tool is tested

This topic generates the highest volume of scenario-based items because every tool has a context in which it applies and several in which it does not. Items reward selecting the right tool for the described situation, reading the output of that tool correctly, and drawing only the conclusions the data supports.

Hypothesis testing items frequently hinge on p-value interpretation. Candidates who treat a significant p-value as proof of a large or causal effect choose the wrong option. The strongest response acknowledges what the p-value actually licenses: a conclusion about the null, not about magnitude.

Regression items reward reading R-squared and slope values in plain business language. Decision tree items reward computing expected values correctly and identifying the optimal branch. The consistent trap across all of this topic is over-claiming: drawing a stronger conclusion than the statistic supports. The exam is built around that trap. Read the scenario, identify what the statistic can and cannot establish, and let that boundary guide your choice.

The tools in this topic extend the statistical toolkit from the previous one into areas that carry specific procedural rules. Confidence intervals, chi-square tests, and experimental design each have conditions that must be met before they apply, and the exam tests those conditions as much as it tests the outputs. Sampling and forecasting add a temporal and design dimension that earlier topics did not require.

– Inference, association, and experimental control

A confidence interval gives you a range, not a point. The exam tests whether you can interpret what a 95% confidence interval actually means: if you repeated the sampling process many times, 95% of the intervals constructed that way would contain the true population parameter. The trap is reading it as a probability that the true value is in this particular interval. That framing is subtly wrong, and the exam knows it.

Chi-square tests assess whether an observed distribution of counts differs from what you would expect, or whether two categorical variables are associated. The exam distinguishes the goodness-of-fit version from the test of independence. Know which question each answers.

Experimental design items test blinding and double-blinding as controls for bias. Single blinding keeps participants unaware of their group assignment; double blinding extends that to the researchers. Control groups establish the baseline. Items will describe a study design and ask whether a specific bias has been adequately controlled. The trap is assuming that any form of blinding eliminates all bias.

– Sampling, forecasting, and financial thresholds

Sampling methods determine whether your data represents the population you care about. Sampling bias enters when the selection process systematically excludes or overrepresents certain groups. The exam tests your ability to identify the bias in a described sampling approach, not just name the sampling method.

Time-series analysis applies when data is ordered by time and you want to identify trends, seasonality, or cycles. Forecasting extends that pattern forward. Items will describe a data set and ask which analytical approach fits; the signal is always whether time is the organizing variable.

Sensitivity analysis tests what happens to a conclusion when one input assumption changes. It is the tool for understanding how fragile a recommendation is. Break-even analysis identifies the point at which revenue equals total cost. Both are tested in business scenario form: given a described situation, what does the analysis tell the manager? The exam rewards interpreting the result in decision terms, not just computing the number.

How More Statistical Tools is tested

Items from this topic most often test the conditions under which a tool applies and what its output means in a business context. Confidence interval items reward precise language about what the interval licenses you to say. Chi-square items reward distinguishing between the two main variants and knowing which one addresses the question described.

Experimental design items are traps for candidates who treat blinding as a binary: either you have it or you do not. The exam will describe a partially controlled study and ask what bias remains. The strongest response names the specific uncontrolled source.

Sampling bias items reward the ability to look at a described recruitment or data-collection method and identify the structural reason certain groups are excluded. Break-even and sensitivity analysis items reward translating a numerical result into a managerial decision. Candidates who stop at computing the number and do not interpret it for the decision at hand are leaving points behind. The topic as a whole rewards knowing not just what a tool produces, but what you are and are not entitled to conclude from it.

Quality is where statistical thinking meets operational practice. The tools in this topic are not abstract: they are used to decide whether a process is in control, where defects originate, and which problems to fix first. The exam tests whether you can select the right tool for a described quality problem and read its output correctly, not whether you can build one from scratch.

– Process control and improvement methodologies

Quality assurance and quality control are not interchangeable. QA is proactive: it builds quality into the process. QC is reactive: it inspects output to catch defects. The exam will place a described activity in one category and ask you to identify it. Getting this backwards is a reliable way to lose the point.

Statistical Process Control uses control charts to distinguish variation that is inherent to a process (common cause) from variation that signals something has changed (special cause). Items will describe a data point on a chart and ask what action is warranted. The trap is overreacting to common-cause variation by adjusting a process that does not need adjustment.

Six Sigma's DMAIC methodology (Define, Measure, Analyze, Improve, Control) is tested stage by stage. Items will describe what a team is doing and ask which DMAIC phase it represents. Total Quality Management is broader: it is an organizational philosophy, not a project method. The exam distinguishes between using a tool within DMAIC and embedding quality as a cultural commitment under TQM.

– Diagnosing and prioritizing quality problems

The Ishikawa diagram (also called a fishbone or cause-and-effect diagram) structures the investigation of why a defect occurs. The categories (machine, method, material, measurement, environment, people) give teams a framework for brainstorming root causes. The exam tests whether you know what the diagram is for (root cause analysis) and what it does not do (quantify how often each cause occurs).

Pareto charts rank problems by frequency or impact, and the 80/20 rule says roughly 80% of defects come from 20% of causes. The exam uses Pareto in prioritization scenarios: given a list of defect types and their frequencies, which should the team address first? The trap is choosing the most dramatic-sounding problem rather than the one with the highest cumulative impact.

Process capability analysis measures whether a process can consistently produce output within specification limits. The key indices (Cp and Cpk) compare process spread to specification width. The exam tests interpretation: a Cpk below 1 means the process is producing out-of-spec output. Know what that threshold means for a managerial decision.

– Performance standards and measurement

Benchmarking compares your process or output to a reference standard, which can be internal (comparing departments), competitive (comparing to rivals), or best-in-class (comparing to the best performer regardless of industry). The exam will describe a comparison and ask which type of benchmarking it represents.

KPIs are the quantified targets that tell an organization whether it is moving in the right direction. Dashboards aggregate KPIs into a visual display for decision makers. The exam tests the relationship between KPIs and strategy: a KPI should be directly connected to an objective that matters, not just easy to measure. Items will describe a KPI and ask whether it is well-aligned to a stated goal. The trap is accepting any measurable metric as a good KPI. The exam rewards the distinction between measuring what is convenient and measuring what is consequential.

How Quality Metrics and Tools is tested

Items from this topic are heavily scenario-based. You will read a description of a quality problem or a team activity and be asked to identify the tool, the phase, or the interpretation that fits. The demand is applied recognition, not recall of definitions.

Control chart items reward knowing the difference between common-cause and special-cause variation and what action each warrants. DMAIC items reward placing a described activity in the right phase. The trap there is confusing Analyze with Measure: measurement collects data, analysis interprets it.

Pareto items reward prioritization logic. The strongest response identifies the cause that accounts for the largest cumulative share of defects, not the one that sounds most urgent. KPI items reward alignment thinking: the strongest response connects the metric to the strategic objective it is supposed to reflect.

Process capability items reward reading Cp and Cpk values and translating them into a managerial conclusion. Candidates who can compute the index but cannot say what it means for the decision are not fully prepared for this topic.

The previous topics gave you the tools; this one tests whether you can use them in the messy, context-laden situations managers actually face. The exam here is less about whether you know a definition and more about whether you can recognize when a tool is being misused, when data governance has failed, or when a visualization is misleading. The applied judgment demand is higher than in any earlier topic.

– Business intelligence and knowledge infrastructure

Business intelligence systems turn raw organizational data into information that supports decisions. The exam tests the difference between a BI system and the decisions it informs: the system does not make the decision, it structures the evidence. Knowledge management systems capture and distribute organizational expertise, including tacit knowledge that is not easily stored in a database. Items will describe a system and ask which category it belongs to or what gap it addresses.

Big Data is characterized by volume (the amount of data), velocity (the speed at which it arrives), and variety (the range of data types and sources). The exam tests whether you can identify which characteristic a described situation illustrates. The trap is treating volume as the defining feature when the scenario is really about the speed of incoming data streams or the challenge of integrating structured and unstructured sources.

– Visualization, governance, and integrity

Data visualization best practices are tested as a judgment task: given a described data set and audience, which visualization type communicates the right thing? The exam also tests what makes a visualization misleading: truncated axes, inappropriate chart types, and visual encodings that distort relative magnitude.

Data governance is the set of policies and accountabilities that determine who can access, modify, and use organizational data. Data integrity means the data accurately reflects what it is supposed to represent. The exam connects both to decision quality: a decision made on data that has not been governed or validated is only as good as the data. Items will describe a governance failure and ask what risk it creates for the organization's decisions.

– Performance frameworks and ethical boundaries

The Balanced Scorecard organizes performance measurement across four perspectives: financial, customer, internal process, and learning and growth. In this topic it appears as an enterprise-level framework for connecting strategy to measurement. Items will describe a metric and ask which perspective it belongs to, or describe a strategic objective and ask which perspective it reflects.

Interpreting decision trees and regression in real-world contexts means reading the output and saying what it implies for a business decision, not just reporting the number. The exam tests whether you can move from the statistic to the recommendation.

Ethical use of data covers privacy, consent, and the boundaries of what data should be used for even when using it is technically possible. Items will describe a data practice and ask whether it is appropriate. The trap is assuming that legal compliance equals ethical acceptability. The exam rewards the distinction.

How Real World Data-Driven Decisions is tested

This topic's items are the most integrative in the exam. They draw on tools and concepts from earlier topics and place them in realistic business scenarios where the analytical task and the ethical or governance dimension appear together. The demand is judgment, not just technique.

Visualization items reward recognizing misleading design choices and knowing which chart type fits the data and the audience. Big Data items reward classifying a described challenge by the correct characteristic. Governance items reward tracing a described failure to its consequence for decision quality.

Balanced Scorecard items reward perspective classification and the connection between a metric and the strategy it is supposed to serve. Ethical items reward drawing the line between what data can do and what it should do in the described context. Candidates who treat this topic as a review of earlier material miss the applied judgment layer. The exam is testing whether you can think like a manager who has internalized these tools, not just a student who can recall them.

This final topic pulls together the measurement and analytical tools from across the course and asks how they drive sustained organizational improvement. The shift is from knowing how to analyze to knowing how to act on analysis over time. Performance frameworks, continuous improvement methods, and workforce analytics all appear here, and the WGU-specific context of competency-based evaluation adds a dimension unique to this exam.

– Improvement methodologies and performance frameworks

Continuous improvement methodologies are tested as a contrast pair. Kaizen is an incremental, ongoing approach to improvement embedded in daily work. Lean focuses on eliminating waste from processes to improve flow and reduce cost. The exam will describe an organizational initiative and ask which methodology it reflects. The trap is treating them as interchangeable because both aim at improvement: the distinguishing factor is what each targets and at what pace.

Performance management frameworks give organizations a structure for setting targets, measuring results, and taking corrective action. The exam tests whether you can identify what a described framework is doing and whether it is well-designed. A framework that measures what is easy rather than what matters is a common scenario. The Balanced Scorecard reappears here in its strategic execution role, connecting the four perspectives to long-term organizational goals rather than just reporting metrics.

– Analytics, workforce data, and strategic decisions

Using analytics to drive strategic decisions is the applied synthesis of the entire course. Items will describe a business situation with available data and ask what analytical approach supports the decision. The exam rewards selecting the approach that fits the decision type: a forecasting problem calls for time-series tools, a root-cause problem calls for diagnostic analytics, a resource-allocation problem may call for decision tree analysis.

Workforce analytics applies the quantitative methods from earlier topics to HR data: turnover rates, time-to-hire, training effectiveness, and productivity metrics. Items will describe an HR challenge and ask which metric or analytical approach addresses it. The trap is selecting a metric that sounds relevant but does not directly measure the outcome the organization cares about.

Criterion-referenced assessment and competency benchmarking evaluate performance against a defined standard rather than against other test-takers. WGU's competency-based model is a direct application: passing means demonstrating mastery, not outperforming peers. The exam may test this distinction in the context of assessment design.

How Improving Organizational Performance is tested

Items in this topic are the most synthesis-heavy in the exam. They assume you can apply tools from earlier topics and ask you to evaluate whether they are being used correctly in an organizational improvement context. The demand is evaluative, not just applicative.

Kaizen versus Lean items reward identifying the distinguishing feature of each in a described scenario, not just knowing both exist. Balanced Scorecard items here test strategic alignment: does the described set of metrics actually cover all four perspectives, and does each metric connect to the right one?

Workforce analytics items reward selecting the metric that directly addresses the described organizational problem. The trap is choosing a metric that is related to the problem but measures a proxy rather than the outcome itself.

The WGU competency-based context appears in items about criterion-referenced testing. These items reward understanding that the standard is external and fixed, not norm-referenced. Candidates who conflate criterion-referenced with norm-referenced assessment lose those points. Finishing strong on this topic means carrying the analytical precision of the earlier topics into organizational and strategic territory.