Overview
Data Analysis is where prepared data becomes an answer. Objective 3.1 is about communication: turning requirements into an approach that fits the audience, using mock-ups, accessibility, the right level of detail, internal versus external and sensitive versus non-sensitive handling, user personas and key performance indicators (KPIs). Objective 3.2 covers choosing among descriptive, inferential, predictive and prescriptive methods and using functions and measures, including central tendency, dispersion and logical, date and string functions. Objective 3.3 covers troubleshooting connectivity, user-reported, basic SQL and corrupted-data issues with logging, source validation and vendor communities.
The domain exists because a correct number delivered the wrong way, or computed with the wrong method, does not support a decision. It consumes the clean data from Domain 2 and feeds the visual and reporting choices in Domain 4.
How it's tested
Expect requirement-driven and 'given a scenario' items. A stem describes a stakeholder, a data distribution or a symptom and asks for the BEST communication approach, the right statistical method or measure, or the first troubleshooting step. The right answer fits the stated audience, data shape or symptom, not the most sophisticated technique available.
-
Fit the message to the audience
A stakeholder type, sensitivity level or accessibility need is described and you choose the level of detail, format, KPI set or mock-up step that fits.
-
Select the method or measure
A business question or distribution is given and you choose descriptive, inferential, predictive or prescriptive analysis, or the right measure such as median, standard deviation or a date function.
-
Troubleshoot in order
A failing connection, a user complaint, a SQL error or corrupted data is described and you choose the first or most effective step, such as enabling logging or validating the source.
A stem says executives want a monthly view of whether customer churn is on target, and an analyst proposes a 40-page model appendix. Options include adding more statistical tests, a KPI summary against target with a brief explanation, a raw data extract, and a self-built pivot table. The executive audience and the decision they make point to a concise KPI summary; the technical detail belongs in an appendix for analysts.
Objectives 3
Official objectives from the CompTIA outline, in outline order.
3.1 Given a set of requirements, determine the appropriate communication approach for data analysis.Mock-ups, accessibility, audience, detail, sensitivity, personas and KPIs.
- What it means
- Before building outputs, the analyst matches the approach to the requirements: a mock-up to agree layout early, accessible design, technical or nontechnical language, the level of detail, internal or external distribution, sensitive or non-sensitive content, the user persona's goals, and which key performance indicators (KPIs) measure success.
- What to understand
- Executives need a few KPIs tied to targets and the 'so what'; analysts and engineers need method, assumptions and drill-down. External or sensitive audiences need aggregated or masked data and approved content only. A mock-up confirmed with stakeholders before development is the cheapest way to catch a misunderstood requirement.
- How it shows up
- A requirement names an audience, a sensitivity level or an accessibility need, and you choose the approach that fits, such as a summary for the board or a mock-up review before building.
- Common confusion
- Giving every audience the same detailed output, or treating accessibility as only a color issue when it also covers alt text, contrast, font size and screen-reader compatibility.
3.2 Given a scenario, select the appropriate statistical method or function.Descriptive, inferential, predictive and prescriptive methods; measures and functions.
- What it means
- This covers the four basic statistical methods, mathematical measures of central tendency (mean, median, mode) and dispersion (range, variance, standard deviation, interquartile range), and logical, date and string functions used in calculations.
- What to understand
- Descriptive statistics summarize what happened; inferential statistics draw conclusions about a population from a sample, with confidence intervals and hypothesis tests; predictive analysis forecasts what is likely; prescriptive analysis recommends an action. For skewed data the median represents the typical value better than the mean. Logical functions (IF, CASE), date functions (date difference, extracting month) and string functions (concatenate, trim, substring) build the fields the analysis needs.
- How it shows up
- A question names the business goal or the data shape and you pick the method or measure, such as inferential analysis for a sample survey or the median for home prices.
- Common confusion
- Calling a forecast prescriptive, or describing a sample survey result as if it were exact for the whole population.
3.3 Given a scenario, troubleshoot basic issues using the appropriate tool or method.Connectivity, user-reported, SQL and corrupted-data issues; logging, source validation, vendor resources.
- What it means
- Analysts troubleshoot four issue types: connectivity-related failures, problems users report, errors in basic SQL code, and corrupted data, using methods such as enabling logging, validating the data source and consulting vendor communities or online resources.
- What to understand
- Work from evidence: reproduce the issue, enable or read logs to see the actual error, then validate the data source by comparing counts or values with the system of record. Connectivity issues often trace to credentials, expired passwords or tokens, firewalls, or a changed server name. SQL issues often come from syntax, ambiguous column names, wrong join conditions or data type mismatches.
- How it shows up
- A symptom such as 'refresh fails overnight' or 'numbers differ from the finance system' is given and you choose the first or most useful step.
- Common confusion
- Rewriting the query or rebuilding the report before reading the error log or confirming the source data is correct.
High-yield concepts
The terms, rules and formulas this domain turns into questions most often.
-
Descriptive vs. inferential statistics
Descriptive statistics summarize the data you have, such as last quarter's average order value; inferential statistics use a sample to estimate or test something about a larger population, with stated uncertainty.
- Remember
- Sample plus a claim about everyone means inferential.
- Tested as
- Method-selection questions turn on whether the data is the full population or a sample.
- Trap
- Reporting a survey sample's percentage as an exact population fact with no margin of error.
-
Predictive vs. prescriptive analytics
Predictive analytics estimates what is likely to happen, such as next month's demand; prescriptive analytics recommends what to do, such as the reorder quantity that minimizes cost.
- Remember
- Predictive answers 'what will happen'; prescriptive answers 'what should we do'.
- Tested as
- Scenarios phrase the business question and ask for the method type.
- Trap
- Choosing predictive when the stem asks for a recommended action or optimal choice.
-
Mean, median, mode
The mean is the arithmetic average, the median is the middle value when sorted, and the mode is the most frequent value.
- Remember
- Skewed numeric data such as income or home prices calls for the median; categorical data uses the mode.
- Tested as
- Measure-selection items describe the distribution.
- Trap
- Reporting the mean salary when a few executives' pay drags it far above what most employees earn.
-
Range, variance, standard deviation
Range is maximum minus minimum; variance is the average squared distance from the mean; standard deviation is the square root of variance, in the original units.
- Remember
- Standard deviation is the spread measure you can read in the data's own units.
- Tested as
- Questions ask which measure shows consistency or volatility.
- Trap
- Using range, which one extreme value can distort, as the main measure of spread.
-
Interquartile range (IQR)
The IQR is the third quartile minus the first quartile, the spread of the middle 50% of values, and it is resistant to outliers.
- Remember
- Pair the median with the IQR for skewed data, and the mean with the standard deviation for symmetric data.
- Tested as
- Measures of dispersion and outlier detection.
- Trap
- Pairing a median with a standard deviation on heavily skewed data.
-
Normal distribution and z-score
In a normal distribution about 68% of values lie within one standard deviation of the mean, 95% within two and 99.7% within three; a z-score expresses how many standard deviations a value sits from the mean.
- Remember
- z = (value minus mean) divided by standard deviation.
- Tested as
- Interpreting spread and flagging unusual values.
- Trap
- Applying the 68-95-99.7 rule to clearly skewed data.
-
Logical functions
Logical functions such as IF, CASE, AND and OR return values based on conditions, for example labeling orders over 500 as 'large'.
- Remember
- Use them to create categories and flags from raw values.
- Tested as
- Function-selection items describe a rule-based field.
- Trap
- Overlapping conditions in a CASE expression, where the first true branch wins and later ones never fire.
-
Date functions
Date functions extract parts of a date, add intervals or compute differences, such as days between order and shipment.
- Remember
- Durations, aging and period grouping are date-function jobs.
- Tested as
- Scenarios needing tenure, aging buckets or month-level grouping.
- Trap
- Subtracting dates stored as text, which fails or sorts alphabetically.
-
String functions
String functions such as concatenate, trim, substring, upper and replace build or clean text fields.
- Remember
- Trim invisible leading and trailing spaces before matching or joining on text keys.
- Tested as
- Function-selection and troubleshooting items where text values fail to match.
- Trap
- Blaming the join when trailing spaces make 'NY' and 'NY ' different keys.
-
Key performance indicator (KPI)
A KPI is a metric tied to a business objective and a target, such as on-time delivery rate against a 95% goal.
- Remember
- A KPI needs a definition, a target and an owner; a raw count without context is just a metric.
- Tested as
- Communication items ask which measures belong on a stakeholder's view.
- Trap
- Filling an executive view with every available metric instead of the few tied to decisions.
-
Mock-up
A mock-up is a draft of the report or dashboard layout, sometimes with sample data, reviewed with stakeholders before full development.
- Remember
- Mock-ups validate requirements early, when change is cheap.
- Tested as
- Requirement scenarios ask what to do before building.
- Trap
- Building the full dashboard first and discovering the layout or KPIs were wrong.
-
Accessibility
Accessible outputs work for people with visual, motor or cognitive impairments through sufficient contrast, colorblind-safe palettes, alt text, readable fonts and keyboard or screen-reader support.
- Remember
- Never encode meaning by color alone; add labels, patterns or icons.
- Tested as
- Communication items mention a user with a visual impairment or a public-sector audience.
- Trap
- Relying on red and green to signal status.
-
Validate the data source
Source validation compares the analysis input with the system of record, using row counts, control totals or spot checks, to confirm the data is complete and current.
- Remember
- When numbers disagree with another system, check the source and its timing before changing logic.
- Tested as
- Troubleshooting items with mismatched totals.
- Trap
- Rewriting calculations when the real cause is a partial or stale extract.
-
Enable logging
Logging records connection attempts, query errors and refresh steps so the actual failure point and message are visible.
- Remember
- Gather evidence before changing anything.
- Tested as
- First-step troubleshooting questions for intermittent or overnight failures.
- Trap
- Reinstalling drivers or rebuilding the report without knowing what failed.
Side by side
Concepts candidates mix up, compared on the point that decides the answer.
Descriptive vs. inferential vs. predictive vs. prescriptive
| Concept | Meaning | Key difference | Common trap |
|---|---|---|---|
| Descriptive | Summarizes what happened in the data you have | No claim beyond the observed data | Presenting a sample summary as a population fact |
| Inferential | Generalizes from a sample to a population with stated uncertainty | Uses confidence intervals and hypothesis tests | Ignoring that results depend on a representative sample |
| Predictive | Estimates future or unknown outcomes | Answers what is likely to happen | Mistaking a forecast for a recommendation |
| Prescriptive | Recommends an action, often by optimization or simulation | Answers what should be done | Choosing it when the stem only asks for a forecast |
Mean vs. median vs. mode
| Concept | Meaning | Key difference | Common trap |
|---|---|---|---|
| Mean | Sum divided by count | Uses every value, so outliers pull it | Reporting it for skewed data such as income |
| Median | Middle value of sorted data | Resistant to outliers and skew | Forgetting to average the two middle values when the count is even |
| Mode | Most frequent value | The only one of the three that works for categories | Assuming there is always exactly one mode |
Technical vs. nontechnical audience
| Concept | Meaning | Key difference | Common trap |
|---|---|---|---|
| Technical audience | Analysts, engineers, data scientists | Wants methods, assumptions, data definitions and drill-down | Hiding the logic so results cannot be checked |
| Nontechnical audience | Executives, operations and business users | Wants KPIs against targets, the takeaway and the recommended action | Leading with methodology and jargon |
Exam traps
Each trap is an answer that looks right. Learn the clue that should change your mind.
-
Most advanced method wins
- What candidates pick
- A predictive model when the stakeholder asked how last quarter performed.
- Why it looks right
- It sounds more valuable and analytical.
- Why it's wrong
- The question is about what happened, which descriptive statistics answer directly and transparently.
Clue to watch for Past-tense phrasing such as 'how did', 'what was' or 'last quarter'.
-
Sample treated as population
- What candidates pick
- State that exactly 72% of all customers are satisfied based on a 300-person survey.
- Why it looks right
- It is the number the survey produced.
- Why it's wrong
- A sample result estimates the population value; inferential methods attach a confidence interval or margin of error.
Clue to watch for The stem mentions a survey, a sample or a subset of the population.
-
Mean for skewed values
- What candidates pick
- The mean home price to describe a typical home in the area.
- Why it looks right
- Average is the default summary measure.
- Why it's wrong
- A few very expensive homes pull the mean upward; the median describes the typical home.
Clue to watch for Words like skewed, long tail, a few very large values, or income, prices and salaries.
-
Detail for executives
- What candidates pick
- Give leadership the full analysis workbook with every breakdown.
- Why it looks right
- More information seems more transparent.
- Why it's wrong
- Executives need a few KPIs against targets and the implication; detail belongs in an appendix or a drill-down for analysts.
Clue to watch for The audience is senior leadership or a nontechnical stakeholder making a decision.
-
Sensitive data in an external report
- What candidates pick
- Share the same customer-level dashboard with an external partner.
- Why it looks right
- It is already built and accurate.
- Why it's wrong
- External and non-sensitive audiences should receive aggregated or masked data only; customer-level details may breach privacy rules.
Clue to watch for External recipients, partners, the public, or personal data in the source.
-
Fix before diagnosing
- What candidates pick
- Rewrite the SQL as soon as a scheduled refresh fails.
- Why it looks right
- The query is the visible part of the process.
- Why it's wrong
- The failure may be credentials, a changed server or a network issue; enabling logging and reading the error comes first.
Clue to watch for An intermittent or overnight failure with no error message captured yet.
Answer framework
Use this sequence when a Domain 3 item asks how to analyze, present or troubleshoot.
- Identify the decision behind the request and who makes it: executive, operational user, technical peer, external party.
- Classify the question: what happened (descriptive), what is true of the population (inferential), what will happen (predictive) or what to do (prescriptive).
- Check the data shape and type before picking a measure: skew favors median and IQR, categories need the mode, dates and text need the matching functions.
- Match format and detail to the audience, sensitivity and accessibility needs, and confirm KPIs and layout with a mock-up before building.
- When something fails, gather evidence first by reproducing it, enabling logging and validating the source, then change only what the evidence points to.
Practice question
Want more like this? The CompTIA Data+ practice test has questions for every domain.
Checklist
0 of 14
Tick each one you can do without notes. Your ticks stay in this browser.
Study strategy
Start with objective 3.2, because statistics questions reward exact understanding and are the easiest to drill. Compute mean, median, mode, range, variance, standard deviation and IQR by hand on a ten-value data set, then add one extreme value and watch which measures move. Write a one-line business question for each of the four method types.
Next, study objective 3.1. Candidates underestimate it because it reads like soft skills, but the questions hinge on precise clues: the audience type, whether data is sensitive, whether recipients are internal or external, and whether a mock-up or accessibility requirement is stated. Practice by taking one analysis and outlining three versions: executive summary, analyst detail and external partner view.
Then cover objective 3.3 by building a habit of evidence-first troubleshooting. Break a database connection on purpose, read the error, and fix it; introduce a SQL ambiguity and a type mismatch; compare an extract with its source using counts and totals.
Afterward, do mixed practice sets, since this is the heaviest domain. Review each miss by asking which clue in the stem you skipped, then move to Domain 4.
FAQ
Is Data Analysis the largest Data+ V2 domain?
Yes. CompTIA weights it at 24% of DA0-002, the highest of the five domains. CompTIA does not publish per-domain question counts.
How much math is on the Data+ exam?
Expect basic statistics: central tendency, dispersion and the meaning of descriptive, inferential, predictive and prescriptive methods. Being able to compute and interpret a mean, median and standard deviation matters more than advanced formulas.
What is the difference between predictive and prescriptive analytics?
Predictive analytics estimates what is likely to happen, such as next month's sales. Prescriptive analytics recommends an action, such as how much stock to order given that forecast.
Why does a communication objective sit in the analysis domain?
Because the outline treats choosing the audience, detail level, sensitivity handling and KPIs as part of analysis, decided before building outputs. Domain 4 then covers the visual elements and delivery methods themselves.
What should I do first when a report's numbers disagree with another system?
Validate the data source: compare row counts, totals and extract timing against the system of record. Many discrepancies come from partial or stale data rather than faulty logic.
How long should I spend on this domain?
Our planning estimate is about 20–30 hours within a 100-hour plan, more if statistics is new to you. Adjust after a diagnostic practice test.
How does this domain connect to Visualization and Reporting?
The audience, KPI and sensitivity choices made here determine which visuals, delivery methods and validation checks you use in Domain 4.
Other Data & Analytics certifications
Where the same subject shows up in other exams. Cards marked "Same subject" open the matching domain.
Test your Domain 3 knowledge
You have just covered method selection, measures of center and spread, audience-fit communication and evidence-first troubleshooting. Test whether you can spot the deciding clue in a scenario.
Keep going
- Training Training Built on your Weak areas
- Practice test Questions by domain, with explanations
- Mock exam Full length, timed like the real exam
- Readiness test Find your weakest domain in minutes
- Cheat sheet Key terms and formulas on one page
- Ask the CompTIA Data+ AI tutor Get any concept explained your way