CompTIA Data+ DA0-002 · Free study guide

Objective 3.2 — Select statistical methods and functions

Statistical analysis is not a contest to use the most advanced technique. Choose a method that matches the question, data, and decision. When several calculations are possible, start by classifying the question, then choose summaries and functions that preserve meaning.

This objective also covers everyday analytical functions. Logical, mathematical, date, and string operations often prepare or classify data before any formal statistic is calculated.

Classify the analytical question

Four analytical categories appear repeatedly.

Descriptive analysis organizes observed data. A dashboard of last month’s sales is descriptive even if it contains sophisticated calculations. Predictive analysis uses patterns and predictors to estimate an unknown or future outcome. Prescriptive analysis goes one step further by recommending an action under goals and constraints. Inferential analysis uses a sample to reason about a larger population while acknowledging sampling uncertainty.

Classify the work being performed, not a product name. A “forecast dashboard” is descriptive if it merely displays forecasts produced elsewhere.

Choose a measure of center

The mean is the sum divided by the number of observations. It uses every value and works well for a roughly symmetric numeric distribution without influential extremes. The median is the middle value after sorting. It resists extreme values and is often a better center for skewed data such as income, home price, or response time. The mode is the most frequent value and can describe numeric or categorical data. A data set can have one mode, several modes, or no mode.

The measure depends on the question. Mean order value may help revenue planning, while median order value may better describe a typical customer. “Typical” does not automatically mean mean.

Describe spread and unusual values

The range is maximum minus minimum. It is simple but depends entirely on two observations. Variance measures the average squared distance from the mean. Squaring prevents positive and negative deviations from cancelling, but it expresses spread in squared units. Standard deviation is the square root of variance, so it returns to the original unit and is easier to interpret.

A small standard deviation indicates clustering near the mean; a large one indicates greater dispersion. Compare only compatible scales and definitions. Standard deviation does not prove normality.

Percentiles locate a value relative to an ordered distribution. The 90th percentile is at or above roughly 90% of observations, not necessarily 90% of the maximum. Quartiles divide ordered values into four regions. The first quartile is the 25th percentile, the second is the median, and the third is the 75th percentile. The interquartile range, or IQR, is Q3 - Q1 and describes the middle half of the data.

A common outlier rule flags values below Q1 - 1.5 × IQR or above Q3 + 1.5 × IQR. A flagged value is a candidate for investigation, not automatic proof of an error. It may be a valid rare event, a different population, or a defect.

Distribution shape affects interpretation. A symmetric distribution has similar shape on either side of its center. A right-skewed distribution has a long high-value tail and often has a mean above its median. A left-skewed distribution has a long low-value tail. Multiple peaks may indicate mixed groups that should be examined separately.

Use logical and mathematical functions

Logical functions turn business rules into repeatable classifications. IF or CASE can label a record based on conditions. AND requires all joined conditions to be true, OR requires at least one, and NOT reverses a condition. Parentheses matter when AND and OR are combined. Null requires explicit handling such as IS NULL; an ordinary equality comparison with null does not evaluate like comparison with a known value.

Common mathematical functions include ABS, ROUND, FLOOR, CEILING, POWER, and SQRT. Aggregates such as SUM, COUNT, MIN, MAX, and AVG operate across rows, while scalar functions operate on one value. Round near the end; repeated intermediate rounding can distort totals.

Work with dates and strings

Date arithmetic supports questions such as days since signup, duration between events, or month of service. Common operations add or subtract an interval, calculate a difference, truncate a timestamp to a period, or extract a part such as year or weekday. Function names differ between SQL dialects, so focus on the operation described. Normalize time zones before comparing timestamps, and distinguish a calendar date from a precise instant.

String functions include TRIM, UPPER, LOWER, LENGTH, SUBSTRING, REPLACE, and concatenation. They standardize casing, remove outer spaces, extract codes, or build labels. Trimming does not correct spelling, and lowercasing does not prove entity identity. Watch nulls and substring-index differences.

Worked scenario: service response times

A support team records response times of 10, 12, 12, 13, 15, and 78 minutes. The mean is approximately 23.3 minutes, the median is 12.5, the mode is 12, and the range is 68. The high value pulls the mean far above most observations, so the median is the more honest description of a typical response. With Q1 = 12 and Q3 = 15, the IQR is 3 and the upper outlier fence is 19.5; 78 should be investigated.

The analyst discovers that the 78-minute case arrived before a regional holiday and is valid. It should not be deleted merely because it is an outlier. A descriptive report can show the median, IQR, and exception count. A predictive model might estimate which new cases risk a late response. A prescriptive process might recommend staffing changes. An inferential study might use a representative sample to estimate the median response time for all cases.

To prepare the records, the analyst uses TRIM and UPPER on region codes, calculates minutes between opened and answered timestamps after time-zone normalization, and uses CASE to label responses above the service target. Each function serves a defined analytical purpose.

Common exam traps

• Choosing the mean for strongly skewed data merely because it uses every value.
• Saying the median is affected more by an extreme value than the mean.
• Treating mode as numeric-only or assuming every data set has exactly one mode.
• Confusing variance with standard deviation; variance uses squared units.
• Interpreting the 90th percentile as 90% of the maximum value.
• Automatically deleting every IQR outlier.
• Calling a recommendation predictive rather than prescriptive.
• Using = NULL instead of explicit null logic.
• Combining AND and OR without parentheses that express the intended rule.
• Comparing timestamps before resolving their time zones.
• Assuming function names and substring indexing are identical in every tool.

Readiness checklist

• I can distinguish descriptive, predictive, prescriptive, and inferential questions.
• I can choose mean, median, or mode based on distribution and purpose.
• I can explain range, variance, standard deviation, percentiles, quartiles, and IQR.
• I can recognize symmetric, skewed, multimodal, and outlier-containing distributions.
• I can use the IQR rule without assuming every flagged value is invalid.
• I can apply logical conditions with correct null handling and precedence.
• I can distinguish scalar mathematical functions from aggregates.
• I can choose date arithmetic and string functions for a stated preparation task.
• I can explain how an analytical function changes the meaning or usability of the result.