Data Journalism

• A. To replace reporters with automated software
• B. To use data analysis and visualization to find, verify, and tell news stories ✓
• C. To publish only numerical tables without context
• D. To predict future election results with certainty

Why: Data journalism uses data collection, analysis, and visualization to uncover and explain stories, combining traditional reporting skills with quantitative methods.

• A. Publish a chart immediately to attract readers
• B. Check where the data came from and how it was collected ✓
• C. Remove all outliers without investigating them
• D. Convert every value into a percentage

Why: Understanding a dataset's source, collection methods, and limitations is essential to judging its reliability before drawing any conclusions.

• A. It contains only data from government sources
• B. It is large enough to fill a full-page graphic
• C. It reflects the characteristics of the larger population it is meant to describe ✓
• D. It has been reviewed by a senior editor

Why: A representative sample mirrors the population of interest, so conclusions drawn from it can reasonably be generalized. Unrepresentative data can lead to misleading stories.

• A. The color scheme of the accompanying chart
• B. The absolute numbers and population size, since small counts can exaggerate change ✓
• C. Whether the reporter has won awards before
• D. The font used in the headline

Why: Percentage changes based on very small numbers can sound dramatic but be statistically meaningless. Checking absolute counts and the base population gives proper context.

• A. Hiding the methodology to protect a scoop
• B. Publishing the data source and explaining how the analysis was done ✓
• C. Rounding numbers so readers cannot check them
• D. Only sharing findings with the newsroom's sponsors

Why: Transparency about sources and methods lets readers and other journalists verify the work, building credibility and allowing errors to be caught and corrected.

• A. A social media post summarizing the figures
• B. The national statistics agency's published data portal ✓
• C. An anonymous forum comment citing numbers
• D. A blog post that paraphrases a news article

Why: Official statistics agencies publish primary data with documented methodology, making them the most authoritative and verifiable source. Secondary sources like blogs or social posts may introduce errors or distortions.

• A. The raw numerical values in the dataset
• B. Information describing the dataset, such as its source, collection date, and variable definitions ✓
• C. A visualization created from the dataset
• D. The software used to analyze the dataset

Why: Metadata is data about data. It documents where the dataset came from, when it was collected, how variables are defined, and any known limitations, all of which are essential for assessing its reliability.

• A. To find out how many people downloaded it
• B. To confirm they are using the most current version and understand any revisions ✓
• C. To change the color scheme of the charts
• D. To increase the file size for faster loading

Why: Datasets are often revised or corrected over time. Using an outdated version can lead to inaccurate conclusions, so checking the update history ensures the analysis reflects the latest verified data.

• A. A list of spelling errors found in the data
• B. A file explaining each variable's name, meaning, units, and possible values ✓
• C. A translation tool for converting files between formats
• D. A summary of the journalist's reporting notes

Why: A data dictionary defines every field in a dataset, including what it measures, its units, and allowed values. It helps journalists interpret the data correctly and avoid misreading columns.

• A. Whether the file uses a popular spreadsheet format
• B. Whether the company has a potential conflict of interest in how the data was collected or presented ✓
• C. Whether the dataset has more than ten columns
• D. Whether the company's logo appears on the download page

Why: Companies may have commercial incentives to present data favorably or selectively. Journalists should assess potential bias, the collection methodology, and whether the data can be independently verified.

• A. Publishing the raw data alongside the story
• B. Inspecting the data to identify missing values, duplicates, and inconsistent formats ✓
• C. Creating charts before checking the data's structure
• D. Merging the dataset with other sources immediately

Why: Inspecting the data first reveals problems such as missing values, duplicate rows, and inconsistent formatting, which must be understood before any analysis or merging can be trusted.

• A. Delete all rows that do not match the first format
• B. Leave the column as is, since spreadsheets will sort it automatically
• C. Convert all entries to a single standard date format ✓
• D. Convert the dates to text so they are easier to read

Why: Mixed date formats cause errors in sorting, filtering, and time-based analysis. Converting everything to one standard format, such as ISO 8601 (YYYY-MM-DD), ensures consistency. Note that '03/04/2021' can be ambiguous (March 4 or April 3), so the source should be checked.

• A. Two different records that share the same column name
• B. A record that has been saved in two different file formats
• C. Two or more rows representing the same entity or event ✓
• D. A column that contains the same value in every row

Why: Duplicate records are multiple rows describing the same entity or event. They can inflate counts and totals, so identifying and removing them is an important cleaning step.

• A. Documentation is required only for academic research, not journalism
• B. It allows readers and editors to verify the analysis and reproduce the findings ✓
• C. It makes the raw data file smaller and easier to share
• D. It replaces the need to check the original source

Why: Transparent documentation of cleaning decisions, such as how missing values were handled or which rows were removed, lets others verify and reproduce the story's findings, which builds credibility and accountability.

• A. Keep the value, since it is part of the original data
• B. Average it with the other salaries to reduce its impact
• C. Replace it with zero automatically
• D. Investigate whether it is a data entry error or a placeholder code before deciding how to handle it ✓

Why: Values like 9999999 are often placeholders or typos. Investigating the cause before acting prevents incorrect removal or inclusion of data, and may require checking the original source or contacting the data provider.

• A. To publish final findings to the public as quickly as possible
• B. To understand the structure, patterns, and quality of a dataset before drawing conclusions
• C. To design the visual layout of a published article
• D. To confirm a story angle decided in advance by the editor

• A. Create a headline based on the most dramatic number
• B. Check the column names, data types, and summary statistics to understand what the data contains
• C. Send the raw file to readers so they can explore it
• D. Delete all rows with missing values without looking at them

• A. Missing values always indicate fraud in government data
• B. Missing values can bias results and lead to misleading conclusions if not understood
• C. Missing values must be filled with zeros so charts display correctly
• D. Missing values only matter when working with text data

• A. The maximum value
• B. The median
• C. The total sum
• D. The number of unique entries

• A. Publish immediately, since averages are always accurate
• B. Check whether a few extreme cases or a change in how data was recorded explain the increase
• C. Replace the average with the highest single value in the dataset
• D. Ignore the time dimension and report only the current year

• A. Pie chart
• B. Line chart ✓
• C. Scatter plot
• D. Choropleth map

Why: Line charts connect data points in chronological order, making trends and changes over time easy to see. Pie charts show parts of a whole at one moment, and scatter plots and maps serve different purposes.

• A. To make the chart look more colorful
• B. To reduce the number of gridlines needed
• C. To avoid visually exaggerating differences between bars ✓
• D. To match the layout of line charts

Why: Bar length is interpreted as a quantity. If the axis starts above zero, small differences look dramatic, which can mislead readers about the true size of the differences.

• A. Using many bright colors with no defined meaning
• B. Using color only to decorate the chart background
• C. Using a single hue to highlight the key data point against muted gray bars ✓
• D. Using red and green together as the only way to distinguish categories

Why: Highlighting one element with a strong color while muting the rest directs attention to the story. Decorative color adds clutter, and red-green pairings are hard for colorblind readers to distinguish.

• A. The composition of a total broken into parts
• B. The relationship or correlation between two numerical variables ✓
• C. The ranking of categories from highest to lowest
• D. The geographic distribution of events

Why: Scatter plots place each observation at the intersection of two numeric values, revealing patterns such as positive or negative correlation, clusters, and outliers.

• A. It is required only for printed charts
• B. It allows readers to assess credibility and check or verify the data ✓
• C. It makes the chart load faster online
• D. It replaces the need for a headline

Why: Sourcing and dating establish trust and transparency, letting readers judge the reliability of the data and understand how current it is. This is a core ethical practice in data journalism.

• A. Checking whether the chart colors match the newsroom's style guide
• B. Tracing the data back to its original collector and reviewing its methodology
• C. Rounding all figures to the nearest whole number for readability
• D. Asking a colleague whether the headline sounds interesting

• A. Publishing the change immediately because the numbers are official
• B. Checking whether the police department changed how it classifies or records crimes
• C. Averaging the two years together to smooth out the difference
• D. Removing the older year from the story to avoid confusion

• A. It makes the article longer, which improves search engine ranking
• B. It allows readers and other experts to understand, check, and reproduce the analysis
• C. It is required by law for all news articles that contain numbers
• D. It replaces the need to verify the underlying data

• A. Report only the 200% figure, since it is mathematically correct and eye-catching
• B. State the raw counts (1 to 3) alongside the percentage so readers can judge the scale
• C. Replace the percentage with the word 'huge' to avoid numerical confusion
• D. Omit the change entirely because small numbers are unreliable

• A. Deleting the raw data file once the story goes live to protect privacy in all cases
• B. Keeping the dataset, code, and notes accessible and updating the story if errors are found
• C. Assuming the story is final and never revisiting it, since the data was already checked
• D. Sharing only the final chart and discarding the spreadsheet used to build it