Top 10 Rules for Sourcing

1. Track down the original source
Use original sources. It’s okay to read a Wikipedia article – or one published on a blog or major news site, like CNN, Forbes, the New York Times, or the Washington Post – but these are not original sources for the data or information in question.

Rather than taking the information you read on the web for granted, track it down to the organization, company, or government authority that released it. That is your source, not the media outlet that published it.

2. Verify the data
A newspaper, magazine, or blog article could misinterpret data or present it in the wrong context. This is why, again, taking statistics you read in an online article or blog at face value is problematic. Track down the original source, then confirm the data. Make doubly sure it is appropriate for your topic and for the context in which it will appear.

3. Make sure you’re using the most recent data available
Surveys and reports get outdated, fast. A good rule of thumb: if a data set is older than 2010, chances are there is more recent data available. (Even from the government.)

Note: this is especially true with social media data. Quoting Facebook user statistics from 2009 could be like citing population demographics from the 1700s. [We just made that up – don’t quote us on it.]

4. Do not, ever, source information from user-generated content websites
When it comes to sourcing, consider user-generated content websites totally off limits. That includes Wikipedia (never, ever include that in a source list), Yahoo Answers, Google Answers, and even sites like eHow and About.com, where “experts” (of questionable origin) publish text that is rarely sourced.

You can never be 100% sure who wrote the article or answer you are looking at, where they got their information, and whether it’s the best one out there. It could even be completely wrong!

5. 99% of the Web is just your starting point
You can use Wikipedia as a starting point: as long as you click through the resources at the bottom of the article. Follow the data path (i.e. rule #1: track down the original source) and make sure it is the most recent data available (i.e. rule #2: verify the data). The same goes for information you see on blogs or websites: Find out where they got their facts, verify them, and credit the appropriate source.

6. Never use data out of context
Always keep the big picture in mind. Data should be presented in the exact same context intended by the original source. If the survey talks about Facebook users ages 18-34, make sure you mention that. Don’t just say “Facebook users” because that misrepresents the data entirely.

7. Don’t combine data from multiple sources
If you ever combine statistics from different surveys to illustrate a certain trend, you have to make sure that they reference the same time period and demographic. And then be absolutely clear in your copy that the sources are different.

For example, if one source says that 25% of Facebook users drink wine and another says that 45% of Twitter users drink liquor, saying “while 25% of Facebook users prefer wine, a whopping 45% of Twitter users go for hard liquor” is just wrong. You must always be clear about the context (time period, demographic) of the data you use.

8. Limit the number of sources
Source lists that are as long as the infographic itself are unsightly. More importantly, the more sources you use, the higher the risk of breaking one of the rules above. While there isn’t a pre-determined number of sources you should use, three to five is a good target range. Chances are, if you have followed the rules above and have found credible, original sources, each of them (a report or survey, for example) will have plenty of data to choose from.

9. Use proper formatting
A list of lengthy URLs in your source list is far from user-friendly. Many of the projects you’ll work on for Visually will be static (i.e. no active hyperlinks). Who would even try to retype the URL in their browser? Make it easier for people to check your sources by listing them in easily “Google-able” format.

For example, instead of http://www.wired.com/wired/archive/11.09/ppt2.html in your source list, you should write: PowerPoint Is Evil, by Edward Tufte (WIRED, 2003). (Except, of course, this source is too old. Remember: Try to stick to sources no older than 2010.)

Always include as many details (title, author name) as necessary to find the information.

10. Never use other infographics as sources
We can’t stress this enough, so we left it for last. The web has been littered with poorly sourced infographics, many of which violate not just some, but all of our rules. By grabbing a statistic from an infographic, you do the same. You use information from an unreliable source: unconfirmed, out of context, possibly outdated, and maybe even fake.

Just don’t do it. Ever.

SOURCES: The Good, The Bad, & The Ugly

Good:

• Analyst reports
• White papers
• Government documents/ reports
• Company documents filed with government organizations (such as SEC filings)
• Survey data (recent!)
• Academic research (papers, books by professors/ instructors – but NOT student work)
• Corporate research (beware of company bias, seek out neutral sources if possible)

Bad:

• Traditional media (reputable newspaper and magazine articles, sourced online or off). Exception: articles bylined by known experts in a field, for example: an article on data visualization practices by Edward Tufte would be OK to quote/ source in an infographic about data visualization.
• Blogs (corporate, media, non-profit organizations, government)
• Any data that is outdated, especially in fast-paced industries (mobile, social media, Internet)

Ugly:

• Feel free to try to find another source for the data (see our “good” list), but do not use any of these as sources.
• Personal blog posts - by authors with unknown or questionable credentials
• Wikipedia
• “Google”
• Other infographics
• Any UGC website (Yahoo Answers, Google Answers, etc)
• SEO content, incl. that from eHow, About.com, Demand Media
• The number one or two Google search result, cited with no further research into the matter.

Last, but not least...
Please respect the intelligence of your client. Remember that they probably have expertise in the subject matter at hand. They will know if your research was perfunctory and will call it out to us.

On the other hand, well conducted, thorough research will leave a good impression. It could result in positive feedback; repeat work; and a strong, long-standing relationship with Visually!

Also read Visually's Plagiarism Policy.