How do you determine that your information is accurate and valid?

Do you take the time to evaluate where and how the information got to you?

What makes you believe that it is so? Is it the source of the information? The author of the article? The website it came from? The article validates your beliefs, so it must be true? The fact that a best friend from high school shared the article on Facebook, and they got better grades than you in school, so they must know what they’re sharing?

There are various checklists that one can use to begin to evaluate information and websites, such as CARS, which stands for Credibility, Accuracy, Reasonableness, and Support. Purdue University has a checklist page for evaluating print versus internet resources. The University of Arizona Library has a checklist page for evaluating websites. There may be a generational gap in the knowledge of and use of website checklists as those who have grown up with the internet are more likely to have more experience with the internet as well as instructional time reviewing checklists such as these, while older adults may have less experience and may not be aware that such checklists exist. Checklists can be a good start, but they are often based on what you can see or evaluate quickly. As such, they can be limited in evaluating how accurate and verifiable website information can be. The checklist might only help one discriminate a valid website from a completely bogus website. What else can we do to verify information?

An example for verification evaluation can come from currency. (see this website for how to validate a US $100 bill). The website shows six defined security features included on the $100 bill that verify that it is genuine American currency and not a counterfeit. These security features include special printing techniques, color changing ink, security threads, and watermarks.

Validating information isn’t quite as easy as validating currency. A webpage doesn’t have a special printing feature that will validate its claims as authentic and accurate. Even if it did, it might not be long before those signals were mimicked, copied, and/or spoofed to add credibility to websites that may be intentionally misleading. For example, some websites have “badges” that show their credibility; and others have taken to using spoofs of similar icons.

What sorts of indicators and good signals can be used?

One thought is the difference between the ‘free web’ – articles and content that are freely available – and the ‘paid web’ – news, articles, and scholarly work, for example, that require pay access. Many of the paid web access articles may be part of databases that available to the general public or students at public or university libraries for free. Subscription or paid access can be a signal that the information has been through a vetting process prior to being published. It is also likely that the publishers and authors of the information are not anonymous, which can be an issue for other sources. This could be similar to a security ribbon in currency.

What signals does the information itself provide? Does the information hold together on its own? Is it comprehensive and logical? Do other websites refer to this information, or corroborate it? Is it inflammatory and designed to illicit an emotional response? What sort of writing or verbal style is used to communicate the information? What is the intended audience for the information? Are there citations to other sources? These sorts of questions can reveal the ‘watermark’ in the information, as much as holding a $100 bill to the light reveals Ben Franklin’s face.

Being an engineer, one of the other ways I consider the meaning of good signals is in terms of meeting the correct thresholds for use. Even digital signals, interpreted as values of either zero or one, do not always translate exactly to the digital voltages of zero or positive 5 volts. Even binary signals are not truly binary when examined closely with a meter. There can be thresholds between 0.5 and 1.0 volt depending on the circuit, meaning that in some applications, 4.0 volts is ‘good enough’ to declare a value of one, digitally. There may always be margins of error in information evaluation, but does it cross the threshold when measured?

For discussion and comments:

• Do you agree or disagree with the analogies proposed here? Why?
• What other good signals can be used that have not been mentioned?