The short answer is that an Infographic should be able to tell a story that is supported by the data/information that underlies it. Have your seen any of the Edward Tufte books, such as Visual Explanations or The Visual Display of Quantitative Information (see links below)? He uses lots of examples to reinforce the message, "Above all else show the data". Like many things, it is often easier to spot a good infographic from a bad one than it is to create a good infographic and not a bad one.
I agree with above. However if you want to get back to fundamentals this question lies firmly in the field of the ergonomics of display. Basically ergonomics has another field of control, what you do to make something happen, whereas display is about conveying what does happen. Gross oversimplification of course. There are countless references in display ergonomics but googling that might throw up some interesting leads. The other implied strand to your question is how might we measure success. This in turn raises several issues. Success is surely measured by a mental state change in the observer. Are you interested in how well they remember the information later, or how accurately they perceive it immediately and so on. Huge field of cognitive measurement there. An example of this...research has shown that we pay more attention in video to gesture than words, so for example someone gesturing 'large' is more likely to convey the largeness of soemthing than using the word either written or spoken. There are so many parallels her to investigate. If you had a more specific question I might be able to help more directly
I think to evaluate of visual products depends on the type of visual products, but as overall you can use bradzel's Methods to consider evaluation, by the first step of his method (reading the Visual Items) you can identity the most important criteria. and on the base of these factors you can make a questionnaire. certainly you must standardize your questionnaire base on some statistic factors, like Cronbach's alpha. till now, i think this method is so practical and lots of new papers have written on this vision. but most of them are not familiar with Bradzel's Method.
other method is base on the Semiotic Critical. as you know some methods like this is base on the Qualitative research methods. this is safe but it needs more cognition to know context and so on, and also more thinking on the part of discussion. there are two files of bradzel's paper are attached to download.
I hope my explanation is useful and it will be helpful.
There is another aspects that are very important to evaluate infographics. The AUDIENCE and the INFORMATION DENSITY. Jenn and Ken Visocky O'Grady explain this in their 2008 book The Information Design Handbook with a very nice scheme.
It´s a matrix with two dimensions, one that reads about low-high desity of information and another that goes wide-specific audiences. In the "low-wide" area you can include the airport way-finding images, while "high-wide" you can set the yellow pages, a lot of information but for a lot of persons.
In the other aspects, "high-specific" you can think about a medical journal or in the "low-specific" they set the example of professional software icons, like the ones you can see in autocad or maya.
I don´t know if I made myself clear but, maybe you should check this book. Good luck!
The creation of an infographic that is as intuitively simple to understand is what we strive to do in our studio. Context is critical; who is your audience? Recommendations above have been very interesting and thoughtful; Tufte is essential. Other good references are The Visual Miscellaneum by David McCandless, and Design for Information by Isabel Meirelles. Good luck!