We are undertaking a research project to determine whether existing guidelines for formatting text for museum exhibitions are rooted in unconscious neural predispositions to select certain formats over others. A reader would experience this as increased or decreased willingness to read text when presented in different formats. Our hypothesis is that the more the text format conforms to certain consensus guidelines, the more willing anyone will be to read it, regardless of age, gender, education level, or learning (dis)ability.
We will focus on five key aspects of text format to test unconscious reader reactions: font size, letter form thickness, foreground/background contrast, line length, and letter spacing (kerning).
We assume three basic components will have to be created: 1) a web portal that can be accessed by either PC or Mac devices to run the test, e.g. laptops, desktop computers, iPads or similar devices (the nature of the test makes running it on a smart phone not feasible); 2) the test itself, which would include user interface and data collection; and 3) a means of immediately uploading the results of each test to a secure cloud site, which we can access as needed.
The website would send the users to the first page of the test sequence.
The test has three basic parts: an intro/training section; the test itself; and personal info collection section at the end.
We have the test itself already broken down into specific pages, with specific instructions about what exact text and images would appear on each page, and with specifications about how users would move forward and backward through the pages. We also have specified the data collection categories for the user's profile, separate from the data collected in the test itself. It is important that the data be collected confidentially, but that each user's test results be "married" to the personal data entered by the user and not altered when uploaded.
Also, the users will have the option to have all text read out loud to them orally, so that option needs to be programmed into the intro/training section of the test, and the personal info collection section afterwards. The reading test itself occurs too quickly to be read out loud; it is designed to be a rapid visual reaction test only.
Essentially, once the user is walked through a series of instructions--and also gives explicit consent to have the test results used for research purposes--each user will be presented with a split screen in which the same line of text appears, but in different iterations of the five aspects of text format being tested. The screens will flash for 400 ms, and then the user will be asked to select which screen s/he feels is "easier" to read.
We will be providing for each of the five key text formats five iterations: one which is most conforming to existing consensus guidelines, and four which deviate (two in one direction and two in the opposite direction). For example, standard font size should be 24 point, so there needs to be an iteration in 24 pt., but also 30 and 36, and 18 and 12. The job of the designer is to create a program that randomizes how all five variations of the text format interact with each other for any one screen. The only proviso is that for each flashing of the split screen, one of the screens must ALWAYS be in sum more conforming to the standard than the other. We also ask that for EACH test, for one flashing of the split screens, one of the screens ALWAYS be fully conforming, and for one flashing of the split screens, one of the screens ALWAYS be fully violating. That is, each time, we want to test the extremes of the range, and then randomized variations within those extremes. We are thinking at present that users will be go through 10-12 selection sequences of split screens. With the intro/training, and with the end data collection, the entire time for the user should be roughly 15 minutes.