The importance of hypotheses

When you want to conduct a study, you first need to know what question you want to find an answer to. So you need to put some idea (based, for example, on observations or previous knowledge) to the test. Such an idea is called a hypothesis. The asked question should lead to testable predictions - the more specific they are, the fewer possibilities to explain the results (which is, of course, better for researchers). Eye-tracking test studies bring a huge amount of data, so it’s necessary to understand the problem to answer the research question properly. 

Example 1

You have two versions of your website design, and you want to know which draws the most attention from viewers. You can show your participants both versions (in random order for each one) or make two studies, so two groups of participants and show each group one of the designs and compare the results.

Note: when participants freely browse a website higher number of fixations means that this area is more interesting, but if the participants are given a task to find something on the page the higher number of fixations could mean uncertainty and difficulties in recognizing the elements.

Example 2

You want to find out the difference between novice and experienced drivers. Show these two groups the same set of images/video and analyze their eye movement measures and patterns to get information on where they were looking most intensively and in what order and compare it.

Example 3

There is some text to analyze if it’s understandable. Give it to your group of participants and analyze their eye movements. A hot point in some areas or frequent returns to a fragment of the text may indicate that there were some difficulties in understanding the given text at that point.

There’s also a rule of thumb to make a pilot study - small-scale versions of the main study to make sure that the study is feasible. It consumes fewer resources (participants, time spent analyzing, etc.) than a full study, but it gives a good notion about the hypothesis. Plus, it can undercover any pitfalls and mistakes in the early stadium of the project. It can also convince investors why this project is worth funding. 

It’s also important to remember that offset (number of tests that are failed) is about 15%. RealEye is aware of this fact and always orders about 15-20% of participants more, so the data it provides to you is without this offset. However, you have to remember that when you use your own participants you need to take it into account.