The Target Selection Task
Target selection is a heavily studied research area in HCI. A simple target selection task involves a user clicking on an icon, where the icon acts as the target. Researchers in this area focuses on developing interaction techniques to help users select targets more efficiently.
An example you might have seen is shown in the picture below. The image on the left: Suppose you were asked to use the interface below to first click on the start button then select the the purple icon (among the others in the set). The icons all look pretty similar, except for the colour and the drawing on it. The image on the right: Suppose you were asked to do the same task, but as your mouse moves closer and closer to an icon, that icon gets bigger (even if you happen to move the mouse to the wrong icon). This interaction technique is known as expanding targets. You might have seen/used it in the Mac OS some years back. (This is one of my friend's old research work, read Acquisition of Expanding Targets (McGuffin and Balakrishnan, 2002) for details.) Which interaction technique do you think is easier for selecting icons? To explain the intuitions behind your answer, we turn to Fitts' law.
The Initial Experiment
The most widely used model in HCI is called Fitts' law. This model was developed by a psychologist Paul Fitts back in 1954. (Yes, psychologists do serious math too.) Fitts' law is designed to model people's rapid aimed movements when they know what they need to select and they just move the mouse to select it.
An example is when there is a target icon on the screen that you want to select using the mouse cursor. This model predicts the time it takes for you to move your mouse from one location to the target icon. The images below are adapted from the images in the original papers by Fitts. The images illustrate an abstract task that the researcher gets users to do. In the serial task on the left, the user holds a stylus and clicks on a starting point (anywhere in the black bar on the right), the user moves to the left side and clicks the target area (anywhere in the black bar on the left). In the discrete task on the left, you will see a similar setup, but the user starts in the middle, and depending on which stimulus light shines, the user moves and clicks on the respective target bar.
The Most Important Equation in HCI
Fitts' law was derived from information theory and is considered the most important equation in HCI due to the influence that it has on design. There are two key terms to notice in Fitts' law: the index of difficulty (ID) and the movement time (MT). ID is a measure of the difficulty of a target selection task. The picture and formula below illustrate the model for ID.
Suppose we have the abstract layout where two areas represent the location where the user's pointing device starts and the location where the user's pointing device ends. (The pointing device can be a mouse, pen, etc.) Without loss of generality, let the starting point be the middle of the starting area and the ending point be the middle of the target area. Then A represents the amplitude, or the distance, between the starting point and the ending point. Moreover, suppose we could vary the width, W, of the target area. The ID model, as indicated by the equation, illustrates that it gets exponentially harder as W gets smaller. Conversely, a target where W is big is very easy to select.
Given the model for ID, Fitts hypothesized that the relationship between MT and ID is linear. This means, if a target is hard to select, then it will take an average user more time to select it, and if a target is easy to select, then it will take an average user less time to select it. Specifically, MT = a + b x ID, where units are in milliseconds and the constants a and b are empirically derived. This equation that defines movement time is what we call Fitts' law.
Note: To have an equation where values are empirically derived means that one needs to collect data to see how a range of users behave in a specific environment and application in order to completely determine the value on the left hand side of the equation.
An Intuitive Explanation
To sum up, the basic interpretation that you should understand about Fitts' law is that it is faster to hit larger targets closer to you than smaller targets that are further from you. Here's an intuitive video that explains how Fitts' law works with specific interface examples.
Isn't that cool about the position under your mouse? It's pretty obvious, but we don't really think about it because it's not there. Now, what about the corners of the screen? Did you come up with that answer? Another answer you might have thought of was probably the top and bottom of the screen -- that's also very close, since you can fling the mouse in that direction without having to slow down and concentrate on whether you hit a small target.
Application of Fitts' Law
There are three ways to use Fitts' law in HCI. First, you can use Fitts' law to analyze and compare design alternatives. Because Fitts' law predicts movement time, you can use it to compare which design alternative affords more efficient movement for the user.
Second, you can use Fitts' to determine the index of difficulty of a design, and re-design as needed. This can also be done to compare and evaluate multiple designs to explain which design is easier for users to work with.
Third, you can determine if a new device or technique conforms to Fitts' law. New research explores how the Fitts model can be adapted to other application contexts, such as 3D target selection, mobile interfaces, tilt-based interaction, and gesture-based interaction, just to name a few. Research on Fitts' law is still active today.