Predictive Models

There are 6 behaviour models that help HCI designers to predict the way an interface will behave, and if it is effective enough to be used on a computer or device. These 6 behaviour models are split into 2 catorgories, predictive and descriptive.

The predictive models are:

•  Keystroad-level model (KLM)
• Throughput (TP)
• Fitt's law

The discriptive models are:

• Key-action model 
• buxton's three state model 
• Guiard's model of bimanual skill 

I am going to look at 1 predictive model and one descriptive model. 

keystroke level model  (predictive model)


The keystroke level model was described by Card, Moran, and Newell in the early 1980s. The model focuses on how long it takes users to actually use the HCI via hardware. The keystroke level model features 11 steps that is used by individual people and organisations, they use this to estimate how long it takes to perform simple tasks involving the input of a human via hardware. Normally companies who cannot afford specialists use this method. 


The keystroke level model defines and measures how long it takes to press and release a key on the keyboard (measured in words per minute, and categorised into fast, novice and slow typists), how long it takes to point the mouse on the screen, how long it takes to press or release a mouse click, how long it takes to switch hardware devices, ie keyboard and mouse, how long it takes for the human brain to prepare to peform an action within an HCI, how long it takes to type a string of characters,how long the user has to wait for the system to perform the action in the HCI. 

The key-action model (KAM) (discriptive)

Computer keyboards today contain a vast array of buttons, the buttons are either symbol keys, executive keys, or modifier keys. Symbol keys deliver graphic symbols — typically, letters, numbers, or punctuation symbols — to an application such as a text editor. Executive keys invoke actions in the application or at the system-level or meta-level. Examples include ENTER, F1, or ESC. Modifier keys do not generate symbols or invoke actions, but, rather, set up a condition necessary to modify the effect of a subsequently pressed key. Examples include SHIFT or ALT. Basicly, this model looks at how users interact with HCI's using keyboards and what shortcut keys HCI's use.