Basic Terminology (Sections 1 and 3.1)
(Research)
Hypothesis
Ex: "Visualization A and B are equally good, i.e.
there
will
be
no
difference
between users who use
visualization
A and those who use B"
"Users who use
visualization
A
will perform easy tasks
faster than those who use visualization B"
Experiment
Treatment
variables (aka independent
variables, factors)
controlled
(systematically varied) by experimentors
Ex: kind of visualization interface, difficulty of task
Dependent
variables Ex: task performance time, user satisfaction,
error
rate
Nuisance
variables (aka confounding factors)
Ex: age, fatigue, motivation,
gender, time of the day
May be hidden: Ex: Visualization
interface A has design flaws
(that
are independent of the visualization principle)
Interface
A is always tried in the mornings,
Interface
B in the evenings
* Balancing
these variables across treatments or
* keeping
them constant across treatments or
* (for
subject factors): random assignment to treatments or
*
factoring them out
during the statistical analysis
Subject sampling
Assignment to treatment (block randomization!)