group will receive information on the same event through other sources of
information (texts, radio, etc.).
The reactions of the participants in an experiment can be measured in various
ways. According to Annie Lang ‘[t]he beating of a heart can tell us something
about attention, arousal, effort, and emotion… details about a person’s
physiological and cognitive state… [and] can be recorded with very little
interference with the subject’ (Lang, 1994: 99). Emotion is, in other words,
actually best assessed through a simple and often-overlooked method: measuring
the heartbeat. Another, perhaps simpler, way to track emotions and cognitive
processes are so-called ‘self-report measures’ (Biocca, David & West, 1994: 20).
Starting with the most basic ‘pen-and-paper version’, where the participants write
down their evaluations of a particular message, self-reporting has developed in line
with advancements in computer technology to be a form of ‘continuous response
measurement’ (CRM). CRM involves showing pictures to an audience and letting
this audience, through the usage of hand-held devices like a push-button box or
joystick, continuously evaluate a visual. The viewers’ reactions are simultaneously
recorded in a computer and displayed to the researcher (Ibid.: 19).
Although this part of the project still needs to be developed, it provides a
preliminary sample of the methodological direction and basic techniques that will
structure the empirical part of this research project.
7
Being the most common design in communication studies, the ‘between-subjects design’ is not without
its critics. For one, individual characteristics and differences among the participants can lead to biased or
even erroneous results. Nonetheless, we would still argue that this strategy is preferable to the ‘within-
subject design’, where the same test groups views different versions of the same message. In that design
there is a great risk that the ‘juxtaposition of treatments will sensitise subjects to the intent of an
experiment and interfere with responses’ (Ibid.: 175). For instance, when the experiment concerns issues
like measurement of emotions, we believe that showing different versions of the same message to one test
group would be far too obvious and influence the respondents’ reactions.
19
Convention