Since public speaking anxiety has been commonly experienced by many language learners, many studies have tried to identify the reasons which cause oral anxiety during a public speech. Young (1991) has identified six sources of anxiety, including personal and interpersonal anxieties, learner beliefs about language learning, instructor beliefs about language learning, instructor-learner interaction, classroom procedures, and language testing. Based on these sources, public speaking anxiety belongs to one of the personal and interpersonal anxieties. Furthermore, perceived competence (Chesebro et al., 1992; MacIntyre & MacDonald), perceived audience congeniality (MacIntyre & MacDonald, 1998), audience expectation (Ayres 1986), audience interest (MacIntyre, MacDonald, & Thivierge, 1997), preparation (Daly, Vangelisti, & Weber, 1995), the affective contagion from the former speaker (Behnke
& Sawyer, 1994 ), and speakers’ beliefs (Daly, Vangelisti, Neel, & Cavanaugh, 1989) were all found to be associated with public speaking anxiety. Bippus and Daly (1999) also found nine factors which people believed would cause public speaking anxiety:
Humiliation, preparation, physical appearance, rigid rules, personal traits, audience interest, unfamiliar role, mistakes, and negative results.
According to Scovel (1978), there are usually three ways to measure anxiety:
by behavioral tests, by the subjects’ self-report, and by physiological tests. Through behavioral tests, others rather than the subjects themselves are required to observe the subjects’ behavior when they are experiencing anxiety. The second way is to have the subjects self-report how they feel about their anxiety. Measurements of heart rate, blood pressure, and palmar sweating are examples of the third way which measures how a subject physically feels.
As far as the three ways of measurements are concerned, there are some specific concerns to which attention need to be paid in order to conduct a valid research on public speaking. As for the first way, it is necessary that the observers must be capable of measuring the phenomenon required to be observed. For example, the speaker’s performance and the audience’s response are usually two of the major focuses of observation during a speech performance. Then how to train the observers to precisely measure the speaker’s and the audience’s behavior may be a problem.
Behnke and Sawyer’s study (1994), which required the audience to rate each
speaker’s state anxiety, is an example like this. In the study, the observation task was done at the undergraduate level. It will be more difficult if it is done with junior high school students since they only have limited knowledge about what actually happens in the context of a public speech. As for the second way of measurement, there are two points that need to be dealt with caution. First, by self-reports, the subjects are often required to make a connection between their speech performance and their speech anxiety. This task will be more easily done when the subjects are provided with a well-established scale for speaking performance and anxiety. If the research is conducted with a questionnaire in an open format, it will then solely depend on the subject’s knowledge to think of the relationship between the two variables. Then if the participant does not posses adequate background information for the observation, the results of the study may be limited to a certain degree. An example like this is Bippus
and Daly’s research (1999). In their preliminary study, they required their participants to think of as many reasons that people generally believe will cause public speaking anxiety as possible. The study was done without too many difficulties since the participants were all native speakers who had attended communication courses at college. However, if the research is conducted in a foreign language environment and with a group of younger participants, then the limited background knowledge will cause problems. Second, some studies involving self-reports in studies on public speaking will require their subjects to immerse in their imaginations to construct the settings of the public speaking context. Then the more concrete the contexts needed to arouse anxiety are, the more persuasive the research may be. MacIntyre et al.’s
research (1997), for example, had their subjects rate their anxiety, willingness to speak, and quality of presentations through their imaginations of eight contexts.
Although merely anticipation of a public speech can even cause anxiety (as cited in MacIntyre et al., 1997), their research process through imaginations was still
unobservable and might be criticized that it was too subjective and unreliable. As for the third way of measurement, there are two concerns worth noticing. First, Scovel (1978) suggested that the physiological measurements which claim to measure anxiety must be highly related to anxiety. Second, it should be noticed that there are still individual differences in the physiological responses among various participants (Scovel, 1978). Take McCullough, Russell, Behnke, Sawyer, and Witt’s research (2006) for example. They examined the relationship among body sensations, state of mind, and anticipatory public speaking state anxiety. For the study to be valid, it must be conducted in the assumption in which body sensations and state of mind are akin to public speaking anxiety. Moreover, even if the relationships between the three variables are proven to be highly correlated, variations among individuals still exist.
That is, it may be true that some high anxious speakers do experience high levels of
body sensations before a speech while some others do not.
Unlike Behnke and Sawyer’s study (1994), which asked their audience to rate the speakers’ anxiety, this study was conducted through two well-established
self-reporting scales (Behnke & Sawyer, 1998). One was a narrow-band version of the STAI (A-State) at the three stages (i.e. anticipation, confrontation, and release), and the other was a narrow-band anticipatory public speaking trait anxiety version.
Contrary to physiological tests, the two scales directly measured the relationship between anxiety and speech performance, instead of the relationship between
behavior and anxiety. With the STAI, it is not necessary for participants to think of the factors associated with anxiety themselves and to heavily rely on their background knowledge. In addition, the study was held in a formal English speech contest with junior high school students rather in imagined contexts. It provided a more concrete context for the arousal of anxiety than the imagined settings in the study of MacIntyre et al. (1997).