Over the past three decades work has been carried out to demonstrate the idea that both teachers and students have personality characteristics that can affect the way they teach or learn and interact (Abrami et al, 1982; Murray et al, 1990; Wolk & Nikolai, 1997; Radmacher & Martin, 2001; Worthington, 2002; Zhang, 2004; Rushton et al, 2007). This work has been used to suggest strategies for optimising teaching effectiveness and for explaining the range of evaluations that teachers receive from their diverse classes.
A number of researchers have taken this concept one step further and used the multidimensional nature of the student evaluation of teaching surveys to show that teachers can have a characteristic teaching style that can be represented as a profile when the results of student evaluations of teaching are compiled. Marsh & Bailey (1993) illustrated the concept of multidimensionality by explaining that the same instructor might be rated well by students in terms of organisation, but poorly for enthusiasm. They went on to argue that it might be possible to optimise teaching by determining a profile that led to optimum results. For example, student learning might be optimised when both enthusiasm and organisation were high whereas teaching that scored well in only one out of the two might not be sufficient. Gibney & Wiersma (1986) made a similar argument about the ideal teacher based on measured student-teacher competencies, which could also be plotted as a profile. The teacher profile has also been identified in work by Stiggins (1988) and more recently by Pan et al (2009), who relied on the written feedback from students on the course evaluation responses.
If teachers have a characteristic profile that persists as they present different courses to different students, it may be important to know if students also have a characteristic profile that they apply when evaluating teaching since any bias in the responses will add to the variance in the returns. Weber & Frary (1982) raised the possibility of this being true when they argued that it was "not difficult to imagine the existence of groups within a class who react differently with respect to a questionnaire item". However, without an identifiable survey return it has not been possible to pursue this in any reliable way as students are protected by anonymity in most evaluations of teaching.
If students do have a characteristic profile this might be useful in explaining some of the scatter that is observed in evaluations of teaching. It is worth investigating because, after all, any characteristic profile assigned to a teacher is only there because it has been assigned by groups of students and it would be wrong to put much reliance on a teacher profile, while at the same time ignoring the students who are responsible for generating it. These profiles, if they exist, may be dependent on a number of factors such as student background and personality.
While the effect of disparate personalities in a class may have no overall effect on student survey responses, particularly in large classes, if the class were filled with students who typically respond to student satisfaction survey questions in a fixed pattern, this could have an impact on teachers who teach those classes. Certainly personality traits such as learning goal orientation have a significant effect on student responses (Moghaddam et al, 2009). Some of the so-called "Big 5" personality traits defined in modern psychology (John et al, 2008) imply direct consequences for survey responses. For example, one of the Big 5 personality traits is Agreeableness. One example provided by John et al (2008, pp. 120) suggests that one aspect of this personality trait is "emphasizing the good qualities of other people", which is likely to have a direct impact on student surveys. It has been shown that personality has an effect on the response rate of students to surveys (Porter & Whitcomb, 2005). …