Gender Bias in Statistical Anxiety Research Report

A basic understanding of statistics is a requirement for achieving academic and professional success in the behavioural sciences (Birenbaum & Eylath, 1994; Quinn, 2006). Statistics-related courses are therefore a compulsory component of most of such behavioural-science degree programmes as psychology and social work (Schram, 1996; Tremblay, Gardner, & Heipel, 2000). Statistical anxiety is, however, a pervasive problem for many students who are attracted to the counselling aspects of such degrees (Baloglu, 2004; Rodarte-Luna & Sherry, 2008; Royse & Rompf, 1992 ). This paper aims to investigate the differences between males and females in regard to statistical anxiety and their attitudes towards studying statistics.

Statistical anxiety is a situation-specific response to any encounter with statistics that is characterised by feelings of fear, nervousness, and apprehension (Baloglu, 2004; Onwuegbuzie, 1995; Onwuegbuzie, DaRos, & Ryan, 1997; White, Hayes, & Livesey, 2010). Fitzgerald, Jurs, and Hudson (1996) found that statistical anxiety is the best indicator of achievement in statistics-related courses, and Onwuegbuzie and Wilson (2003) found that up to 80% of graduate students suffer an uncomfortable level of anxiety in regard to statistics. It is reasonable to assume that this percentage also applies to undergraduate students, as such factors as prior experience, skill, competence, interest, and motivation influence the degree to which people suffer statistical anxiety (Hong & Karstensson, 2002; Manning, Zachar, Ray & LoBello, 2006; Sizemore & Lowandowski, 2009), and graduate students have more academic experience.

Much interplay takes place among anxiety, attitude, and ability, which Lalonde and Gardner (1993) termed the "three classes of variables" involved in statistical learning (p. 108). Such demographic variables as sex, age, and previous mathematical experience also play a significant role in the learning of quantitative methods (Bell, 2003; Benson, 1989; Eriksson & Lindholm, 2007; Rodarte-Luna & Sherry, 2008).

Both Baloglu (2003) and Gal, Ginsburg, and Schau (1997), however, have found that people's attitudes toward studying statistics are closely related to statistical anxiety, as these can either hinder or enhance anxiety levels, course performance, or both regardless of true ability, and Onwuegbuzie (2000) argued that the degree to which people suffer statistical anxiety may be a determinant of their attitude toward statistics. Similarly, Tomazic and Katz (1988) found a bidirectional influence between attitude and anxiety in which either can be the forerunner. They also found that the time that has elapsed since people's previous mathematical courses and the perception of prior success in mathematics are two key factors in statistical anxiety and in forming attitudes towards statistics. These factors are important because such aspects of male and female differences are related to mathematical learning, aptitude, and liking as differences in maturational timing and brain structure (Halpern et al., 2007), social climates (Mulhern & Wylie, 2004), and gender-specific stereotypes and preferences (Frenzel et al., 2007; Spencer, Steele, & Quinn, 1999).

Baloglu (2004), DeVaney (2010), and Finney and Schraw (2003) have found a negative relationship between people's levels of statistical anxiety and their attitudes toward statistics, with higher levels of statistical anxiety tending to correspond with less positive attitudes. Papanastasiou and Zembylas (2008), however, found higher levels of statistical anxiety among those who consider research methods to be an important aspect of their profession. These findings are interesting because statistics' usefulness in terms of value or worth is a common factor in attitudinal measures of statistical learning (Roberts & Bilderback, 1980; Tremblay, Gardner, & Heipel, 2000). It is also the first of six factors that constitute the Statistics Anxiety Rating Scale, which asks students about their perceptions of statistics courses' value and worth (Bell, 2003; Cruise, Cash & Bolton, 1985).

Statistical anxiety is also closely related to the concept of mathematical anxiety (Baloglu, 2003, 2004; Benson, 1989; Cruise, Cash, & Bolton, 1985). Although it is important to acknowledge that they do differ considerably, it is also important to highlight Baloglu's (2004) finding that many students overestimate the mathematical nature of statistics courses, so the residual or carryover effects of mathematical anxiety may be the source of some students' statistical anxiety (Hong & Karstensson, 2002; Rodarte-Luna & Sherry, 2007; Royse & Rompf, 1992), and Zeidner (1991) found that previous negative experiences with mathematics tend to be antecedent correlates of both statistical anxiety and attitudes toward studying statistics. Notable examples of this include past mathematical performance and achievement and negative or poor feedback from teachers (Cooper & Robinson, 1991; Hackett, 1985; Pajares & Miller, 1994). Consistent with this finding is repeated findings that female students tend to experience higher levels of statistical anxiety than do their male counterparts (Bradley & Wygant, 1998; Onwuegbuzie, 1993, 1995; Stroup & Jordon, 1982; Zeidner, 1991), largely due to the long-standing gender-biased stereotype that "girls and mathematics are a bad-fit" (Frenzel, Pekrun, & Goetz, 2007, p. 497).

An early report by Tobias (1978) offered sound explanations for why females tend to report higher levels of statistical anxiety and why males often have more positive attitudes toward statistics (Bradley & Wygant, 1998; Cashin & Elmore, 2005; Onwuegbuzie, 1993, 1995; Stroup & Jordon, 1982; Zeidner, 1991), finding that only 8% of the female students in her US sample had taken all four years of high school mathematics, while 57% of the male students had. It is possible that radical changes have occurred in mathematics education during the subsequent 32 years that have lessened male dominance of both secondary-school and higher-education mathematics. The finding is noteworthy, however, in the context of the structure of schooling in the seventies, when minimal subject choice was available and males still managed to complete significantly more courses than females.

A recent longitudinal study, furthermore, suggested that Tobias's findings are still valid, as it found statistically significant sex differences in both planned and actual enrolment among Australian adolescents, with females planning and actually enrolling in fewer mathematics courses than males, these differences being consistent from year 9 through year 11 (Watt, 2006). Females tended to score lower on mathematical self-perceptions than males, to see mathematics as