This study examined sex differences in, and the influence of IQ test experience on, the self and partner estimation of Gardner's ten multiple intelligences. Over 600 students from New Zealand completed a brief questionnaire based on those used in previous research (Furnham, 2001). Three of the ten self-estimates yielded significant sex differences. Males believed they were more intelligent than females on mathematical (logical), spatial and existential intelligence. Those who had previously completed an IQ test gave higher self-estimates on eight of the ten estimates. Self-estimates were compared to university samples from America and Britain, and results tended to show New Zealand students gave lower self-estimates particularly on mathematical, body-kinetic, existential, spiritual and naturalistic intelligence. Factor analysis showed the ten multiple intelligences fell into three interpretable factors which were predicted by both gender and test experience.
A study, published by Beloff (1992) ten years ago on sex differences in self-estimated intelligence in Scottish students, has provoked a good deal of research. Similar studies have been carried out in America, Germany, Hong Kong, Iran, Japan, Singapore, South Africa, and Uganda (Bennett, 1996, 1997, 2000; Byrd & Stacey, 1993; Furnham, 1999, 2000; Furnham & Baguma, 1999; Furnham, Clark & Bailey, 1999; Furnham & Fong, 2000; Furnham, Fong & Martin, 1999; Furnham, Hosoe & Tang, 2001; Furnham & Mhkize, 2002; Furnham, Rakow & Mak, 2002; Furnham & Rawles, 1995; Furnham, Shahidi & Baluch, 2002; Rammstedt & Rammsayer, 2000).
The study of self-estimated intelligence and implicit theories of intelligence is of considerable research importance but also practical significance. Dweck and Bempechat (1983) note that personal beliefs about intelligence are unrelated to actual ability but have a behavioural and cognitive impact in academic situations. Mueller and Dweck (1998) showed that, paradoxically, being praised for being intelligent can actually undermine children's motivation and performance because they believe academic success and failure are a function of "native intelligence" rather than effort and hard work. Beyer (1990, 1998, 1999) has demonstrated sex differences in intelligence in terms of expectations, self-evaluations and performance on ability tests. She notes that self-evaluations affect expectancies of success and failure as well as actual performance on these tests. Her particular interest is in gender differences in the accuracy of self-evaluation but notes its practical implication, particularly in suggesting ways of eliminating women's underestimation of their own ability. In this sense studies on self-estimated intelligence are linked to the literature on implicit theories, self-perception and expectancy theory as well as attitude-behaviour research on the behavioural consequences of holding particular beliefs.
The results of the studies on self-estimated intelligence are remarkably consistent. When asked to estimate overall IQ (g), nearly every study has demonstrated a consistent sex difference of usually between 3 to 6 points, with males giving higher estimates than females. This is true of both student and non-student populations (see Table 1). Only two studies had adult non-student populations (Furnham & Gasson, 1998; Furnham, Reeves & Budhani, 2001). However there has been one exception, which was the study by Byrd and Stacey (1993) in New Zealand, which found no sex differences in self-estimates but very large differences in estimates of parents' and siblings' intelligences, with females rating their parents (mother and father) and sister (but not brother) as more intelligent than did males. Participants in that study were second year students at the University of Canterbury in Christchurch.
The current study replicates and extends this literature in a number of ways. …