Exploring the Prototypical Definitions of Intelligent Engineers Held by Irish and Swedish Higher Education Engineering Students

Males are generally overrepresented in higher education engineering. However, the magnitude of this variance differs between countries and engineering fields. Evidence associated with the field-specific ability beliefs hypothesis suggests that perceptions of intelligence held by actors within engineering affects the engagement of underrepresented groups. This study examined perceptions of an intelligent engineer held by undergraduate and postgraduate engineering students in Ireland and Sweden, countries selected based on their levels of female representation in engineering education. It was hypothesised that there would be a significant difference in perceptions between countries. A survey methodology was employed in which a random sample of Irish and Swedish university students completed two surveys. The first asked respondents to list characteristics of an intelligent engineer, and the second asked for ratings of importance for each unique characteristic. The results indicate that an intelligent engineer was perceived to be described by seven factors; practical problem solving, conscientiousness, drive, discipline knowledge, reasoning, negative attributes, and inquisitiveness when the data was analysed collectively, but only the five factors of practical problem solving, conscientiousness, drive, discipline knowledge and negative attributes were theoretically interpretable when the data from each country was analysed independently. A gender × country interaction effect was observed for each of these five factors. The results suggest that the factors which denote intelligence in engineering between Irish and Swedish males and females are similar, but differences exist in terms of how important these factors are in terms group level definitions. Future work should consider the self-concepts held by underrepresented groups with respect to engineering relative to the factors observed in this study.


Independent EFA solutions for the Irish and Swedish data
In comparing the factor structures between the EFA analysis where the responses from both the Irish and Swedish participants were included and the pursuant EFA analyses where the data from the Irish and Swedish participants were included separately, the EFA solution where the data from both countries were included will henceforth be referred to as the 'consolidated factor structure'. To examine the EFA structures using both the Irish and Swedish data separately, identical processes of determining the factorability of the data, the number of factors to extract and conducting an EFA to those used in determining the consolidated factor structure were implemented. Additionally, as with the previous EFA, factors were theoretically interpreted based only on characteristics which loaded on them that had pattern coefficients greater than .4 and less than -.4.
For the Irish sample, of the 3916 correlations in the correlation matrix, 314 were greater than or equal to .3. An examination of the anti-image correlation matrix revealed that 73 of the 89 anti-image correlations were above .5, and the off-diagonal elements were mostly small (M = -.0104, SD = .1388). The KMO measure of sampling adequacy was .617, above the recommended value of .6 (Kaiser, 1974), and Bartlett's test of sphericity was significant (χ 2 (3916) = 7356.199, p < .001). Therefore, there was a reasonable level of factorability within the data. A parallel analysis (Horn, 1965) suggested a seven factor solution for the Irish sample Figure S1. The EFA solution is presented in Table S3.

FIGURE S1
Factor eigenvalues and parallel analysis for the EFA with data from the Irish sample only. The first factor which was extracted is loaded most heavily on by the characteristic of being able to understand complex information. This is then followed by five qualitatively similar personality characteristics; being solution orientation, methodical, practical orientated, responsible and realistic. While not exactly the same, it shares many of the higher loading characteristics as Factor 5 'Reasoning' in the consolidated factor structure. The second factor which was extracted very closely resembles Factor 6 'Negative attributes' from the consolidated factor structure as it is most highly loaded on by the characteristics of being disorganised, lazy and stressed. The third factor which was extracted is most highly loaded on by characteristics associated with being creative and having a desire to learn, which taken together do not clearly reflect any of the factors from the consolidated factor structure. The fourth factor is primarily associated with being supportive, positive and being driven. Again, it does not clearly resemble one of the earlier factors but can be seen to have some alignment Factor 2 'Conscientiousness' and Factor 3 'Drive'. The fifth factor, however, more clearly reflects Factor 2 'Conscientiousness' from the consolidated factor structure as it is most highly loaded on by the characteristics of detail orientated, thoughtful, economic and empathetic and as the sixth extracted factor is loaded on by the characteristics relating to competence in discipline areas, it closely resembles the previous Factor 4 'Discipline Knowledge'. The final factor to be extracted is only loaded on highly by the three characteristics of being responsible, dedicated and disciplined. It does have low reliability (α = .589) but appears to be reflective of Factor 3 'Drive', at least more so than the fourth extracted factor. Based on these results, while there was not exact alignment with the consolidated factor structure, there were many similarities particularly with the factors associated with negative attributes, conscientiousness, discipline knowledge, and to a lesser degree with the factors of practical problem solving, and drive. Notably, Factor 5 'Reasoning' which had a significant main effect of the participants' country and was rated more important by the Swedish sample was not clearly observed in the EFA solution from the Irish sample and neither was Factor 7 'Inquisitiveness'.
For the Swedish sample, of the 3916 correlations in the correlation matrix, 395 were greater than or equal to .3. An examination of the anti-image correlation matrix revealed that 86 of the 89 anti-image correlations were above .5, and the off-diagonal elements were mostly small (M = -.0105, SD = .1193). The KMO measure of sampling adequacy was .714, above the recommended value of .6 (Kaiser, 1974), and Bartlett's test of sphericity was significant (χ 2 (3916) = 8084.649, p < .001). Therefore, there was a reasonable level of factorability within the data. A parallel analysis suggested a five factor solution for the Swedish sample ( Figure   S2). Please see Table S4 for the full EFA solution.

FIGURE S2
Factor eigenvalues and parallel analysis for the EFA with data from the Swedish sample only. The EFA solution from the Swedish sample, while a five factor solution as opposed to a seven factor solution, has factors with a clearer alignment with the consolidated factor structure than the Irish sample EFA solution. All factors also have acceptable reliability (α > .6). The first factor which was extracted is most heavily loaded on by the characteristics of being ethical, empathetic, humble, supportive, open-minded and honest, and appears to align with Factor 2 'Conscientiousness' from the consolidated factor structure. The second factor from the Swedish sample is loaded on most heavily by the characteristics relating to disciplinary competence, resembling Factor 4 'Discipline knowledge, from the consolidated factor structure. The third extracted factor is primarily described by the characteristics of being strange, nerdy, easily bored, stressed and pessimistic, reflecting Factor 6 'Negative attributes' and the fourth extracted factor pertains to the characteristics of being ambitious, having a good work ethic, and being motivated, determined and positive, and is therefore very similar to F3 'Drive' from the consolidated factor structure. The final factor which was extracted relates to the five characteristics of being quick thinking, intuitive, adaptable, creatively brave and having leadership skills. All of these except for being adaptable were characteristics which loaded on F1 'Practical problem solving' in the consolidated factor structure. Being adaptable did not load heavily on any of the factors in that solution but being resourceful, which is arguably similar, also loaded on F1 'Practical problem solving', and therefore these factors appear to be describing a similar construct.