Disciplinarity and sport science in Europe: A statistical and sociological study of ECSS conference abstracts

Abstract

Abstracts of European College of Sports Science conferences (1995–2014) are studied. The number of abstracts has been increasing regularly (+90 per year). This rise is in recent years largely due to extra-European countries. The magnitude and accumulation of the different topics of discussion are examined. An operational criterion determines four stages of evolution of a topic: social network, cluster, specialty, and discipline. The scientific production can, therefore, be classified as disciplinary or non-disciplinary. The disciplinary part is more important but has been less dynamic recently. The cognitive content of sport science is then explored through a multidimensional scaling of the topics based on the keywords used in the abstracts. Three areas are visible: social sciences and humanities, sports medicine and physiology, and biomechanics and neurophysiology. According to the field theory of Bourdieu (1975), three scientific habitus are distinguished. The logic of academic disciplinary excellence is the consequence of the autonomy of this scientific field, its closure, peer-review process, and barriers to entry. The distribution of scientific capital and professional capital is unequal across the three areas. Basically, conservation strategies of academic disciplinary excellence are predicted in biomechanics and neurophysiology, subversion strategies of interdisciplinarity based on professional concerns can appear in the sports medicine and physiology area, and critical strategies of interdisciplinarity based on social utility in social sciences and humanities. Moreover, additional tensions within these areas are depicted. Lastly methods based on co-citations of disciplines and boundary objects are proposed to find tangible patterns of multidisciplinarity confirming these strategies.

Keywords:

• Sociology of science
• sport science
• Europe
• disciplines
• interdisciplinarity
• Bourdieu

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. The full list is given in Table I.

2. Both databases are maintained by Sportools GmbH.

3. Name and surname are separated by a comma; moreover, authors are also distinguished by commas. That is to say, the number of authors is the number of commas plus one divided by two.

4. No statistically significant autocorrelation as tested by a Durbin–Watson test.

5. The production of the tenth congress was divided into two equal parts, one including the first years and the other the last years.

6. In a Bonferroni-adjusted studentised residuals test (Cook & Weisberg, 1982) of the previous linear model.

7. For the first congress and the final one, the denominator was the participation of nationals in the closest congress.

8. w is Cohen’s effect size (Cohen, 1988).

9. Considering only the 36 countries counting more than 100 abstracts.

10. Eurostat estimation (2014).

11. French National Institute of statistical and Economic Information estimation (2013).

12. At least in Europe, yet the section “Statistics in Sports” of the American Statistical Association was founded in 1992, a dedicated Journal of Quantitative Analysis in Sports exists, and several collective works about Sports statistics were published.

13. The total number of different words in the 23,074 abstracts was counted and is thereafter called m. For each topic (numbered i from 1 to 20), the frequencies of each word employed (j = 1 … m) were computed then the corresponding proportions p_ij (Σ_jp_ij = 1). A similarity s_ii′= Σ_jmin(p_ij,p_i′j) between topic i and topic i′ can thus be calculated and, therefore, a distance d_ii′ = 1 – s_ii′. The (20 × 20) matrix of distances is finally analysed through multidimensional scaling (Cox & Cox, 2001). Word j can then be located onto the factor map by barycentric calculations using weights p_ij. Only the 10 most-cited words for each topic were used, leading to 101 unique words in Figure 3.

14. Simmonot (1988).

15. See also Renson (1991) though with a different meaning.

16. Interestingly, there is no reference to clinical studies.

17. Peretti-Watel (2013).

18. The chi-squared statistic is the sum of the squared Pearson’s residuals. A Pearson’s residual greater than 2 in absolute value is considered as important (Agresti, 2007).