Reliability of a field based 2D:4D measurement technique in children R.M. Ranson a , S.R. Taylor a, ⁎, G. Stratton b, c a Sports and Exercise Sciences, Glyndwr University, Wrexham LL11 2AW, UK b Applied Sport Technology Exercise Medicine Research Centre (A-STEM), Swansea University, Singleton Park, Swansea SA2 8PP, UK c School of Sport Exercise and Health Sciences, University of Western Australia, Crawley, WA 6160, Australia abstract article info Article history: Received 30 October 2012 Received in revised form 27 February 2013 Accepted 5 March 2013 Keywords: Digit ratio Coefficient of variation Correlation coefficients Limits of agreement Children Health-related fitness Background: There is limited literature on the relationship between second to fourth finger digit ratio (2D:4D) and health- and skill-related fitness in children. To examine this relationship it is important to estab- lish a reliable method of assessing 2D:4D for use with large groups of children. Aim: The aim of the study was to examine the reliability of a field-based 2D:4D measure in children. Methods/research design: Fifty 8–11 year olds had 2D:4D of the right hand measured using a Perspex table top, a digital camera, and Adobe Photoshop software. Second to fourth finger digit ratio (and 2D and 4D) intra-observer and inter-observer reliabilities were assessed on the same day and intraobserver reliability was measured between days. Limits of agreement (LoA), coefficient of variation (CV) and Pearson's correla- tion coefficient were used for statistical analysis. Results: High correlation coefficients (r = 0.95–0.99) and low CV's (0.4–1.2%) were reported for intra- and inter-observer reliabilities on the same day and between days. LoA revealed negligible systematic bias with random error ranging from 0.02 to 0.12. Conclusion: These findings suggest that 2D:4D (and 2D and 4D) assessment in children using digital photog- raphy provides a reliable measure of 2D:4D that can be used during field-based testing. © 2013 Elsevier Ireland Ltd. All rights reserved. 1. Introduction The second and fourth digit ratio (2D:4D) of the hand is established during foetal development and caused by prenatal androgens [1,2]. The digit ratio has been related to a number of attributes and behaviours in health-related fitness and sporting attributes in both males and females [3], in females only [4] and males only [5]. Digit ratio has been shown to be negatively associated with ability in sports such as skiing, football, middle distance and endurance running, which are dependent upon an efficient cardiovascular system [6]. Most of the current literature on 2D:4D and sporting performance are based on rankings or selection which may be prone to bias. Research investigating performance out- comes are rare but 2D:4D was positively related to maximum voluntary contraction using a hand grip dynamometer in men but not women [7] but 2D:4D was not related to _ VO2max in boys [8]. The 2D:4D literature is limited in children, particularly in large cohorts. In order to examine the relationship between digit ratio and fitness in children it is impor- tant to establish a reliable method of measuring 2D:4D that can be used in the field when undertaking fitness testing with a large group of children. Previous methods are unsuitable for measuring 2D:4D in the field, as it is well established that children are less cooperative than adults [9]. The measurement of 2D:4D needs to be sensitive enough to detect the lengths of the second (2D) and fourth (4D) digits accurately to enable the precise calculation of 2D:4D [2]. Second digit and 4D lengths have been measured in a number of ways, both directly and indirectly, for example using callipers [10,11,5], rulers [12], market tubes [13], radiography [14], or computer software [15–17], from ac- tual hands, their scans, photocopies, or photographs, and all have their limitations (financial and feasibility), especially for use with children. A range of statistical analyses have been employed to assess the reliability of the various 2D:4D measurement techniques. These have included intraclass correlation coefficients (ICC) ranging from r = 0.91 to 0.98 for the measurement of hands using photographs [10,11,5] and correlation coefficients r = 0.93–0.97 using scans of the hand and Adobe Photoshop software [18]. Similarly Manning et al. [2] reported correlation coefficients of r = 0.96–0.97 for 2D and 4D measured separately using callipers directly on the hand. Histori- cally a significant test–retest correlation coefficient of r ≥ 0.8 has been used to indicate that equipment and tests are suitably reliable [19]. However, this method of assessing reliability has been criticised, since a high correlation coefficient indicates the strength of the rela- tionship between the test–retest values, not the agreement between them [19–21]. Statistical methods to assess reliability have evolved and authors advocate the use of various tests, and calculations, to de- termine whether a method is reliable. Atkinson and Nevill [22] sup- port the use of limits of agreement (LoA) to assess reliability but this has been criticised as the method has been reported to be biased Early Human Development 89 (2013) 589–592 ⁎ Corresponding author at: Sports and Exercise Sciences, Glyndwr University, Mold Road, Wrexham LL11 2AW, Wales, UK. Tel.: +44 1978 293092. E-mail address: s.taylor@glyndwr.ac.uk (S.R. Taylor). 0378-3782/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.earlhumdev.2013.03.002 Contents lists available at SciVerse ScienceDirect Early Human Development journal homepage: www.elsevier.com/locate/earlhumdev