*2.3. Protocol*

Each participant performed two tasks, anterior and lateral single-leg jump landings, in one day. Therefore, participants were asked to implement either the anterior single-leg jump landing or the lateral single-leg jump landing, while wearing either low, elastic, or high collar football shoes. All of the tasks were first randomized, and then the shoe order was randomized. Prior to data collection, anatomical and tracking reflective markers were placed on the lower limbs, according to the Istituto Ortopedico Rizzoli (IOR) lower limb model [31]. Meanwhile, the shoelaces were tied by an experimenter and the same type of sport socks were worn, in order to avoid the effects of various shoelaces and socks on the results. Participants were provided five practice trials for each task, to become familiar with the reflective markers and tasks. The anterior and lateral single-leg jump landings were normalized by jump distance according to body height, which was 40% and 33% of body height, respectively [33,34]. Additionally, 30 cm and 15 cm hurdles were placed at 10 cm from the edge of the force plate in anterior and lateral single-leg jump landings, respectively. During data collection, participants were positioned at a normalized distance, then they jumped onto the center of the force plate and landed on their dominant leg after receiving the "start" signal from the researcher. For each condition, each participant was required to stabilize as quickly as possible, place their hands on their waist during landing, and remain motionless on the landing leg for 10 s. Trials were discarded

and repeated for the following reasons: (1) moving the foot before jumping, (2) touching or collapsing the hurdle during jumping, or (3) losing balance or removing hands from the waist during landing. To prevent fatigue, 2 min and 5 min breaks were provided between trials and tasks. Trials of each condition were collected for three successful jump landings tasks.

#### *2.4. Data Analysis*

Visual3D software (C-motion, Inc.; Germantown, MD, USA) was used to analyze the marker positions and force plate data, which were filtered with a low-pass Butterworth filter with cut-o ff frequencies of 14 Hz and 50 Hz, respectively. The ankle joint angle was defined using the segmen<sup>t</sup> coordinate system for the virtual foot segment, which set the ankle joint angle to zero degrees in the static standing, to be aligned with the segmen<sup>t</sup> coordinate system for the shank. The ankle joint moment was calculated using Newton–Euler inverse dynamics with the proximal segmen<sup>t</sup> of the shank as the reference segment, which was normalized to each participant's body mass. Ankle joint sti ffness was calculated as the change in ankle joint moment divided by the change in ankle joint angle from initial contact to peak dorsiflexion [35].

The DPSI is the composite of the vertical (VSI), anteroposterior (APSI), and medial-lateral (MLSI) components, and was computed following the method of Wikstrom et al. [30] using the customized Visual3D software. The square root of the mean square deviation of force, which was the fluctuation from the baseline along each axis of the force plate, was calculated. The APSI and MLSI were assessed using the fluctuations from 0, and the VSI was calculated using the fluctuation from the subject's body weight. The square root of the sum of the squares of APSI, MLSI, and VSI constituted total DPSI.

These variables were calculated using the first 3 s following initial contact, identified as the force threshold exceeding 10 N. The time interval of 3 s is recommended by Wikstrom et al. for studies of sports performance [36]. For anterior single-leg jump landings, the variables of interest included: (1) ankle dorsiflexion ROM, which refers to the total ankle dorsiflexion excursion; (2) ankle eversion ROM, which refers to the total ankle eversion excursion; (3) total ankle ROM in the sagittal and frontal planes, which refers to the total angle changes in the ankle joint in both planes; (4) peak ankle plantarflexion moment, which refers to the maximum plantarflexion moment; (5) peak ankle inversion moment, which refers to the maximum inversion moment; (6) ankle joint sti ffness; and (7) APSI, MLSI, VSI, and DPSI, which refer to the assessments of dynamic postural stability. For lateral single-leg jump landings, the variables of interest were similar to the anterior single-leg jump landing, but with two extra variables: (1) ankle inversion ROM, which is the total ankle inversion excursion; and (2) peak eversion moment, which is the maximum eversion moment. The variables of interest are listed in Tables 1 and 2.


**Table 1.** Mean (standard deviation) of biomechanical variables and pairwise post hoc *p*-value (Cohen's dz) in ankle joint during tasks in the high-, elastic-, and low collar shoe conditions.

Note. \* represents a significant difference within a subject factor. High (H), Elastic (E), and Low (L) represent three football shoe conditions: high collar, elastic collar, and low collar, respectively.


**Table 2.** Mean (standard deviation) of dynamic postural stability index and pairwise post hoc *p*-value (Cohen's dz) during tasks in the high-, elastic-, and low collar shoe conditions.

Note. \* represents a significant difference within a subject factor. High (H), Elastic (E), and Low (L) represent three football shoe conditions: high collar, elastic collar, and low collar, respectively.

#### *2.5. Statistical Analyses*

The residual of each dependent variable was assessed for normality using a one-sample Kolmogorov–Smirnov test ( α = 0.05). Di fferences between shoe conditions were examined using two (for anterior and lateral single-leg jump landings) one-way within-subject analyses of variance (ANOVA). Pairwise post hoc analyses were conducted to assess significant di fferences in the main effects. Wilks's Λ and e ffect size (ηp2) were calculated, and Cohen's dz e ffect sizes were used to interpret the e ffect of pairwise comparisons. An alpha level of 0.05 was used for statistical analysis. SPSS (19.0, IBM Inc.; Chicago, IL, USA) was used to conduct all statistical analyses.
