Prior to conducting the study, sample size calculations were performed according to the recommendations of several authors (17–19) and the G*Power for simulations (20). Fritz and MacKinnon’s recommendations for small a and β paths (the most restrictive condition when 1-β = .80) indicate a minimum of 558 individuals. For a simple moderation simulation (one moderator) using f2 = .15, α = .05, and 1-β = .80, a sample size greater than 77 is recommended. Considering that no clear indication exists on how to calculate an adequate sample size for the model under study, we selected a sample size that would (i) surpass the simulations performed under the most restrictive conditions of mediation and (ii) the minimum number of cases for the number of interactions to be tested.

Participants were included in the study if they met the following inclusion criteria: aged between 60- and 80-years old living in the community, not suffering from cognitive disorders, not having medical contraindications for physical fitness and body composition tests, voluntarily participating, and being able to move autonomously or with the use of a manual mobility device.

Before data collection began, approval was obtained from the ethics committee (numbers omitted for review purposes). Subsequently, an existing community program in central Portugal with over 1000 registered participants, was contacted for convenience. After obtaining authorization from the municipal council, the program directors were contacted, and the objectives of the study were explained to them. All the participants met the inclusion criteria. Potential participants were informed of the study objective, anonymity, voluntary participation, and withdrawal at any time. Physical fitness tests and body composition measurements were also performed. The technical team received specific training on how to collect the data. All potential participants who wished to participate in this study signed an informed consent form before undergoing the physical fitness tests and body composition measurements. All participants who consented to participate underwent the tests at the beginning of the sport season, namely in September and October, following a one-month break during the vacation program.

The 30-second chair stand test (21) was used to assess lower body strength. During the test, the participant was seated in a chair, crossed their arms, and instructed to perform repeated movements of standing up and sitting down in the chair for 30 s, with the aim of completing as many repetitions as possible.

The chair sit and reach test (21) was used to assess lower body flexibility. During the test, the participant sat in a chair, extended the dominant leg forward with the foot flexed, then leaned forward, trying to reach as far as possible towards the feet, keeping the knee in complete extension with one palm overlapped on the other, and the reached (or not reached) distance was recorded.

The 2-minute step test (21) was employed to evaluate aerobic endurance. The participants were instructed to step up and down a 12-inch step for 2 minutes continuously. This test assessed the efficiency of the cardiovascular system in sustaining physical activity over a prolonged period, and the maximum steps were recorded.

The timed up-and-go test (21) allows the evaluation of agility and dynamic balance, indicating the risk of falls. It consisted of walking a distance in a straight line, starting from a seated position in a chair, going to a cone placed 2.44m away, circumventing it, and returning to the initial chair. During the test explanation, the importance of rising, walking quickly without running, circling the cone, returning to the chair, and sitting down again was emphasized. An initial measurement was conducted for familiarization. The time required to complete the test was determined.

Height was measured using a portable Seca 213 stadiometer with an integrated leveler (GmbH & Co. KG; Hamburg, Germany). Participants were positioned with their backs on a wall where the stadiometer was fixed, ensuring contact of the back of the head, back, and buttocks with the wall. With the weight evenly distributed on both bare feet and heels together and touching the wall, participants were instructed to look straight ahead, inhale, and hold their breath. The stadiometer gauge was then lowered to the highest point of the head, and measurements were recorded. Weight was measured using the Seca 761 brand scale (GmbH & Co. KG, Hamburg, Germany) was placed on a stable surface. The participants were weighed without shoes or coats, keeping the rest of their clothing. They were instructed to step on the scale with their feet on the scale’s reference, looking straight ahead, and remaining still during the measurement. Body mass index (BMI) was calculated using the following formula: BMI (kg/m2) = weight (kg) / height2 (m).

All statistical analyses were performed using the IBM SPSS Statistics 29.0 (IBM, Armonk, USA). Descriptive statistics were calculated using the mean and standard deviation. Bivariate correlations were also calculated to understand possible associations between variables. Subsequently, a mediated-moderation model was developed following the recommendations proposed by Hayes et al. (18) using PROCESS version 4.2, and model 5 for hypothesis testing. This model assumes the existence of an independent variable (lower body strength), parallel mediators (lower body flexibility and aerobic endurance), a dependent variable (agility and dynamic balance), and a moderator (BMI as a continuous variable). The values were conditioned at the 16th, 50th, and 84th percentiles. A bootstrap with 5,000 samples was used for the 95% confidence interval estimation, and significant effects were considered if the interval did not encompass zero.