| contrast | estimate | SE | df | p.value | lower.CL | upper.CL |
|---|---|---|---|---|---|---|
| altitude - sea | 7.07 | 4.24 | 28.09 | 0.11 | -1.6 | 15.75 |
Additional Information
Linear Mixed-Effects Model Examining the Impact of Altitude on RPE Breathe Scores on Match Day, Accounting for Individual Repeated Measures
The p-value was not statistically significant, P = 0.11, and the 95% confidence interval -1.6 to 15.75 is wide and includes zero, suggesting that the observed effect could be due to chance. If you looked at the p-values and confidence intervals alone, it would suggest that, based on this data, altitude does not have a clear impact on RPE during matches. Although recent academic opinion has criticised the limitations of relying solely on p-values to determine the usefulness of results (Jafari & Ansari-Pour, 2018).
RPE Legs was looked at as well as RPE Breathe but no significant changes were found that would benefit passing onto the team coach other than the correlation between an increase in RPE Breathe and an increase of RPE Legs.
This graph shows the impact on RPE Legs, sea level to altitude on a match day.
The table shows the mean RPE legs, sea level to altitude.
| altitude_code | mean_rpe_leg |
|---|---|
| sea | 82.9 |
| altitude | 77.8 |
Scatter plot of RPE Legs at altitude on match day.
RPE Legs - Players on longer than 70 minutes versus those on less than 70 minutes
Mean RPE Legs for players on >70 minutes and <70 minutes match day.
| time_group | altitude_code | mean_rpe_leg |
|---|---|---|
| 70+ mins | altitude | 81.1 |
| 70+ mins | sea | 82.9 |
| <70 mins | altitude | 61.7 |
| <70 mins | sea | NaN |
Analyzing the Effect of Altitude on RPE Legs Scores Using a Linear Mixed-Effects Model
| contrast | estimate | SE | df | p.value | lower.CL | upper.CL |
|---|---|---|---|---|---|---|
| altitude - sea | -4.25 | 3.85 | 25.38 | 0.28 | -12.18 | 3.67 |
RPE Legs scores were slightly lower at altitude compared to sea level, with an average difference of -4.25, which given the results from RPE Breathe were not expected. However, this difference isn’t statistically significant P = 0.28 and the confidence interval -12.18 to 3.67 is wide and includes zero, meaning the true effect may be minimal, absent or even reversed. Again, academics have cautioned against the reliance on p-values alone to determine the significance of results (Pineda & Sirota, 2018).
Playing at altitude
Football is a global sport played year-round across varying environments, meaning players are frequently exposed to environmental triggers such as temperature and altitude, which can significantly impact performance (Draper et al., 2023). High altitude, in particular, poses primarily physiological challenges for athletes and due to this FIFA initially banned international matches from being played above 2,500 meters (8,200 feet) citing player health and the potential distortion of competition outcomes (BBC News, 2007).
This decision was controversial, especially in high altitude countries and as a result, FIFA later revised the rule, allowing matches at high altitude provided players had sufficient time to acclimatise—one week for altitudes above 2,500 meters and 15 days for altitudes above 3,000 meters (BBC News, 2008). This decision was made based on evidence that the effects of high altitude despite being more pronounced initially, can lessen with acclimatisation (Levine, Stray‐Gundersen & Mehta, 2008). Now while these acclimatisation guidelines are ideal, experts acknowledged that due to the hectic football calendar, extended periods of acclimatisation are unrealistic to most teams (D’Hooghe, 2013).
However, in the current data set it is based on Match Day 2, Match Day 1 and then a match giving the players 96 hours to acclimatise. It was also stated in the data set that inclusion was based on a match at altitude >1000m not falling on a congested week where another match was played. Whilst this excludes factors such as fatigue due to too many matches played in a short period, which would make the results more reliable; possible inadequate acclimatisation time, could be considered a limitation and should be considered when interpreting findings related to altitude performance.
So why would altitude impact RPE, particularly breathing
At high altitudes, lower oxygen availability reduces aerobic exercise capacity, resulting in decreased maximal oxygen uptake (VO₂max), greater perceived exertion, and prolonged recovery (Levine, Stray‐Gundersen & Mehta, 2008). The body compensates for hypoxia through elevated breathing and heart rates, yet oxygen delivery to muscles remains insufficient, negatively impacting performance (Gore et al., 2008).
A study by Aliverti et al. (2011) found that the RPE Breathe during exercise is directly related to how hard the breathing muscles are working, regardless of altitude or oxygen availability. In other words, according to Aliverti et al. (2011), at high altitudes people reported the same breathing effort if the workload on their breathing muscles remained constant. The study was conducted on healthy but non-athletic individuals, meaning the findings may not apply to trained athletes, who may experience different responses due to changes in VO₂max at altitude. Additionally, the small sample size of only 10 participants limits the generalisability of the results. The group also had varying fitness levels, potentially influenced by age or training history, which may have affected the outcomes. This suggests that results may differ for elite athletic populations.
Fulton et al. (2018) conducted a study exploring how trained distance runners maintain a breathing pattern synced with their stride, known as locomotor-respiratory coupling. The study found that even when running at a simulated moderate altitude the runner did need to breathe more frequently due to reduced oxygen but were still able to coordinate their breathing with their steps (Fulton et al., 2018). Fulton et al. (2018) suggests that their training helps them adapt their breathing efficiently, potentially conserving energy. These finding may not be applicable to footballers due to the sustained aerobic nature of endurance running, unlike football which has aerobic/anaerobic activity, with periods of high intensity and short, rapid recovery. It is also simulated altitude which may leave out other environmental factors impacting performance.
Nassis (2013) conducted an observational study used secondary data to examine the effects of altitude on football performance. Nassis (2013) states that while lab studies consistently show that altitude impairs endurance due to reduced oxygen availability, there is limited research on its real-world impact in football and other team sports. Nassis (2013) goes on to conclude that interestingly, the negative effects of higher altitude on VO2max, tend to be more pronounced in well-trained athletes, who may experience greater perceived effort and fatigue when competing at altitude compared to sea level. While this is not the strongest evidence coming from a collection of general articles, the author does point out the high volume of laboratory-based research on altitude compared to real world research, particularly around football, despite FIFA’s previous health concerns.
Substitution reasoning
Substitutions that are made in the second half of a match are usually done to counteract player fatigue, and can be influenced by other factors such as match status, opposition strength, and location (Hills et al., 2018l; Gómez et al., 2016),
Data was extracted for RPE Breathe and RPE Legs, comparing those who played fewer than 70 minutes with those who played more than 70 minutes during a match. The aim was to determine whether these data could be utilised to inform and optimise substitution strategies.
RPE Breathe showed an effect in players who played longer than 70 minutes, while RPE Legs did not show any increase. These findings suggest that RPE Breathe results could potentially be used to inform substitution decisions in the latter part of a match played at altitude.
Silva and Swartz, (2016) in their analysis of substitution times based on data from the 2010 world cup and the 2009/10 season from 4 top football leagues in Europe and found no distinct period in the second half where substitutions provide a clear performance benefit. While Silva and Swartz (2016) collected extensive data in their analysis, the data was from Europe and it is unclear whether any data was gathered at altitude. There is a need for more refined, data-driven decision-making support, particularly when optimising substitution strategies to enhance overall team performance (Van Roy et al., 2023).”
Number of observations and how it impacts the reliability of results.
Match day
Both sea and altitude had the same number of observations on a match day adding more reliability to the results that altitude does in fact impact RPE Breathe, although higher numbers of observations overall would correlate this and perhaps more statistically significant results would be achieved.
| altitude_code | n |
|---|---|
| altitude | 21 |
| sea | 21 |
Challenges in data collection
On of the biggest challenges in sports science research is the reliance on small sample sizes, which undermines the replicability and generalisability of findings, small sample sizes can also reduce the likelihood of detecting small, but meaningful effects (Mesquida et al., 2022; Schweizer and Furley, 2016). Collecting data from larger cohorts enhances the accuracy and reliability of research findings by minimising the impact of individual variability, making results more representative of true population values (Hecksteden et al., 2022).
Sports science studies conducted with small sample sizes could be comparable to the small number of observations recorded at both sea level and altitude in this data set. Without collecting a higher number of observations it is difficult to achieve reliable and statistically significant results on the impact of altitude on the perceived rate of exertion. Although not statistically significant, RPE Breathe did show a trend where it increased at altitude, it would be wise not to dismiss this and use the results of this data to manage load whilst playing at altitude.
P values
Both p-values for RPE Breathe and RPE legs were above the threshold to be statistically significant. P-values are widely used to assess the statistical significance of observed results and estimates how likely it is that the observed results occurred by chance under the assumption that the null hypothesis is true (Vidgen & Yasseri, 2016; Concato & Hartigan, 2016). Traditionally, results with a p-value below 0.05 are considered statistically significant, originally intended as a guideline it has become a rigid standard potentially leading to the dismissal of meaningful effects of results with higher p-values (Concato & Hartigan, 2016).
There has been increased criticism in recent years of the reliance on p-values, especially in fields like biomedicine where large datasets, reproducibility issues, and evolving analytic methods complicate interpretation (Jafari & Ansari-Pour, 2018; Pineda & Sirota, 2018). P-values are often mistaken for proof of a real effect, but they just reflect the likelihood of the data if no effect exists and can be distorted by sample size or multiple testing (Lytsy, 2018). Vidgen and Yasseri (2016) state that this over-reliance on p-values can lead to false positive, results that appear significant by chance alone, going on to suggest that findings should not hinge on a single p-value.
In this analysis, the p-values related to perceived exertion at altitude do not meet conventional thresholds for statistical significance, which may be due to the small sample size. However, given the evidence around the physiological effects of altitude and the concerns from organisations like FIFA about player health in matches played at high altitude, these findings perhaps should not be dismissed outright. It could be argued that with a larger sample, the p-values might decrease, potentially revealing statistically significant effects.
References:
Abbott, H. and Taber, C., (2021). ‘How to collect rating of perceived exertion to monitor athlete training load’. Journal of Physical Education, Recreation & Dance, 92(9), pp.5-10.
Aliverti, A., Kayser, B., Mauro, A.L., Quaranta, M., Pompilio, P., Dellacà, R.L., Ora, J., Biasco, L., Cavalleri, L., Pomidori, L. and Cogo, A., (2011). ‘Respiratory and leg muscles perceived exertion during exercise at altitude’. Respiratory physiology & neurobiology, 177(2), pp.162-168.
BBC News (2007) Fifa bans high-altitude football. Available at: http://news.bbc.co.uk/1/hi/world/americas/6697159.stm (Accessed: 23 April 2025).
BBC News (2008) Fifa suspends altitude match ban. Available at: http://news.bbc.co.uk/1/hi/world/americas/7422293.stm (Accessed: 23 April 2025).
Concato, J. and Hartigan, J.A., (2016). ‘P values: from suggestion to superstition’. Journal of Investigative Medicine, 64(7), pp.1166-1171.
D’Hooghe, M., (2013). ‘Football and altitude: a FIFA vision’. British Journal of Sports Medicine, 47(Suppl 1), p.i1. Available at: https://doi.org/10.1136/bjsports-2013-093006
Draper, G., Wright, M.D., Ishida, A., Chesterton, P., Portas, M. and Atkinson, G., (2023). ‘Do environmental temperatures and altitudes affect physical outputs of elite football athletes in match conditions? A systematic review of the ‘real world’studies’. Science and Medicine in Football, 7(1), pp.81-92.
Fulton, T.J., Paris, H.L., Stickford, A.S., Gruber, A.H., Mickleborough, T.D. and Chapman, R.F., (2018). ‘Locomotor-respiratory coupling is maintained in simulated moderate altitude in trained distance runners’. Journal of Applied Physiology, 125(1), pp.1-7.
Gomez, M.A., Lago-Peñas, C. and Owen, L.A., (2016). ‘The influence of substitutions on elite soccer teams’ performance’. International Journal of Performance Analysis in Sport, 16(2), pp.553-568.
Gore, C.J., McSharry, P.E., Hewitt, A.J. and Saunders, P.U., (2008). ‘Preparation for football competition at moderate to high altitude’. Scandinavian journal of medicine & science in sports, 18, pp.85-95.
Hecksteden, A., Kellner, R. and Donath, L., (2022). ‘Dealing with small samples in football research’. Science and medicine in football, 6(3), pp.389-397.
Hills, S.P., Barwood, M.J., Radcliffe, J.N., Cooke, C.B., Kilduff, L.P., Cook, C.J. and Russell, M., (2018). ‘Profiling the responses of soccer substitutes: A review of current literature’. Sports Medicine, 48, pp.2255-2269.
Jafari, M. and Ansari-Pour, N., (2018). ‘Why, when and how to adjust your P values?’. Cell Journal (Yakhteh), 20(4), p.604.
Levine, B.D., Stray‐Gundersen, J. and Mehta, R.D. (2008) ‘Effect of altitude on football performance’, Scandinavian Journal of Medicine & Science in Sports, 18, pp. 76–84.
Lorenzo-Martínez, M., Padrón-Cabo, A., Rey, E. and Memmert, D., (2010). ‘Analysis of physical and technical performance of substitute players in professional soccer’. Research Quarterly for Exercise and Sport, 92(4), pp.599-606.
Lytsy, P., (2018). ‘P in the right place: Revisiting the evidential value of P‐values’. Journal of Evidence‐Based Medicine, 11(4), pp.288-291.
Mesquida, C., Murphy, J., Lakens, D. and Warne, J., (2022). ‘Replication concerns in sports and exercise science: a narrative review of selected methodological issues in the field’. Royal Society Open Science, 9(12), p.220946.
Nakamura, F.Y., Moreira, A. and Aoki, M.S., (2010). Monitoramento da carga de treinamento: a percepção subjetiva do esforço da sessão é um método confiável. Revista da Educação Física/UEM, 21(1), pp.1-11.
Nassis, G.P., 2013. Effect of altitude on football performance: analysis of the (2010) ‘FIFA World Cup Data’. The Journal of Strength & Conditioning Research, 27(3), pp.703-707.
Pineda, S. and Sirota, M., (2018). ‘Determining significance in the new era for p values’. Journal of pediatric gastroenterology and nutrition, 67(5), pp.547-548.
Schweizer, G. and Furley, P., (2016). ‘Reproducible research in sport and exercise psychology: The role of sample sizes’. Psychology of Sport and Exercise, 23, pp.114-122.
Silva, R.M. and Swartz, T.B., (2016). ‘Analysis of substitution times in soccer’. Journal of Quantitative Analysis in Sports, 12(3), pp.113-122.
Vidgen, B. and Yasseri, T., (2016). ‘P-values: misunderstood and misused’. Frontiers in Physics, 4, p.6.
Van Roy, M., Robberechts, P., Yang, W.C., De Raedt, L. and Davis, J., (2023). ‘A Markov framework for learning and reasoning about strategies in professional soccer’. Journal of Artificial Intelligence Research, 77, pp.517-562.