Share this article on:

Monitoring the Intended and Perceived Training Load of a Professional Futsal Team Over 45 Weeks: A Case Study

Rabelo, Felipe N.1; Pasquarelli, Bruno N.2,3; Gonçalves, Bruno3; Matzenbacher, Fernando4; Campos, Fábio A.D.5,6; Sampaio, Jaime3; Nakamura, Fábio Y.7,8

The Journal of Strength & Conditioning Research: January 2016 - Volume 30 - Issue 1 - p 134–140
doi: 10.1519/JSC.0000000000001090
Original Research

Rabelo, FN, Pasquarelli, BN, Gonçalves, B, Matzenbacher, F, Campos, FAD, Sampaio, J, and Nakamura, FY. Monitoring the intended and perceived training load of a professional futsal team over 45 weeks: a case study. J Strength Cond Res 30(1): 134–140, 2016—The aims of this study were to compare the training load intended by a coach with the training load perceived by the players, over a 45-week professional futsal team season and to compare the variation of session rating of perceived exertion (sRPE) across different periods of the season. Eighteen Brazilian professional futsal players participated in the study. The players' rating of perceived exertion (RPE) and coach's rating of intended exertion (RIE) were collected daily throughout the study. To compare the sRPE variation, the season was divided into 4 periods as follows: preseason (PRE-SEASON), first competitive period (COMP1), intercompetition period (INTER-COMP), and second competitive period (COMP2). Based on the cluster analyses, the training sessions were classified into 3 different intensity zones (low, moderate, and high). In all the season periods and intensity zones, the players' RPEs were lower than the coach's RIE. In the low-intensity zone, the INTER-COMP demonstrated higher sRPE values than did the other periods. In the moderate-intensity zones, the INTER-COMP and COMP2 were similar. Finally, for the high-zone intensity, the PRE-SEASON demonstrated higher values than did the other periods. We concluded that the coach overestimated the training load reported by the players in almost all intensity training zones and season periods. The RPE scale does not seem to be a suitable tool when used by the coach for intended training load. Therefore, technical staff should constantly review the training goals in each season period through daily training load control, always taking into consideration the possibilities and limitations of the RPE method.

1Sport Club Corinthians Paulista, São Paulo, Brazil;

2Campinas State University, Campinas, São Paulo, Brazil;

3Research Centre in Sports Sciences, Health Sciences and Human Development, CIDESD, CreativeLab Research Community, University of Tra[Combining Acute Accent]s-os-Montes e Alto Douro, Vila Real, Portugal;

4Passo Fundo University, Passo Fundo, São Paulo, Brazil;

5Air Force Academy, Pirassununga, São Paulo, Brazil;

6Methodist University of Piracicaba, Piracicaba, São Paulo, Brazil;

7Londrina State University, Londrina, PR, Brazil; and

8Nucleus of High Performance in Sport, São Paulo, São Paulo, Brazil

Address correspondence to Felipe N. Rabelo,

Back to Top | Article Outline


From a theoretical standpoint, sports performance is expected to improve when training load and recovery alternate effectively (22). When the players train harder than planned for prolonged periods, they may exhibit signs of underrecovery and maladaptation (24). In contrast, if they do not perform sufficiently on intended intensified days, the training stimuli may not be sufficient to elicit adequate adaptations (34). In addition, inaccurate training load prescription and monitoring may lead to highly monotonous training programs and sessions, which is a key determinant to increase the risk of over-training, injuries, and illness (7,11,33).

Rating of perceived exertion (RPE) is a commonly used measure to provide information about training intensity (5,10,12,15). It is highly correlated with markers of physiological strain, such as heart rate and blood lactate concentration (4,19,27). Although some debate exists in the literature, studies performed with team sports players provide support to the validity and suitability of the training load derived from RPE measure (8,12,19,25,27). Because the adoption of traditional periodization models is not straightforward (22), and given the high degree of unpredictability in team sports routines (e.g., difficulties in individualizing loads, traveling, and tournament characteristics), monitoring perceived training, competitive loads, and the effects is both complex and challenging. Added to this, the players' ability to report the perceived effort can vary both within the team and individually according to different training workloads. For coaches, this presents a difficult challenge because they do not perceive the players' effort. Thus, the intended training load is the expected internal stimulus in accordance with the handling of external variables (i.e., task constraints, mode of exercise, intensity, duration, number of repetitions, and series). Recent research has shown that players may have a tendency to report higher intensity in low and moderate training sessions and report lower intensity in high-intensity sessions, when compared with the coaches' intended training intensity (6). In fact, a daily control should provide information to adequately adjust the training loads during the microcycle periods. Available research proposes discriminating the daily training intensities using the RPE scale (1,6); however, sports scientists constantly seek more accurate and valid approaches.

Incorrect interpretation of RPE data may lead to errors in training control and subsequent planning. Thereby, daily control with feedback to the coaches is a key to enhancing performance and decreasing the risk of injury and deleterious training effects (7,22). Accordingly, the comparison of the training load intended by the coaching staff and the actual session rating of perceived exertion (sRPE) reported by the athletes has become a relevant topic of research (6,29,32,33). Available studies on this theme have been performed with individual sports athletes, in which case it is easier to quantify and manage the training loads (29,33). In team sports, the longest reported period of observation comparing coach and player ratings lasted for 15 weeks (32). Therefore, it remains to be established whether the agreement of sRPE between team sports coaches and players is consistent over prolonged periods of time, such as an entire season. Accordingly, the first aim of this study was to describe and compare the training load intended by the coach with the training load perceived by the players, over a 45-week professional futsal team season.

Moreover, different characteristics were shown for each team sports season period, considering the preseason (8,20,26,28) and the competitive season (20,21,26,28). Depending on the competitive schedule, the tactical-technical-physical aims of the training sessions can change over the season, whereas the athletes' physical condition levels can influence the reported perception of exertion (25). Thereby, each period of the season should be investigated more accurately. Hence, the second aim of this study was to compare the variation of sRPE across different periods of the season.

Back to Top | Article Outline


Experimental Approach to the Problem

The study comprised 45 weeks of planned training (3–6 training session days per week), resulting in 314 days of training and friendly/competitive matches based on the competitive calendar. The season was divided into 4 periods as follows: preseason (PRE-SEASON), first competitive period (COMP1), intercompetition period (INTER-COMP), and second competitive period (COMP2). Within each period, the training microcycles were divided into training session day, day off, and match day (Table 1). The players' RPE and coach's rating of intended exertion (RIE) were collected throughout the study (total training sessions, n = 157).

Table 1

Table 1

Back to Top | Article Outline


Eighteen professional futsal players, playing at the highest level in the Brazilian futsal league, participated in this study. The players had been involved in a planned training process for at least 8 years (age: 24.6 ± 3.8 years; height: 1.76 ± 0.1 m; weight: 77.0 ± 9.6 kg, percentage body fat: 14.9 ± 4.1%). The team competed in 2 regional competitions and one national competition, covering distances ranging from 300 to 2000 km per away match by bus over the competitive periods. The study was approved by the Ethical Committee for Research Involving Human Subjects, and written informed consent was obtained from subjects and the club's directive board.

Back to Top | Article Outline


Before each training session, the coach's RIE (7,29,32,33) was registered using 0–10 Borg's scale (5) translated to the Portuguese version (30). The coach had more than 3 years' experience of using the tool and was familiarized with the scale. The predicted load was planned weekly according to the season period and match events. Players' RPE was obtained using the same scale (5), approximately 30 minutes after each training session to ensure a global perceived effort for the entire session. The players were familiarized with the scale, the data were collected individually, and they did not have access to either the training load intended by the coach nor the perceived exertion of their teammates, avoiding an interference of the training intensity judgment.

Back to Top | Article Outline

Statistical Analyses

A 2-step cluster with log-likelihood as the distance measure and Schwartz's Bayesian criterion was performed to classify the training load into low-, moderate- and high-session efforts according to: (a) Coach's RIE and players' RPE and (b) players' RPE. This method differs from traditional clustering techniques through the handling of categorical variables (assuming variables to be independent), automatic selection of the number of clusters (it automatically determines the optimal number of clusters), and scalability (by constructing a cluster membership) (35). Also, the RPE scores should be interpreted according to the different season periods. In functional understanding, the same RPE value may present different implications for the players' performance monitoring over the season.

Therefore, both cluster analyses were used to classify the training load into 3 maximally different intensity zones (designated as low, moderate, and high). The first analysis compared the RPE with RIE according to the season periods and session demands (i.e., PRE-SEASON players' RPE vs. PRE-SEASON coach's RIE at low, moderate, and high intensities, and so on). The second cluster analysis, which involved only the players' RPE, compared the same training intensities in different season periods (i.e., low-RPE sessions in the COMP1 vs. low-RPE sessions in the PRE-SEASON, INTER-COMP vs. COMP2 period, and so on).

In the first analysis, the variables were ranked according to the predictors' importance, indicating the relative importance of each variable in estimating the model (the sum of the values for all predictors on the display is 1, and the variables' predictor importance provides different weights to support the cluster distribution). In the second analysis, a single variable was used, and subsequently, each clustering variable's description was presented as mean ± SD.

The differences between players' RPE and coach's RIE (first analysis) and the differences between season periods (second analysis) were assessed using a spreadsheet (Microsoft Excel; Microsoft, USA) and using the procedures developed by Hopkins (16) to compare mean values of 2 groups. As proposed, a threshold value of ±1 in raw units was established for the smallest important or harmful effect because this is the usual approach of psychometric scales. The data were presented as mean ± SD. Comparisons were described by the differences in mean values as a fraction of the between-subjects SD (Δ mean/SD) and 90% of the confidence limit. Uncertainty in true differences of comparisons was assessed using mechanistic inferences, based on threshold chances of 5% for substantial magnitudes, thus, when the effect was >5%, it was reported as unclear; the effect was otherwise clear and reported as the magnitude of the observed value. The scale was as follows: 25–75%, possible; 75–95%, likely; 95–99%, very likely; and >99%, almost certain (17).

Finally, the standardized difference in mean values (effect size [ES], with 90% of confidence intervals) for each comparison parameter was calculated. The limit values for the analyses of ES used were those proposed by Hopkins (18): <0.2 (trivial effect), 0.2–0.6 (small effect), 0.6–1.2 (moderate effect), 1.2–2.0 (large effect), and >2.0 (very large effect).

Back to Top | Article Outline


Differences Between Player's and Coach's Perceived and Intended Effort

Results from comparisons between players' RPE and coach's RIE are presented in Table 2. The results showed that in sessions with high-load demand, the coach's RIE was overestimated compared with players' RPE throughout the season. Differences were also found in the low PRE-SEASON and moderate COMP2 periods (Table 2). The ES also showed that in high-load sessions the differences were large in the PRE-SEASON and very large in the COMP1, INTER-COMP, and COMP2 (Figure 1). In the moderate-load sessions, the effects of differences were classified as moderate in the PRE-SEASON, large in the COMP1 and INTER-COMP, and very large in the COMP2. In the low session load, a moderate effect was found only in the PRE-SEASON (Figure 1).

Table 2

Table 2

Figure 1

Figure 1

Back to Top | Article Outline

Players' Rating of Perceived Exertion According to the Season Period and Training Load

Table 3 shows the players' RPE according to the season period and training load. Differences in season periods were more evident when training intensities were low and high. The mean RPE values in the INTER-COMP for the low-intensity training sessions were higher than the COMP1 and COMP2 periods (Table 3). In higher intensity sessions, the PRE-SEASON period was higher than the COMP1 and INTER-COMP periods (Table 3). Furthermore, the COMP2 presented higher RPE mean values and a very large effect compared with COMP1 (Figure 2). Finally, for the moderate training sessions, the COMP1 showed the lowest RPE mean values and a moderate effect when compared with the INTER-COMP and COMP2 periods (Figure 2).

Table 3

Table 3

Figure 2

Figure 2

Back to Top | Article Outline


The aim of this study was to describe and compare the training load intended by the coach with the training load perceived by the players, over a 45-week professional futsal team season. Research currently available is limited to data collected over substantially shorter periods; thus, this study demonstrates an advance in training load research as controlled for a prolonged period. Overall, the results demonstrated that in all season periods and zone intensities, the players perceived a lower training load than that intended by the coach. In addition, a greater discrepancy was found between the coach's RIE and players' RPE toward the end of the season, as well as for moderate and high training loads.

In contrast to this study, results of some studies have shown that coaches underestimated the sRPE values related to professional soccer players for low and moderate training session days (6). In addition, it has been demonstrated that coaches underestimated the athletes' sRPE in volleyball physical training (32) and elite junior tennis players (29). Also, in this study, on high-intensity training days, coaches overestimated the athletes' sRPE (6). Similar results were found with well-trained swimmers (36), moderately trained swimmers (2), and high-level volleyball athletes (31).

Intended training load is a complex procedure for coaches and requires an accurate analysis of habitual data collection through objective and subjective measures, while also taking into account the coaches' experience and perspective (22). Accordingly, some intervening issues must be outlined to explain the difficulties in prescribing and controlling training load, such as the psychological factors and travel schedule. In this study, the team covered distances ranging from 300 to 2000 km per away match by bus. As any disruption in routines, such as eating and sleeping patterns, can inhibit the use of recovery interventions, the recovery process after away matches could have been impaired (13,14). In addition, the wide use of ball-based exercises in groups (i.e., small-sided games and conditioning games) may bias perception of effort because of the possibility of self-controlling the pacing strategy. The awareness of self-positive (or negative) training performance may also have affected players' perception and their behavior concerning the goals of the primary task. Thus, using sRPE as a training intensity control framework should consider more multidimensional player-related information, especially during high-intensity and self-paced activities. This issue warrants future investigation.

The second aim of the study was to compare the training load of the season periods. In the low-intensity zone, the INTER-COMP demonstrated higher sRPE values compared with those of the other periods. In the moderate-intensity zones, the INTER-COMP and COMP2 were similar. Finally, for the high-zone intensity, the PRE-SEASON demonstrated higher values compared with those of the other periods.

The PRE-SEASON was the period with the highest number of training sessions (68.8%, averaging one session every 1.5 days), lasting 120 minutes on average and including only 11 days off. This might have contributed to eliciting a higher perception of effort by the players. Such data indicate that these adjustments in load were a direct attempt to deliver training aimed at promoting physical adaptations along with technical and tactical improvements. In fact, recent findings have demonstrated that PRE-SEASON training is more intense than in-season training in professional football (20) and futsal players (28).

The COMP1 and COMP2 periods presented 46.8% of training session days (averaging one session every 2.1 days) and 38% of total (one session every 2.6 days), respectively. Compared with the COMP1, there were more days off in the COMP2 period (46% vs. 30.5%) and a greater relative number of games (1 game per 6.3 days vs. 1 game per 4.4 days). The COMP1 period presented the lowest values of training intensity in all 3 zones, which could be explained by the greater number of games—38 in total—and the strenuous travel schedule during the national and regional competitions. In contrast, the COMP2 period presented a higher prevalence of high-intensity training (57%), which was different from all the other periods. This training intensification at the end of the season may be due to a lower number of games in the schedule, after elimination from the national championship, followed by focusing on the regional competition playoffs. Thus, during this competitive period, there were more official games and days off, in addition to less training sessions. Therefore, it was necessary for the coach to reduce training load, aiming to provide an adequate recovery process from the previous game and an optimal physical state for the next game (13,21).

In the INTER-COMP period, the training session days corresponded to 61% of total, averaging one session every 1.6 days, presenting more moderate- (63%) and high- (30%) intensity training than low intensity (7%). The INTER-COMP also presented a short duration (39 days) and only one friendly game, allowing greater training load intensification. However, despite similar characteristics between the PRE-SEASON and INTER-COMP periods regarding training load intensification, in the INTER-COMP period, the players already presented greater training load adaptation and tactical-technical-physical condition, which could explain the lower RPE values found, mainly in response to the high-intensity training days. A study verified that physical condition influences RPE values when comparing futsal players of the same team, exposed to similar external training loads (25).

Moreover, a previous study reported that the brain area activated in effort situations is linked to the area of emotional reactions (23). In addition, some factors such as mental fatigue may significantly alter RPE, independent of peripheral signs of metabolic and neuromuscular strain (9). In contrast, self-talk may act to reduce RPE during exercise and increase tolerance to fatigue (3). In view of these neuroscience aspects, further studies should be performed to support the specific mechanisms (and psychological influences) that contribute to RPE.

Therefore, when planning training, coaches should keep in mind the individual characteristics (physical and psychosocial) that affect the internal load of each individual player on the team (6). Despite the results commonly presented as central tendency measures, the coach's staff should consider the interindividual response variability because the risk of overreaching and over-training or underadaptation may be hidden by the players' collective variable responses. Monitoring of the planned and perceived training load of coaches and players may optimize performance and prevent players from under-training or over-training.

It is noteworthy that having only one coach in the study is a limitation and could reduce the interpretability of the data. Therefore, further studies should investigate the discrepancy between futsal players and a group of coaches' perceptions of internal load. Nevertheless, it should be highlighted the relevance of this study considering the longitudinal approach and the team competitive level.

Back to Top | Article Outline

Practical Applications

The coach's RIE and players' RPE differed within the intensity training zones and during the season periods. The RPE scale does not seem to be a suitable tool when used by the coach for intended training load. Meanwhile, it was demonstrated that it could be sensitive to monitoring the daily training loads over a whole futsal season. As an alternative, it is suggested the use of RIE by the coach through controlled anchoring procedures (both exercise and memory) before RPE assessment. Therefore, technical staff should constantly review the training goals in each season period through daily training load control always considering the possibilities and limitations of the RPE method. The coach should consider, especially during high-intensity training loads, the players' response variability (and possible mismatches) according to the demands of the season period.

Back to Top | Article Outline


1. Algroy EA, Hetlelid KJ, Seiler S, Stray Pedersen JI. Quantifying training intensity distribution in a group of Norwegian professional soccer players. Int J Sports Physiol Perform 6: 70–81, 2011.
2. Barroso R, Salgueiro DF, do Carmo EC, Nakamura FY. Training volume and repetition distance affect session rating of perceived exertion and internal load in swimmers. Int J Sports Physiol Perform 10: 848–852, 2015.
3. Blanchfield AW, Hardy J, De Morree HM, Staiano W, Marcora SM. Talking yourself out of exhaustion: The effects of self-talk on endurance performance. Med Sci Sports Exerc 46: 998–1007, 2014.
4. Borg G, Hassmen P, Lagerstrom M. Perceived exertion related to heart rate and blood lactate during arm and leg exercise. Eur J Appl Physiol Occup Physiol 56: 679–685, 1987.
5. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc 14: 377–381, 1982.
6. Brink MS, Frencken WG, Jordet G, Lemmink KA. Coaches' and players' perceptions of training dose: Not a perfect match. Int J Sports Physiol Perform 9: 497–502, 2014.
7. Brink MS, Visscher C, Arends S, Zwerver J, Post WJ, Lemmink KA. Monitoring stress and recovery: New insights for the prevention of injuries and illnesses in elite youth soccer players. Br J Sports Med 44: 809–815, 2010.
8. Buchheit M, Racinais S, Bilsborough JC, Bourdon PC, Voss SC, Hocking J, Cordy J, Mendez-Villanueva A, Coutts AJ. Monitoring fitness, fatigue and running performance during a pre-season training camp in elite football players. J Sci Med Sport 16: 550–555, 2013.
9. de Morree HM, Marcora SM. The face of effort: Frowning muscle activity reflects effort during a physical task. Biol Psychol 85: 377–382, 2010.
10. Eston R. Use of ratings of perceived exertion in sports. Int J Sports Physiol Perform 7: 175–182, 2012.
11. Foster C. Monitoring training in athletes with reference to overtraining syndrome. Med Sci Sports Exerc 30: 1164–1168, 1998.
12. Foster C, Florhaug JA, Franklin J, Gottschall L, Hrovatin LA, Parker S, Doleshal P, Dodge C. A new approach to monitoring exercise training. J Strength Cond Res 15: 109–115, 2001.
13. Fowler P, Duffield R, Vaile J. Effects of simulated domestic and international air travel on sleep, performance, and recovery for team sports. Scand J Med Sci Sports, 2014.
14. Fullagar HH, Skorski S, Duffield R, Hammes D, Coutts AJ, Meyer T. Sleep and athletic performance: The effects of sleep loss on exercise performance, and physiological and cognitive responses to exercise. Sports Med 45: 161–186, 2015.
15. Haddad M, Padulo J, Chamari K. The usefulness of session rating of perceived exertion for monitoring training load despite several influences on perceived exertion. Int J Sports Physiol Perform 9: 882–883, 2014.
16. Hopkins WA. Spreadsheet to compare means of two groups. Available at: Accessed January 21, 2015.
17. Hopkins W. Statistical Analysis and Data Interpretation: An Introduction. Available at: Accessed January 21, 2015.
18. Hopkins WG, Marshall SW, Batterham AM, Hanin J. Progressive statistics for studies in sports med and exercise science. Med Sci Sports Exerc 41: 3–13, 2009.
19. Impellizzeri FM, Rampinini E, Coutts AJ, Sassi A, Marcora SM. Use of RPE-based training load in soccer. Med Sci Sports Exerc 36: 1042–1047, 2004.
20. Jeong TS, Reilly T, Morton J, Bae SW, Drust B. Quantification of the physiological loading of one week of “pre-season” and one week of “in-season” training in professional soccer players. J Sports Sci 29: 1161–1166, 2011.
21. Kelly VG, Coutts AJ. Planning and monitoring training loads during the competition phase in team sports. Strength Conditioning J 29: 32–37, 2007.
22. Kiely J. Periodization paradigms in the 21st century: Evidence-led or tradition-driven? Int J Sports Physiol Perform 7: 242–250, 2012.
23. Marcora SM, Staiano W, Manning V. Mental fatigue impairs physical performance in humans. J Applied Physiol (1985) 106: 857–864, 2009.
24. Meeusen R, Duclos M, Foster C, Fry A, Gleeson M, Nieman D, Raglin J, Rietjens G, Steinacker J, Urhausen A. European college of sport S, and American college of sports m. Prevention, diagnosis, and treatment of the overtraining syndrome: Joint consensus statement of the European college of sport science and the American college of sports med. Med Sci Sports Exerc 45: 186–205, 2013.
25. Milanez VF, Pedro RE, Moreira A, Boullosa DA, Salle-Neto F, Nakamura FY. The role of aerobic fitness on session rating of perceived exertion in futsal players. Int J Sports Physiol Perform 6: 358–366, 2011.
26. Milanez VF, Ramos SP, Okuno NM, Boullosa DA, Nakamura FY. Evidence of a non-linear dose-response relationship between training load and stress markers in elite Female futsal players. J Sports Sci Med 13: 22–29, 2014.
27. Milanez VF, Ramos SP, Salle-Neto F, Machado FA, Nakamura FY. Relationship between training load monitoring based on perceived exertion and heart rate in young futsal players. J of Phys Ed and Sport 26: 17–27, 2012.
28. Miloski B, Freitas VH, Bara-Filho M. Monitoring of the internal training load in futsal players over a season. Rev Bras Cineantropom Desempenho Hum 14: 671–679, 2012.
29. Murphy AP, Duffield R, Kellett A, Reid M. Comparison of athlete-coach perceptions of internal and external load markers for elite junior tennis training. Int J Sports Physiol Perform 9: 751–756, 2014.
30. Nakamura FY, Moreira A, Aoki MS. Training load monitoring: Is the session rating of perceived exertion a reliable method? R da Educação Física/UEM (Impresso) 21: 11, 2010.
31. Nogueira FCA, Nogueira RA, Coimbra DR, Miloski B, Freitas VH, Bara-Filho M. Internal training load: Perception of volleyball coaches and athletes. Rev Bras Cineantropom Desempenho Hum 16: 638–647, 2014.
32. Rodriguez-Marroyo JA, Medina J, Garcia-Lopez J, Garcia-Tormo JV, Foster C. Correspondence between training load executed by volleyball players and the one observed by coaches. J Strength Cond Res 28: 1588–1594, 2014.
33. Stewart AM, Hopkins WG. Swimmers' compliance with training prescription. Med Sci Sports Exerc 29: 1389–1392, 1997.
34. Storen O, Bratland-Sanda S, Haave M, Helgerud J. Improved VO2max and time trial performance with more high aerobic intensity interval training and reduced training volume: A case study on an elite national cyclist. J Strength Cond Res 26: 2705–2711, 2012.
35. Tabachnick BG, Fidell LS. Using Multivariate Statistics. Boston. MA: Allyn & Bacon, 2007.
36. Wallace LK, Slattery KM, Coutts AJ. The ecological validity and application of the session-RPE method for quantifying training loads in swimming. J Strength Cond Res 23: 33–38, 2009.

coach; season; team sport; intensity training; performance; rating of perceived exertion

Copyright © 2016 by the National Strength & Conditioning Association.