Exploring the Games’ Intangible Legacy on Individuals: A Longitudinal Study of Teacher’s Community

Abstract

Hosting the Olympics is subject to socio-educational outcomes, which can represent intangible and peripheral assets for host communities. The current study explores the Games’ intangible legacy on teachers’ attitudes at different points in time. Data were collected among teachers who attended the Rio 2016 Education Program at three different stages: 2016 (n = 611), 2017 (n = 451), and 2020 (n = 286). A longitudinal trend study was designed using multivariate analysis of variance MANOVA tests and latent growth modelling. Results show that the teachers’ perceptions of Olympic knowledge had a significant growth rate, while skills development and network/social exchange do not show significant changes over the time periods. Longitudinal findings suggest the continuity of the Olympic education programs as the basis for strengthening the Olympic intellect and social capital formation.

1. Introduction

The Olympic and Paralympic Games teach values, promote ideals, and contribute toward humanity development (IOC 2016). The Olympics bidding cities in recent decades have included education initiatives within their tangible and intangible aims (e.g., “Generation 2024” in Paris or “Becoming a Man” in Chicago; Monnin 2020). The education programs and their pedagogical tools tend to appear in the years leading up to hosting the Olympic Games (Teetzel 2012) and are mainly promoted by the local organizing committees. For example, several years before the London Olympics, the LOC offered a plethora of activities, classroom ideas, and resources for teachers through the “Get Set” education program. Similarly, the Rio 2016 organizing committee provided key resources such as an education toolkit, online courses, and a digital platform for teacher training through the “Transforma” program. These programs reinforce the role of Olympism as an educational concept that seeks to value good examples, social responsibility, and respect for universal ethical principles (IOC 2015). Blending sport with culture and education, Olympism draws from divergent approaches that are altered according to the pedagogical orientations and cultural contexts of each host country (Binder and Naul 2017).

The current approach to the Olympics legacy includes education as a way to promote the social capital created by cultural practices and social exchanges (Prüschenk and Kurscheidt 2020) such as education programs. Every host city commits to carrying out Olympic education programs as part of their responsibilities for organizing the Games (Binder and Naul 2017) and as a long-term social development process (Maya 2017), with three assumptions underlying these programs: First, Olympism should be used as an educational tool to influence behaviors based on cognitive (intellectual), affective (social/emotional), and kinesthetic (physical) learning (IOC 2016). Second, these educational interventions should disseminate a values-based curriculum (excellence, respect, and friendship) and ideals (fairness, equality, and ethical behavior) to shape the younger generation in their adulthood (Teetzel 2012). Third, they should promote leadership, engagement, and social inclusion through innovative approaches that foster social behavior change (Kirakosyan 2020). To apply these aspects in Olympic education programs is a long-term educational challenge (Georgiadis 2010) in which teachers play a key role. According to Rio’s official strategy, a strong plan was provided “to motivate and engage teachers and to promote the adoption of the programs activities by schools” (ROCOG 2012, p. 4). This stakeholder group received educational materials (e.g., an Olympic toolkit), participated in Olympic courses (e.g., OVER and PSD), and was encouraged to share experiences in a network/social exchange (e.g., Educopedia). By engaging with the education program, the teachers are involved in propagating the Olympic ideology, and as result, unintentional or intangible benefits have emerged.

Social value creation can be substantiated by a range of positive outcomes apportioned between different actors within the Olympics program. These outcomes can be evidenced in intangible and peripheral forms, such as enhancing Olympic knowledge among teachers (Ribeiro et al. 2020), prosocial behavior among students (Sukys et al. 2017), or by other intangible benefits such as gaining experience and personal talents within the host community (Leopkey and Parent 2012). These social effects are heightened by a strategic alliance between the Olympic Games and the school community. For instance, the Sochi Olympic Winter Games in 2014 embraced the importance of educational legacy and created the Russian International Olympic University (IOC 2016). PyeongChang, the host of the 2018 Winter Olympic Games, has continued to develop the “Dream Program” since 2004, as a part of its educational leverage strategy. More recently, in Tokyo, the Tsukuba University, in its teaching policy for Olympic education (e.g., Master’s in Sport and Olympic Studies), offers a significant contribution to creating social capital and lifelong learning experiences (Hwang 2017).

Scholars and practitioners have been debating the scope, social effects, and value creation of Olympic education programs. For Schnitzer et al. (2018), the perceptions about Olympic values depend mainly on the social capital created (norms, institutional trust) provided by the Youth Olympic Games (YOG). Prüschenk and Kurscheidt (2020) formed detailed reasoning on the formation of social capital on the basis of the Olympic values among spectators, and Ribeiro et al. (2020) demonstrated that when Olympism is embedded in pedagogical practices, social capital formation occurs, leading to a positive education impact among teachers. As defined by Coleman (1988), social capital is created by the quality of the interactions between different social networks. As a result, participating in Olympic education programs may be seen as a form of “social interaction” leading to developing skills, networks, and human capital (Kay and Bradbury 2009). This is apparent in the Olympics context, given the multiplicity of actors involved (Leopkey and Parent 2017) and the engagement opportunities these events represent (e.g., via education initiatives).

However, an analysis of existing evidence and social capital theory suggests that delivering a positive legacy to host communities may be difficult to prove. Griffiths and Armour (2013) were skeptical about the Olympic legacy aspirations and suggested adopting a more critical view of their contribution to developing social capital. In the same vein, Mackintosh et al. (2015) noted that the Olympics legacy may remain untested, and they highlighted the need for considering longitudinal studies for its assessment. Despite the intuitive appeal of the link between the Olympic Games and social capital, it is important to consider that there is no empirical evidence to support its long-term outcome (Ribeiro et al. 2020). In this respect, the social capital theory is a useful lens in the Olympic Education context for understanding the Games’ social value for individuals (Coleman 1990), whilst adding palpability to the intangible legacy concept and demonstrating its growing significance to future bidders. By nature, this theory offers a relevant theoretical paradigm for understanding how Olympic Games can be used to build intangible benefits through a community network/social exchange and learning.

Nevertheless, previous attempts to assess the Games’ intangible legacy on individuals, in particular sport and school settings, have some important limitations. First, the assessment models of Games’ intangible legacies fall short of capturing the true nature of the legacy value (Preuss 2019), initiating their legacy framework by building on a multifaceted concept (Rocha 2020; Agha et al. 2012). This assumption can have multiple issues; specifically, if too many variables are analyzed together, it can be difficult to gain deep insights into the real advantages of hosting the Olympics. Thus, a micro-level analysis (e.g., education legacies such as Olympic programs and teachers) can advance the knowledge of value creation within the Olympic Education field. Second, neither the scientific literature nor the International Olympic Committee offer a comprehensive and reliable scale to measure teachers’ attitudes about Olympic education programs after the Games. Although recent studies have started analyzing this approach (e.g., Nordhagen and Fauske 2018; Sukys et al. 2017), their attention has essentially been placed on the educational program’s impact on children and young people, and, therefore, there is a lack of understanding of how intangible assets for teachers should be measured. While the targeted beneficiary group of the program is children and young people, teachers also play an important role as a change agent creating opportunities in teaching the sports (Ribeiro et al. 2020). They help with the educational program delivery even after the Games have ended, relaying their opportunities, political changes, and commercial agendas over time (Kohe and Collison 2019). Thus, a longitudinal analysis is also important to advance the theory and practice of Olympic Education (Ribeiro et al. 2020) because variations in teacher attitudes provide new insights for future Olympic bidders.

Given the need for more research focusing on intangible legacies (Preuss 2019), education program development (Sukys et al. 2017), and value-in-context for teachers (Ribeiro et al. 2020), this study’s purpose is to explore the Rio 2016 Olympic Games’ intangible legacy on teachers’ attitudes from 2016 to 2020. It provides empirical work that can support future bidders in decision-making, as well as exploring the following questions:

-

How do teacher attitudes vary over the years?

-

How does the Rio Olympics contribute to the social capital development of teachers?

2. Methods

2.1. Research Design

A case study approach was selected, and a longitudinal design was conducted with three post-event measurements, to explore the Games’ intangible legacy on teachers who attended the Rio 2016 Education Program, the Transforma Program. In this study, the intangible legacies examined refer to the propagation of Olympic ideology for the development of skills, knowledge, and experiences of teachers.

2.2. Participants and Procedures

Longitudinal surveys were conducted among the teachers who attended the Transforma program. The sampling strategy used an organizational database provided by the Rio Organizing Committee with a total of 1940 teachers registered in the Education Department for the Games. The following teacher selection criteria were considered: (i) Teachers who attended the Transforma program between 2013 and 2016, (ii) teachers who participated in at least 2 Olympic Education activities in the Transforma program, and (iii) teachers whose school was included in the list of schools participating in the Transforma program.

Data were collected using an online questionnaire at 3 different moments: A month after the Games (15 September to 15 October 2016), a year after the Games (15 September to 15 October 2017), and four years after the Games (15 September to 15 October 2020). All participants voluntarily participated and agreed via an informed consent form. Data collection occurred during September/October of each period and the same online survey was conducted at each time point. A banner with the questionnaire link and a description of the purpose of the study was activated using survey administration software (Google Forms), inviting teachers who attended the Transforma program to participate in the study. At each point of data collection, the IP addresses of participants were recorded, and further access from these IP addresses was denied after survey submission to avoid repeat participants. The e-mail address of the participants was requested (not mandatory) to align responses from both data collection waves.

The online surveys were identical for all 3 waves (T1, T2, and T3) with a total of 611 individuals filling out the questionnaire at T1, 451 responding to the questionnaire at T2 (73.8% response rate compared to T1), and 286 filling out the T3 survey (46.8% response rate related to T1). The characteristics of each sample are described in

Table 1. Characteristics of samples.

Sample Size	2016	2017	2020
	611	451	286
Gender
Male	225	181	129
Female	286	270	157
Age
18–29	99	78	30
30–39	225	177	102
40–49	187	141	94
50–59	96	51	54
60 or more	4	4	6
M(SD)	39.3 (9.47)	37.8 (7.53)	41.2 (8.89)
Education
High School	19	18	3
Higher Education	592	433	283
(Bachelor’s)	(554)	(388)	(131)
(Master’s)	(36)	(43)	(148)
(Doctorate)	(2)	(2)	(4)
Residence
North	33	27	12
Northeast	58	41	25
Midwest	40	29	16
Southeast	428	303	196
South	52	49	37
Nationality
Brazilian	591	447	283
Other	20	4	3

M—Mean age; SD—standard deviation.

To obtain the same sampling, new procedures were carried out given that this attribute is the most appropriate for a longitudinal design (Ribeiro and Almeida 2020). The data collected in each moment was then crossed by identifying the e-mail addresses provided by participants. Only respondents who participated in T1, T2, and T3 waves were included in the final database. Subsequently, the data were examined, and the surveys not completely filled out and those containing 10 or more consecutive answers to the same scale number were excluded from further analysis. After these procedures, a total of 218 surveys were deemed usable for data analysis, considering the same individuals who participated in the T1, T2, and T3 survey waves (n = 218). This sample contained 55.1% female and 44.9% male respondents with an average age of 41.16 years old (SD = 8.87). A total of 22 Brazilian states composed the sample and most of the teachers were of Brazilian nationality (99%). Specifically, 72.7% of the teachers reported having a college degree followed by a graduate degree (25.9%), and only 3 had high school as their highest level of education (1.5%). Most of the respondents demonstrated knowing the Impulsiona program (78.5%) and 93.1% of the respondents would like the Transforma program to remain in their schools.

2.3. Measures

An online questionnaire was constructed assessing a pool of 12 items and containing the following socio-educational measures: Experience networks (EN), Olympic knowledge (OK), and skills development (SD). These measures were proposed by Ribeiro et al. (2020), adapted to the Brazilian population using a scale development process (Clark and Watson 1995), and validated in its study. The scale assessed 3 attributes of the intangible legacy as first-order latent variables. EN concerns the opportunity to establish new contacts, volunteer, and share emotional and professional experiences. OK refers to the opportunity for Olympic sports learning, sharing of information, and Olympic culture-based training, and SD represents the opportunity to strengthen personal skills, practice a new sport, and develop new employment prospects and talent in different fields. Four items were explored under each of these 3 first-order latent variables (dimensions of legacy), as shown in Appendix A. Additionally, the participants in this study also responded to a set of socio-demographic and personal queries in the second section of the survey. The variables assessed were age, gender, residence, education level, nationality, and education program support (see Table 1, which lists the sample characteristics).

A guide question invited the respondents to assess the items according to their level of agreement. The stem for the items read, “Assessing the legacy of the Transforma program after the 2016 Olympic Games, please express your level of agreement with the following statements”. All items were measured based on positive statements formulated in Portuguese based on the original scale (Ribeiro et al. 2020), and they were jumbled within each section. The response format for all items was a 5-point Likert-scale ranging from 1 (Strongly Disagree) to 5 (Strongly Agree). The survey items are reported in Appendix A.

2.4. Data Analysis

First, data were analyzed by descriptive statistics (mean scores and standard deviations). Skewness and kurtosis values were also examined to evaluate if the measures were close enough to the regular normal distribution and could be used in further factorial analyses. Second, a MANOVA (multivariate analysis of variance) was conducted to examine possible differences between education legacy perceptions (EN, OK, SD) in different time periods (T1, T2, and T3). The MANOVA test was employed to test the differences in the vector (centroid) by means of multiple interval dependents for one or more independent categories (Lee et al. 2013). In this analysis, the time-period effect was treated as an independent variable and the 3 legacy dimensions as a set of dependent variables. Additionally, to examine the relationships between socio-demographic variables and intangible legacies, a series of MANOVAs were conducted using IBM SPSS 26.0 (see

Table 2. Descriptive statistics and MANOVA results.

Factors		2016 (n = 611)				2017 (n = 451)				2020 (n = 286)				MANOVA
		M	σ	Sk	Ku	M	σ	Sk	Ku	M	σ	Sk	Ku	F	p	Power
1. Experience networks (EN)		3.74	1.09	−0.79	0.03	3.60	1.13	−0.61	−0.46	3.69	1.04	0.90	0.23	1.890	0.151	0.401
2. Olympic knowledge (OK)		3.49 ª	1.32	−0.59	−0.78	3.38 ª	1.28	−0.47	−0.89	4.07	0.90	−1.47	2.20	30.211	0.000 *	1.00
3. Skills development (SD)		3.78	1.16	−0.89	0.04	3.61	1.01	−0.58	−0.22	3.68	1.17	−0.80	−0.34	3.022	0.052 **	0.586
Variable		EN				OK				SD				F	p	Power
		M	SD	F	p	M	SD	F	p	M	SD	F	p
Gender	M	3.63	1.14	1.54	0.215	3.64	1.19	2.37	0.123	3.67	1.15	0.730	0.391	2.888	0.034	0.692
	F	3.71	1.06			3.53	1.29			3.72	1.09
Age	18–29	3.68	1.027	0.740	0.565	3.38	1.237	2.93	0.020	3.66	1.102	0.420	0.795	1.321	0.004	0.684
	30–39	3.64	1.126			3.58	1.290			3.71	1.162
	40–49	3.74	1.087			3.69	1.215			3.75	1.055
	50–59	3.64	1.117			3.48	1.261			3.65	1.131
	60 or more	3.86	1.205			4.10	0.968			3.66	1.316
Education	High school	3.97	1.01	3.22	0.073	3.59	1.35	0.001	0.988	4.08	0.97	5.256	0.022	2.762	0.041	0.670
	Higher education	3.67	1.10			3.58	1.25			3.69	1.12
Residence	North	3.87	0.97	0.585	0.633	3.71	1.24	2.518	0.040	3.88	1.18	1.798	0.127	2.394	0.004	0.944
	Northeast	3.60	1.07			3.63	1.16			3.70	1.10
	Midwest	3.66	1.12			3.88	1.11			3.90	1.10
	Southeast	3.68	1.11			3.52	1.29			3.71	1.11
	South	3.74	1.08			3.80	1.07			3.49	1.16

* p < 0.001; ** p > 0.05 (marginally significant). ª Difference is not statistically significant at p < 0.05. M—Mean; σ—standard deviation; Sk—skewness; Ku—kurtosis.

). Third, the theoretical framework based on the Games’ intangible legacy was designed with a longitudinal outline to estimate a latent growth curve model. The term growth curve model (GCM) comprises statistical models used for repeated measures of data allowing the estimation of the interindividual and intraindividual variability of change patterns throughout time (Curran et al. 2010). The estimation was carried out by applying confirmatory factor analysis (CFA) through a latent growth curve model (LGM) (Marôco 2014).

In the first phase, a series of CFAs were conducted to evaluate the goodness-of-fit of the measuring model considering the following indices: χ²(df) (Chi-Square Statistic and degrees of freedom), CFI (Confirmatory Fit Index), TLI (Tucker–Lewis Index), and RMSEA (Root Mean Square Error of Approximation). CFAs were performed with IBM SPSS AMOS 26.0. Convergent validity was estimated based on factor loadings in the measurement model (Fornell and Larcker 1981) and Cronbach’s alpha (α) was computed to evaluate the reliability of the survey measures (Hair et al. 2009).

In the second phase, univariate LGM analyses were performed to determine the basic shape of the growth curves for the 3 dimensions of intangible legacy. To establish a final model that most adequately depicted the change trajectory, we fitted nested univariate LGM models to the data for each dimension. This fit procedure started to specify the LGM according to the serial correlation of the errors. Eight relations were conducted and justified by the fact that they were the same items correlated in different time-periods (r < 0.20). The correlations between the errors were limited to those statistically significant according to the modification index analysis (IM > 11), as recommended by Marôco (2014), up to the specification of the final model. The fit indices used were as follows: χ², the absolute index χ²/gL, and RMSEA (Curran et al. 2010). Lastly, the values were estimated for all parameters of the LGM model (means, variances, and correlations).

3. Results

The mean scores of the education legacy variables decreased from 2016 to 2017 and increased from 2017 to 2020. Skewness and kurtosis measures indicated no concerns in terms of symmetry and normal distribution of the data. The MANOVA results (Wilks’ lambda = 0.93, F(6,2686) = 15.556, p < 0.001) showed significant changes in teacher perceptions (see Table 2). The results revealed a positive and significant mean difference for Olympic knowledge [F(2,2686) = 3.21, p = 0.000], a marginally significant difference in skills development [F(1,2686) = 3.02, p = 0.052], and a non-significant mean difference for the experience networks factor [F(2,2686) = 1.90, p = 0.151]. In addition, post hoc Tukey tests indicated that there were no statistically significant differences (p < 0.05) in OK between 2016 and 2017 or in SD between 2016 and 2020 as well as 2017 and 2020. Table 2 shows the MANOVA results considering the cross-sectional samples.

Separate MANOVA analyses revealed that teacher perceptions significantly varied across gender [F(3,1344) = 2.88, p < 0.03], age [F(12,3548) = 1.32, p < 0.00], education level [F(3,1343) = 2.76, p < 0.04], and residence [F(2,3548) = 2.39, p < 0.00]. Univariate analysis for each MANOVA analysis was also undertaken to examine the effect of each independent variable on dimensions of intangible legacy. A significant difference between ages was only observed for Olympic knowledge perceptions [F(4,1343) = 2.93, p < 0.02]. The age groups of younger (18–29 years old) and older (60 or more years old) teachers showed statistically significant results with a significant difference also being observed among teachers in different education levels for skills development [F(1,1345) = 5.25, p < 0.02], evidencing that those who had higher education had greater personal skills development. Furthermore, teachers who lived in the Brazilian States further South demonstrated knowledge perceptions [F (4,1343) = 2.55, p < 0.04] that were significantly different than those who lived in North and Midwest States.

The CFA for each latent growth variable showed good internal consistency and composite reliability estimates. The average variance extracted (AVE) of all variables was equal to or larger than 0.50, which is indicative of convergent validity (Fornell and Larcker 1981). All items and their dimensions have substantive importance for the scale.

After that, to identify the best-fitting baseline LGC model, we conducted several nested model comparisons and assessed whether there was a significant improvement in the relative fit (see

Table 3. Latent growth models: Alternative specifications.

Factor	Model	Change Function	χ2	df	χ2/df	CFI	RMSEA	CI
Intangible legacy	Model 1	Non-growth	7563.12	630	12.00	1.00	0.23	32.62–35.37
	Model 2	Linear growth	1782.69	573	3.11	0.83	0.10	0.096–0.101
	Model 3	Non-linear growth a	1124.32	570	1.97	0.92	0.06	0.063–0.075

^a Retained (most parsimonious model); p < 0.001.

). Three models were compared: A no-growth curve model assuming no change in teacher perceptions (model 1), a linear growth model assuming a linear change (model 2), and a nonlinear growth model assuming a non-linear change (model 3). Model 3 was modelled by fixing the first and third coefficients of the slope factor to constants (0, 1) and freeing the second coefficient of the slope factor. This specification allows for estimating an empirical curvilinear trend that optimally fits the data (Guye et al. 2017). Table 3 shows that model 3 fitted the data significantly better than models 1 and 2, justifying a nonlinear growth trajectory. These findings were also illustrated in the following figure, allowing a better scenario analysis to define the measurement moments and cycles to be used in the LGM. Figure 1 shows the average changes in teacher perceptions from 2016 to 2020 measured through the EN, OK, and SD variables.

Then, the LGM fitting was evaluated, and the values of the parameters estimated by the model were tested. The nonlinear LGM indicated an acceptable fit to the data [χ²(570) = 1124.32 (p < 0.001), χ²/df = 1.97, RMSEA = 0.06, and CFI = 0.92]. Even though the χ² was significant, the ratio of the degrees of freedom was below the 3.0 criterion and χ² is known to be sensitive to the sample size (Hair et al. 2009). In addition, the RMSEA was 0.06, indicating a good fit (Byrne 2000), while CFI was greater than the recommended 0.90 criterion (Hair et al. 2009) with a 90% confidence interval [0.063;0.075].

Table 4. LGM estimated parameters.

Parameters/Means	Estimate	Standard Deviation	p
EN (int)	6.473	0.607	***
EN (slo)	0.633	1.005	0.528
OK (int)	3.352	0.093	***
OK (slo)	0.557	0.117	***
SD (int)	3.406	0.077	***
SD (slo)	0.182	0.111	0.103
Parameters/Variances (σ2)	Estimate	Standard Deviation	p
EN (σ2 int)	1.102	0.161	***
EN (σ2 slo)	2.205	0.292	***
OK (σ2 int)	1.323	0.160	***
OK (σ2 slo)	1.806	0.216	***
SD (σ2 int)	0.747	0.114	***
SD (σ2 slo)	1.219	0.411	0.003
Parameters/Correlations (r)	Estimate	Standard Deviation	p
EN (r int, slo)	−0.629	0.184	***
OK (r int, slo)	−0.728	0.168	***
SD (r int, slo)	−0.688	0.142	***

*** p < 0.001; EN—Experience networks; OK—Olympic knowledge; SD—skills development; int—intercept, mean variable value; slo—growth rate, mean variable growth; σ² (int or slo)—variance of growth rate intercept; r (int, slo)—correlation between intercept and growth rate.

shows the estimated LGM parameters. The analysis of the standardized slope parameters of EN (slo), OK (slo), and SD (slo) demonstrated that only Olympic knowledge (OK) revealed a positive and significant growth rate, while EN and SD did not display significant average growth. The results revealed that the growth rate of OK (slo) was significant with an average and positive growth of 0.557 units from the original scale, indicating that teachers showed an increase in their perception about Olympic knowledge over the years (2016 to 2020). In the other factors, EN and SD, a nonsignificant average growth rate was observed in two variables (EN = 0.633, p = 0.528; SD= 0.182, p = 0.103), meaning that teachers did not have a high perception of improvement in relation to new skills development and personal experiences.

Regarding the variance observed between the initial level and the slopes for the three dimensions, the results showed positive and statistically significant variability (σ² EN = 2.20, p < 0.001; σ² OK = 1.80, p < 0.001; σ² SD = 1.21, p < 0.05). This means that teachers did not have the same perception in relation to personal experience, knowledge, and skills development in the Transforma program, and their perceptions did not increase homogeneously. Moreover, the correlation coefficients between the initial status and the slope for the three dimensions were negative and significant (rEN = −0.629, p < 0.001; rOK = −0.728, p < 0.001; rSD = −0.688, p < 0.001). This means that teachers who showed a low perception of Olympic legacy in the initial period had higher growth than teachers who showed a high perception in the same period. Therefore, teachers who had a more negative perception tended to have a higher growth rate. Table 4 shows the complete results of the parameters estimated for EN, OK, and SD.

4. Discussion

In this research, we explore teacher attitudes over time, considering the educational variables as a part of the Games’ intangible legacy. In doing so, this study highlights the changes in teacher attitudes in different stages of the legacy process. Given that the intangible legacy is an important issue and cannot be easily assessed (Preuss 2019), this study represents an important step by clarifying its social effects and suggesting broader educational aspirations for future bidders.

The findings revealed that teacher knowledge had a significant growth rate during the three time periods of observation (see Figure 1). This means that the teachers increased their awareness regarding Olympic Games as a way to contribute to their human development (Coleman 1988). As noted by Binder (2012) and Griffiths and Armour (2013), the Olympic Games are a reference for creating social capital among teachers and youngsters who attend Olympic education programs. At that point, the social connections experienced via working groups, videos, and e-learning courses may have leveraged the bonding and social capital formation among teachers. In comparison with similar education initiatives, Schnitzer et al. (2018) found an increase in local interest in the Olympic Movement in Singapore, while Grammatikopoulos et al. (2005) noted positive effects in teacher relationships and training as a result of the Olympic education program in Greece. This is also consistent with recent studies in which knowledge about teaching Olympism (Ribeiro et al. 2020), simultaneous experience, and social exchange (Prüschenk and Kurscheidt 2020) contribute to the development of social capital as a particularly suitable platform to leverage the Games’ intangible legacy (Griffiths and Armour 2013). In this sense, one may argue that when the Olympic Games are embedded in educational practices, social capital formation occurs over time, leading to a positive and intangible outcome.

This creation of social capital through Olympic Education (i.e., training and information sharing) is a major outcome of this study. However, this finding needs to be discussed beyond the results obtained. Different from the Vancouver 2010 and London 2012 education programs that continued after the Games with local stakeholder support (Kirakosyan 2020), Transforma ended in 2016 without prospects for future continuity (Rocha 2020). Thus, the first explanation for this finding is that the Impulsiona program appears focused on encouraging the use of sports as an educational tool in Brazilian schools (Impulsiona 2020). Such a proposal essentially enriched and complemented the Transforma approach and content and jointly contributed to the pedagogical development of teachers in Brazil (Kirakosyan 2020). An additional explanation for our findings was the free e-learning courses offered (e.g., Olympic Values for Life), provided by the Brazilian Olympic Committee to foster Olympic Education and strengthen social linkages (Ribeiro et al. 2020). In addition, as noted in Table 1, the teacher education level increased during the time periods under analysis, particularly at the graduate degree level, which contributed to enhancing their knowledge of Olympism.

Longitudinal results of this study also revealed that there were no differences in teacher perceptions of experience networks and skills development between 2016 and 2020 (see Table 2). These findings are corroborated by latent growth curve analysis (LGC) that evidenced a non-significant growth rate for the two variables over the time period (see Figure 1). These variables assumed a nonlinear trend, decreasing a year after the Games (EN: −14 and SD: −17) and slowly increased 4 years after the Games (EN: +9 and SD: +7); however, this growth is not statistically significant. Widespread evidence of a lack of resources and pedagogical guidance (dos Santos 2018) as well as a disbelief in the sustainable legacy (Rocha 2020) may explain the non-significant variation of teacher’s perception post-Games. Previous studies have also identified difficulties in integrating its pedagogic contents into mainstream school education in Brazil due to the governmental changes that occurred in secondary education (Kirakosyan 2020). Another alternative explanation for this finding might be the increase in Brazil’s economic and political problems (Rocha 2020), which contributed to halting the Transforma in 2016, and as a result, teachers may have perceived a decrease in network/social exchange in the following post-Olympic years. However, although there was no significant growth rate in teacher perceptions of new skills and experience, their perception increased from 2017 to 2020, which may provide social capital development in the future. If managed effectively, the OEPs could contribute to individual and collective human development (Culpan 2017) using the skills, knowledge, and experience level as a “barometer” for the Games’ social capital.

Lastly, this study provides some practical implications for helping the IOC, LOCs, and future bidders to promote the continuous management of the Olympic education programs, as well as their sustainability and regeneration. First, findings suggest that teachers had different initial perceptions in relation to the Games’ intangible legacy (see variances in Table 4), and over time, those who showed a lower initial perception tended to have a higher growth rate (see correlations in Table 4). In this respect, it is important that Impulsiona’s board strengthen its support for less-engaged teachers (e.g., inviting them to extra school activities or working groups) and embed their related efforts into broader social experiences and talent-focused skills. For instance, using sport to promote education practices such as leadership meetings, teacher exchange, or scholarships might be useful for awakening new personal talents (refereeing, coaching, or sports entrepreneurship).

Second, the success of the OEP has implications in terms of daily physical education and new talent development in school settings. This program is an opportunity to increase awareness of the Olympic values and a pedagogy orientation to foster Olympic Education through Physical Education classes (Nordhagen and Fauske 2018). In this sense, it is crucial that schools’ coordinators and teachers understand it as a continued endeavor beyond the Olympics-hosting time period. In practice, this means including Olympic ideology in physical education curricula, in school sports programs, and in coaching efforts outside of school in many cultural and educational settings. With regards to new talent development, it is important that teachers create events to promote their skills (e.g., Olympic days, sports courses, workshops, or social projects) in order to leverage their talent in different fields (health, technology, volunteering, project management, etc.). This should involve better use of skills learned by teachers (e.g., leadership, communication, respect, or goal setting) as the basis for future professional networks, employment prospects, and volunteering in their lives.

Third, the evaluation of the intangible Olympic legacy in the post-Games epoch requires a departure from past practices. The role and functions of the stakeholders and their accountability after the event need to be redefined (Kohe and Chatziefstathiou 2017). In this sense, it is recommended that the IOC and organizing committees take sustainable steps to manage Olympic education programs such as signing a Host Cities Educational Contract (HCEC) that confirms the accountabilities of each partner as well as defining which external stakeholders can enter the education field (Kohe and Collison 2019). Furthermore, it is important that the IOC remains committed to managing and subsidizing Olympic education programs, at least in the early years after the Games, thereby supporting their future sustainability. Future bidders should learn from the Transforma program outcomes in order to implement a sustainable legacy process that, in turn, maximizes social capital formation in the host communities.

5. Limitations and Future Research

The first limitation has to do with the sample composition and response rates. The online questionnaire strategy that was used may have limited the sample composition given that there are still some teachers from the host regions that are without e-mail or face difficulty in accessing computers. Second, the fact that we only used teachers limits the generalizability of the findings. Future studies should gather opinions on this matter of other stakeholders to better assess the potential benefits and costs of Olympic education programs. Moreover, a qualitative approach to teachers and other stakeholders (e.g., focus groups and in-depth interviews) may also contribute to better understanding other variables that may not have been captured in the current study. Third, the data in this study are correlational and do not allow a causal test of the relations among the variables in our model. Although we controlled demographic characteristics (age, gender, education, or residence), other characteristics might affect teacher perceptions of education programs such as household income, ease of access, or trust in the IOC.