Teaching Skills Assessment in Initial Teacher Training in Physical Education

: This paper assessed the impact of two teacher training degrees in Physical Education, one focusing on primary school teaching and the other one concerning secondary school teaching. Data from students of both degrees were statistically compared and contrasted taking into account four training scopes from the Questionnaire of Perception of Teaching Skills in Physical Education . A three-stage analysis was carried out: (i) a descriptive stage, (ii) an invariance analysis, and (iii) a comparison of latent means across the degrees. The results show signiﬁcant statistical di ﬀ erences between the two degrees regarding the training impact on developing skills for physical and motor development basics, contents for learning, and physical activity promoting health; however, no statistically signiﬁcant di ﬀ erences were found regarding the learning and teaching processes in Physical Education. Moreover, in both degrees, the training proved to be insu ﬃ cient for developing the skill in inclusive education.


Introduction
After the creation of the European Higher Education Area (EHEA), new teacher training programs related to skill development were designed. In this regard, Spain created two new university degrees for the initial teacher training in Physical Education (ITTPE) in order to be more consistent with the existing European ones one was a Primary Education Teacher Degree (6-12 years old students), with a special focus on Physical Education (PETPE); the other degree was in Physical Activity and Sports Sciences (PASS) for 12-16-year-old students [1][2][3]. The first one prioritized training contents with a play-based approach, whereas the other was rooted in sport-oriented content. According to Palacios-Picos et al. [4], the first studies on ITTPE took a technocratic perspective derived from the competency-based teacher education model, which, in turn, prioritized skill development [5][6][7][8][9]. Consequently, it was deemed necessary to assess (i) to what extent the new teacher training programs were having an impact on future teachers' teaching skills and (ii) the differences between both degrees. This study is justified on the basis that its results and conclusions may contribute to the enhancement of similar programs, ones within the European context.

Training Scopes in Initial Teacher Training in Physical Education
According to Goodyear et al. and Zhang et al. [10,11], scopes of content included in ITTPE programs arise from the selection of subjects having conceptual and practical pedagogical applications,

Study Objectives
The general purpose of this study was to assess two teacher training degrees in Physical Education, one focusing on primary school teaching and the other one concerning secondary school teaching. From this, two specific objectives were established. The first one was to statistically compare and contrast data from students of both degrees in four training scopes: (i) the skill for managing teaching-learning processes; (ii) the skill for mastering the basics of physical and psychomotor development; (iii) the skill for managing physical activities and learning contents; and (iv) the skill for promoting body hygiene and healthy habits. The second specific objective was to find out any differences in the results of the two degree programs.

Participants
The sample was composed of 1149 university students (59.7% men and 40.29% women) from 20 university centers from 20 Spanish cities. Out of them, 780 (67.88%) were attending the fourth year of PETPE, whereas 369 (32.12%) were attending PASS.

Procedure
For the purposes of this study, and following the guidelines in the Spanish White Books, the Red Nacional de Evaluación Formativa en Educación Superior (Spanish National Network of Formative Assessment in Higher Education) designed a teaching skills questionnaire to be implemented in both PETPE and PASS degrees [1,2]. The first draft of the questionnaire was reviewed and analyzed by a panel of experts, who grouped, removed, and changed items in terms of content validity and reliability. Afterward, and after including all the suggestions, the second draft of the questionnaire was tested with a random sample of 85 students attending the last year of both degrees in four different universities. The statistical process followed to ensure validity and reliability is described below (see Section 2.3).
The final version of the Questionnaire of Perception of Teaching Skills in Physical Education [4] consisted of 22 items assigned to four subscales. The first subscale consisted of eight items on the organization, design, and planning. The second subscale consisted of eight items on the learning contents inherent to PE and organized physical activities. The third subscale referred to the methodology and included ten items. And lastly, the fourth subscale was made up of three items relating to the assessment. All the items referred to the students' perception of the level of support received by the subjects taken in both degrees and their contribution to the development of the skills acquired.
Furthermore, the ethical rigor was guaranteed as established by the Ethics Committee of the researchers' university. In addition, an informed consent form was signed by all participants who, in turn, were not only informed, verbally and in writing, about the purposes and the procedure of the study but also were ensured anonymity and that the data would exclusively be used for the purposes of the investigation.

Instruments
The Questionnaire of Perception of Teaching Skills in Physical Education [4], made up of four subscales with a total of 22 items, was developed and used. The questionnaire also reflected sociodemographic features: age, gender, location, and studies pursued. The respondents expressed their level of agreement for each of the statements on a Likert scale of five points (0 = Not at all and 4 = Extremely).
The process of reaching validity and reliability was as follows. The internal consistency of the four subscales and of the total score was verified using Cronbach s alpha coefficients for ordinal data and McDonald s omega. Overall, an ordinal alpha value of 0.95 and an omega value of 0.91 were obtained. Both values were considered excellent. Regarding the first subscale, relative to organization, design, and planning, good internal consistency indices were obtained (α = 0.87, ω = 0.87). Regarding the second subscale, relative to the contents, adequate reliability indices were obtained as well (α = 0.84, ω = 0.86). Similarly, good internal consistency indices were shown for both the third subscale, relative to the methodology (α = 0.86, ω = 0.86), and the fourth scale, related to evaluation (α = 0.76, ω = 0.75). Finally, the factorial structure proposed for the questionnaire was ratified, and the fit indices were considered adequate (RMSEA = 0.052, 95% CI [0.048-0.056], CFI = 0.975, p (RMSEA < 0.05 = 0.239)).

Data Analysis
The data analysis process took place in three stages. Firstly, the descriptive analysis of the data set considered whether it was appropriate to compare subgroups taking into account the results of this tool and the measurement invariance based on the degree. Secondly, this was tested in four increasingly restrictive models for the first-and second-level factors: configural invariance (equivalence of the number of factors); metric or weak invariance (equivalence of factor loadings); scalar or strong invariance (equivalence of thresholds); and strict invariance (equivalence of item residuals). Finally, after verifying the scale's factorial invariance, differences in latent means across groups were analyzed. To this end, they were fixed to zero for the PETPE student group (G1) and freely estimated for the PASS student group (G2).

Results
Firstly, a descriptive analysis of the data set and comparison across both groups was compiled. As for the skills related to the teaching-learning processes in PE, the differences across groups were few. The skill for preparing and implementing PE programmes aimed at promoting the effective inclusion of pupils with special educational needs (SEN) (item 1.2) was highlighted, although both groups stated that the subjects taken were the least helpful in developing such skill (Not at all or Slightly 34% in G1 (PETPE); Not at all or Slightly 28% in G2 (PASS)). Table 1 shows the percentages for the answer options selected by both groups. All the differences between groups-for the different answer options-were lower than 5% but for three. Particularly, the differences are 5% higher for item 1.2 (G1 = 37.37%, G2 = 44.35%, diff. = 7.08%), for item 1.14 (G1 = 35.96%, G2 = 29.75%, diff. = 6.21%), and for item 1.17 (G1 = 39.5%, G2 = 34.44%, diff. = 5.06%), all of them showing the neutral answer option Moderately. After checking the homogeneity of the variances using the Levene test, the differences between groups were analyzed using the Mann-Whitney-Wilcoxon tests. The results indicate that none of the differences were statistically significant. Table 1. Percentages for the answer options selected for Factor 1 by Groups 1 and 2. In addition to the skill for the inclusion of students with special needs, a relevant percentage of participants from both degrees considered that the subjects taken contributed little or nothing to developing their skill for responding to diversity in PE practices (item 1.16) (Not at all or Slightly 21.78% for G1, 17.08% for G2). Table 2 shows percentages for the answer options selected by both groups for the items related to Factor 2-the skills on physical and motor development basics. As shown, significant differences for all items were identified but for item 1.4 (Knowing and promoting the different motor expressions as part of your traditional culture).  The Mann-Whitney-Wilcoxon tests indicated that PETPE students perceived that the subjects taken had contributed to developing the competency to know psychomotor development (item 1.5) (Mdn = 3) to a greater extent than PASS students (Mdn = 2) (Z = 5.26, p < 0.001, r = 0.16). In the same way, PETPE students also considered that the subjects taken had contributed to developing the competency to know and understand the bodily and motor evolutionary processes (item 1.18) to a significantly higher degree than PASS students (Z = 3.59, p <0.001, r = 0.11). On the contrary, PASS students considered that the subjects taken had contributed to developing the competencies to know the physical capacities and the factors that determine their evolution, to know how to apply their specific technical foundations (item 1.6) (Md = 3), and to know the biological and physiological foundations of the human body in relation to physical activity (item 1.7) (Md = 3) to a significantly higher degree than PETPE students (item 1.6, Md = 2, Z = 2.71, p < 0.001, r = 0.08) (item 1.7, Md = 2, Z = 7.29, p < 0.001, r = 0.21). There were no significant differences between the two groups with regard to the competency to know and promote the different motor manifestations that are part of one's traditional culture (item 1.4, Z = 1.11, p = 0.263, r = 0.03).
As for Factor 3-on the skills related to the contents of learning and physical activities-most of the students from both degrees considered that the subjects taken in their respective degrees helped them a lot or extremely to develop skills aimed at knowing how to use play as a didactic resource and learning content (item 1.12) and to understand the basics of school sport initiation and to design specific tasks to use in the field of education (item 1.9).
The greatest differences between the groups in Factor 3 were identified in item 1.8 (understanding the elements and basics of body expression and non-verbal communication and their training and cultural value). In this regard, for most of the PETPE students (57.29%) the subjects taken helped them a lot or extremely to develop such skill, while just 37% of PASS students experienced that. Table 3 presents this item (1. 8) showing the greatest differences between the groups. To check whether these differences were statistically significant, the Mann-Whitney-Wilcoxon tests were applied again. The results of these tests indicate that the differences between groups are statistically significant at the 95% confidence level in item 1. As for Factor 4-on the skills related to body hygiene and healthy habits-the greatest differences between the groups were in identifying and preventing health risks arising from practising inappropriate physical activities (item 1.22). Table 4 shows the different answers provided by both groups. The greatest differences between the groups were in items 1.21 and 1.22. The differences in skills for having strategies for implementing health elements on hygiene and nutrition for the educational practice (item 1.10) and having teaching strategies promoting the acquisition of regular physical activity habits (item 1.11) were lower than 6%, and they only took place in two categories. The results of the Mann-Whitney-Wilcoxon tests indicate that the differences between groups are statistically significant at the 95% confidence level in item 1. 10  These results show that PASS students considered that the subjects taken for their degree had been more helpful for developing skills for body hygiene and healthy habits than the subjects taken by PETPE students. However, in regard to having strategies for implementing health elements on hygiene and nutrition for the educational practice (item 1.10), PETPE students considered that the subjects taken had contributed slightly more than the perception provided by PASS students.

Factorial Invariance between Degrees
In order to assess whether the second-order model suggested by Palacios-Picos et al. [4] was kept invariant between degrees, a series of nested models were presented. Firstly, there was a configural invariance model where factorial loadings, intercepts, regression coefficients (second-order factorial loadings), intercepts of first-order factors, and second-order factor means were freely estimated between the groups. Four nested models were then sequentially specified where the factorial loadings, thresholds, means, and error terms were established across the groups. As per recommendations of Chen and Cheung et al. [60,61], increments lower than 0.010 in CFI and TLI and decreases lower than 0.015 in RMSEA suggest there are no relevant changes in fitting the model with respect to the more restrictive following one or in establishing the relevance of the fitting differences between the models.
Indexes showed an adequate fit to the configural invariance model (χ2 (410) = 1046.799, p < 0.001; CFI = 0.975; RMSEA = 0.052, 95% CI [0.048-0.056], p (RMSEA < 0.05 = 0.239)), supporting the assumption that these items account for the same underlying construct for both groups. It can therefore be concluded that both the number of factors and the factorial loadings pattern of the items are similar for both groups. Table 5 shows the information on the goodness-of-fit for each model and the comparison between models.
In order to check the invariance second-order model (metric invariance), the previous model was compared against the factorial loadings invariance model. In this case, non-significant decreases in three of the four fit indices were found (∆CFI = −0.009; ∆TLI = 0.005; ∆RMSEA = 0.007; ∆SRMR = 0). Such evidence-the metric invariance-showed that the contribution of each item to the latent variables remained steady for the different groups, and therefore it allowed to compare the groups with respect to the variances and covariances of the latent variances, being appropriate to proceed with a scalar invariance test. To that end, the factorial loadings invariance model (Model 6) was compared against the threshold invariance model (Model 5). The scalar invariance allowed comparisons between group means and, as showed in Table 5, the strong invariance model (Model 6) did not significantly worsen the immediately prior less restrictive model fit (Model 5). Finally, the strict invariance model (Model 7) compared the null hypothesis that the error variances of each item were equal for both groups. In this case, the differences across the groups in the variances considered were attributable to the differences in latent common factors. Therefore, such differences between groups for the evident variables were the result of differences in latent factors. Again, as shown in Table 5, this model did not significantly worsen the fit compared against the less restrictive previous model (Model 6).

Latent Means Comparison
Comparing the measurement tool's strict invariance for both groups allowed comparing the differences in the latent means across these groups. In order to verify the latent means differences for the four first-order factors, the first-order factors' saturations (except for the ones established in 1.0) were equalized for both groups so that the latent means for the first-order factors were set to zero for G1 and were estimated for G2. In addition, the second-order factor's saturations were estimated for both groups. As shown in Table 6, no significant differences were found regarding Factor 1 (Teaching-learning processes in PE) (z = 0.349; p = 0.727; d = 0.071). Note. CFI = comparative fit index; df = degrees of freedom; RMSEA = root mean squared error of approximation; SRMR = standardized root mean residual; TLI = Tucker-Lewis index; CI = confidence interval.
As for Factor 2 (Physical and motor development basics), there were no significant differences either (z = −0.759; p = 0.448; d = −0.107). In this case, PETPE students considered the subjects had helped them to develop such skills slightly greater than PASS students. However, those differences were not statistically relevant.
As for Factor 3 (Learning contents and physical activities), PETPE students considered the subjects taken helped them to develop this type of skills greater than PASS students (z = −4.734; p < 0.001; d = 1.92; d = 1.92). Likewise, significant differences were identified in the latent mean related to Factor 4 (Body hygiene and healthy habits). In this case, the rating for PASS students was higher than the PETPE students' one (z = 2.125; p = 0.034; d = 0.26). Again, such a difference reflected the results achieved in the data descriptive analysis.
As for the second-order factor (Teaching skills specific to PE teachers), the results of the latent means comparisons did not show significant differences between the students of both groups (z = −0.839; p = 0.401; d = 0.056).

Discussion
With the creation of the EHEA, teaching and transversal competencies programs (ITTPEs) have been developed in Spain. This implies a change in the design of the purposes of these studies that obviously affects the rest of the programming elements. In the same way, a greater differentiation is created between the profile and training that primary education teachers receive through the PETPE degree and secondary education teaching through the PASS degree. In view of this, it was deemed necessary to analyze the assessment made by university students in Physical Education teaching, regarding the impact of their respective degrees on the development of their teaching skills in four main content scopes: (i) manage the teaching-learning processes; (ii) support physical and psychomotor development; (iii) manage physical activities and learning content; and (iv) promote body hygiene and healthy habits.
There was evidence that the four factors analyzed were invariant for both groups. The invariance level achieved allowed for significant comparison of the estimated latent means between both groups, since the differences in the variables observed were the result of differences in common latent factors, instead of differences in the understanding of items, the scale structure, or other external factors. There were no statistically significant differences between the groups observed regarding the impact of the subjects structured in Factor 1-training contributing to the development of skills in managing the teaching-learning processes in PE. Both groups considered that the subjects taken helped them a lot in developing skills related to the design and implementation of didactic interventions and assessment, which are vital aspects for managing such contents [15][16][17], and similarly with students from both degrees [6,22]. However, there was a training shortcoming observed in developing skills for addressing inclusion and diversity by PE, which both groups valued as slightly or not at all trained. Such shortcoming was also assessed in the studies carried out by Hutzler et al., McGrath et al., and Qi et al. [23][24][25] and, as such, needs to be better addressed in ITTPE through inclusive training experiences [26][27][28]. Regarding Factor 2, there were significant differences between the groups toward the PASS degree, since it addressed the need for mastering and managing the evolutive and psychomotor development processes of pupils [1]. In addition, Whittle et al. and Gambau-I Pinasa [32,33] acknowledged such training assurance in ITTPE. For PASS students, unlike for the PETPE ones, the mastering of skills related to the understanding of human body biological and physiological basics related to the physical activity and the understanding of the physical capacities was evident; the need for that knowledge and the focus on it is presumed adequate in order to develop them in teaching secondary education effectively-the teaching context for PASS graduates [2, 19,35,62]. As for Factor 3-learning contents and physical activities-both groups considered a lot the training support received for learning how to use play as a didactic resource and teaching content, as well as for understanding the basics of school sport initiation. There were some slightly higher valuations in PETPE because of its exclusive pedagogic approach to the profile and the focus on play and initiation sport as pedagogic resources, using them to address technical and tactical aspects for the sports instruction [37,38,40]. Nevertheless, according to some scholars, sport must be also considered in ITTPE from a socio-critical and equitable perspective [45] in order to compensate for the prevalent competitive sport model reigning in sports federations and fostered on social media [43,46,63]. The training shortcoming for this scope can be seen, as well, in the limited impact of the subjects taken in PASS for developing the skill for body expression, non-verbal communication, and its cultural training value, which is unable to offset the PE traditional hygienist and biological load [48][49][50] with its aesthetical, expressive, and emotional nature [47,[51][52][53].
Finally, there seems to be a greater impact of subjects from Factor 4 relating to body hygiene and healthy habits on PASS students, when compared to PETPE students' perspectives regarding their training received for developing the skills to analyze and communicate, in a critical and substantiated way, the value of the physical activity and sport, as well as its opportunities to support the development and wellbeing and to identify and prevent health risks from inappropriate physical activities. Gutiérrez et al. [35] agreed that the training for theoretical-practical knowledge in PASS for this type of skill targeting health, risks prevention, and quality of life is sufficient [1]. However, there is an ongoing, and perhaps increasing, need for better training for future PE teachers [57] due to the social relevance these aspects have not only in preventing the use of alcohol, tobacco, and other drugs among teenagers but also for creating comprehensive educational atmospheres where physical activity habits and healthy lifestyles are encouraged [58,59].

Conclusions
The evaluation of the initial teacher training programs in Physical Education allows one to explore the real impact of the training given to future Physical Education teachers and, consequently, address current socio-educational challenges. Among the contributions made in the present study, it is possible to recognize the value of the two degrees that train physical educators in Spain, considering the two particular contexts of primary school education and secondary education.
The mastery of certain competencies necessary to face present challenges in both educational contexts, such as those related to educational inclusion and attention to students with special needs, as well as the development of more equitable and socially-conscious perspectives, need to be developed further if one wishes to foster teacher profiles committed to sustainable, responsible, and pro-active social change. This key challenge requires motivating future teachers at both levels to discover and to create diverse, multicultural, functional, and tolerant educational environments.
In the same way, it is clear the necessity to recognize the importance of non-hegemonic bodily practices, such as bodily expression, to reinforce values such as self-knowledge, cooperation, and a greater awareness of non-competitive practices as resources for the formation of a school model that is tolerant, diverse, and alternative. Finally, it is necessary to strengthen the training of prospective PE teachers focused on healthy habits of physical activity and sustainable and balanced lifestyles.