The data presented here provide preliminary evidence for the construct validity, internal consistency and theoretical alignment of the revised behavioral regulation and need satisfaction scales among primary school-aged children. Bivariate correlations among the motivation subscales supported theoretical tenets of a simplex-like ordering of motivation types on a self-determination continuum . The internal consistency of the three need satisfaction scales and three of the four motivation subscales was acceptable; however the internal consistency of the introjected motivation items was below accepted thresholds . Previous research has reported low internal consistency of scales measuring introjected PA motivation among children  and adolescents . However, other research has found measures of introjected regulation within PE and exercise to be internally consistent .
The measurement of introjected PA regulation among children is challenging as it requires participants to understand and recognise feelings of guilt and shame as a source of motivation. Additionally, participants are required to differentiate introjected motivation from external motivation, the latter of which may be clearer. Introjection may be too abstract for young children, relying on more advanced cognitive / self-perception development. In line with the perspective that children develop more differentiated self-perceptions at approximately 8 years of age  previous studies measuring motivation of younger children (i.e., grades 1–3) have combined introjected and external types into a single controlled motivation indicator . However, in the present study, we rephrased PA introjection items to reflect feeling bad about oneself if not active and showing other people how good I am which are likely understood by 9–11 year olds . Further, introjected and external regulation were only moderately correlated suggesting children could differentiate between these scales. Future work could examine the utility of measuring introjected and external regulation using separate and combined scales among children.
The central aim of the present work was to explore the associations between the motivation types forwarded in SDT and objectively-assessed PA among children. Intrinsic motivation was positively associated with MVPA and shared approximately 4% of its variance with the variance of accelerometer scores. This association is consistent with previous work among adolescents where PA was measured objectively  and by self-report . The magnitude of shared variance between intrinsic motivation and PA is greater than that previously reported between adolescent’s autonomous motivation for exercise and their pedometer scores (1%)  although it still remains small.
Intrinsic motivation was the only type of motivation associated with children’s PA behavior. This suggests that interventions to increase children’s PA should be designed to optimise children’s enjoyment of PA and ensure that they can find inherent satisfaction in being active rather than relying on forms of extrinsic motivation. Designing interventions that achieve this will require substantial formative research and participant engagement to ensure enjoyment is achieved and maintained . Previous research among adolescents has also reported positive associations between identified regulation and PA [9, 16]. The comparative lack of an association in our findings may indicate that internalised extrinsic forms of motivation for PA (e.g., improvements in health) are more central to adolescent’s PA engagement than they are for children.
Lending support to the tenets of SDT that controlling forms of motivation do not underpin meaningful behavioral engagement , both introjected and external regulation were unrelated to children’s PA in the present study. Previous research among adolescents has identified inconsistent cross-sectional associations between introjected regulation and PA [9, 16]. These findings again point to potential developmental differences in PA motivation and highlight the importance of considering children’s and adolescent’s motivation separately. It is important to reiterate that the introjected motivation subscale had low internal consistency and as such, results need to be interpreted with caution.
Within SDT, more autonomous forms of motivation are hypothesised to be underpinned by the satisfaction of needs for autonomy, competence and relatedness . In the present work, a need satisfaction latent factor explained substantial proportions of variance in intrinsic motivation (55%) and identified regulation (44%). The magnitude of these associations is similar to those found between need satisfaction and motivation in PE lessons  and the findings extend previous work in youth PA settings which found that only relatedness and competence needs were weakly associated with an autonomous motivation composite score . Supplementary analysis showed that the association between need satisfaction and intrinsic motivation was primarily carried by perceptions of autonomy need satisfaction. This underpins the importance that children’s social environments (e.g., teachers, parents, coaches) foster perceptions of choice, freewill and volition in order to develop their intrinsic motivation towards PA. While in the bivariate analysis both competence and relatedness were positively and moderately associated with intrinsic motivation, the lack of association between competence and relatedness and intrinsic motivation in the multivariable structural model is not consistent with SDT. A potential explanation is the relatively large factor correlations observed among the need satisfaction subscales. Future work is needed to tease out in more detail how children distinguish between and verbalise the three need satisfaction constructs and to optimise measures to reflect this.
Due to sample size restrictions and our primary focus on exploring the associations between individual motivation types and MVPA, we did not explore associations between individual needs and all behavioral regulations. However, bivariate correlations largely supported SDT, provided further evidence for the validity of the adapted scales and illuminated PA motivational processes among children. Specifically, all three needs displayed moderate-to-strong correlations with identified regulation. Need satisfaction was not correlated with external regulation in either bivariate or SEM analyses. Small-to-moderate correlations were found between all needs and introjected regulation. Such findings are contrary to SDT, but similar associations have been found in the PE motivation literature [25, 33]. In the present study, the need satisfaction-introjected regulation association was much weaker than the association between need satisfaction and both types of autonomous motivation. Further, the shared variance was approximately 10%, lower than for intrinsic motivation and identified regulation.
This unexpected association could reflect the unique nature of need satisfaction and introjection in PA . Specifically, because much of children’s PA is likely to be enacted in social contexts with friends or siblings, children may have interpreted items about feeling bad when not being active as feelings related to disappointing their friends / siblings (e.g., if not able to play out), thereby reflecting a strong connection with active others which was assessed by the relatedness items (i.e., others want me to be active with them). The discordance between theory and empirical findings with regards to need satisfaction and introjected motivation among children and adolescents warrants further investigation.
The findings presented here suggest that psychological need satisfaction is a potential route to autonomous motivation among children. From the SDT perspective, psychological need satisfaction in PA can be manipulated by social agents (teachers, parents, coaches) either adopting autonomy-supportive (need satisfying) or controlling (need thwarting) interpersonal communication strategies . As children’s PA is often facilitated by an adult and such social environmental factors are particularly salient, the psychological needs represent clear targets for interventions seeking to increase children’s autonomous motivation and develop their long-term competence, active friendship groups and enjoyable PA experiences.
Limitations and future directions
While the scales adapted and tested provide theoretically coherent tools with good preliminary evidence of construct validity, scale development is an ongoing process , and there is scope to further test and validate scores derived from the motivation and need satisfaction measures. Particular attention could be given to further test and improve the internal consistency of the introjection items. In addition, in order to reduce participant burden we did not include a measure of amotivation. As such, the motivation measure assumes a presence of motivation, albeit of varying quality (i.e., self-determination) and does not represent children with low quantity motivation (i.e., amotivated). Future work could develop / adapt amotivation items and test a more complete motivational continuum. Central to scale validation is testing the invariance of scale structures between different populations (e.g., children of different ages, genders, ethnicities). We were prevented from doing so due to the sample size of subgroups and future research should seek to do this.
A limitation of the motivation model tested is that it did not include indicators of environmental support for the children’s PA. Previous research suggests that PE teacher autonomy support is associated with need satisfaction  and measures of parent, teacher and friend autonomy support  could facilitate the testing of these associations in the PA context. A further limitation of the tested model was the specification of a composite need satisfaction variable which prevented the examination of associations between the individual need satisfaction variables and each behavioral regulation. Future studies with larger samples are needed to examine this.
Although the Action 3:30 project aimed to recruit children who were not physically active through traditional or after-school sports, the participants were relatively physically active, on average achieving slightly less than 60 minutes of MVPA per day. In addition, participants who complied with the accelerometer data collection protocol reported lower external motivation than children who did not provide valid accelerometer data. It is important to examine motivational models among less active and poorly motivated children as these are arguably the most important to target in interventions. Finally, the data examined were cross-sectional and although causality is inferred by theoretical propositions, the data do not provide evidence for the direction of associations within the motivational model. Previous research supports prospective associations within similar models  and future research is required to examine such relationships within children’s PA motivation.