Issues such as delay, failure, and academic dropout are commonplace in higher education worldwide (e.g., Aina & Pastore, 2020; Álvarez, 2021; Bryson, 2023; Ibáñez-Cubillas et al., 2023; Lorenzo-Quiles et al., 2023). Argentina is far from being in a good place either: on the one hand, 38% of freshmen drop out of university, and 22% change majors in the first or the second year (Ministerio de Educación, Secretaría de Políticas Universitarias, 2021); on the other hand, delays in graduation are also customary, with only 25% of students getting their degrees in the stipulated term (Ministerio de Educación, Secretaría de Políticas Universitarias, 2019). Such delays are partly explained by courses pass-fail: official reports show that 35.7% of students cannot pass any course per academic term whereas 15.3% pass only one; a meager 14.8% pass the required number of courses per year to get their degrees in the stipulated time (Fundación Libertad, 2019). Moreover, delay usually ends up in dropout either in the mean or the long term (Seminara & Graffigna-Vaggione, 2021).

Those academic issues − delay, failure, and dropout − stem from multiple factors such as demographic, environmental, academic, didactic, and psychological, among others (Alonazi, 2018; Çizel et al., 2017; Najimi et al., 2013; Stăiculescu & Richiţeanu-Năstase, 2018; Trejo-Sánchez, 2019; Van Rooij et al., 2017; Yu et al., 2018). Among the abovementioned, psychological resources employed to deal with academic activities stand out − i.e., cognitive, meta-cognitive, motivational, behavioral, emotional, and social (Schunk & Greene, 2018).

Knowing how students use such resources is valuable not only for teachers but also for students themselves. Teachers could plan their methods and strategies accordingly and students can change their study habits and strategies to manage self-regulated learning (Tekkol & Demirel, 2018).

Models of self-regulated learning coincide, in general, in conceiving the notion as a process with different stages: these stages are usually described as task definition, goal setting and planning, execution and monitoring, as well as evaluation. The latter entails a reflection on results and contributes to regulating behavior to adapt it further. The circuit is constantly feedbacking and evolving (Muis & Singh, 2018; Panadero, 2017). It is worth mentioning that not every stage occurs necessarily in every learning situation. Occasionally, students learn some contents implicitly or involuntarily, making the self-regulatory process unnecessary (Boekaerts, 2011; Pintrich, 2000; Winne, 2011).

Even though several models of self-regulated learning were posed, some of them stand out on the grounds of their conceptual and instrumental importance (Hernández-Barrios & Camargo-Uribe, 2017; Panadero, 2017; Zhu et al., 2020). Nevertheless, each one is focused on a unique dimension of self-regulation − either cognition, motivation, or emotion, among others −, disregarding the rest of them (e.g. Boekaerts, 2011; Pintrich, 2000; Winne & Hadwin, 1998; Zimmerman, 1989). Yet, there is a consensus in describing self-regulated learning as a process in which students develop actively their knowledge as a result of setting certain goals. Accordingly, students monitor, regulate, and control not only their cognitions, motives, emotions, and behaviors but also the environment to achieve those goals (Boekaerts, 2011; Pintrich, 2000; Winne, 2011).

Studies repeatedly verified the liaison between self-regulated learning and academic achievement − i.e., significant correlations between them, the role of the former in predicting the latter, and even meta-analyses verifying such relationship as a proven effect (Almalki, 2019; Alotaibi et al., 2017; Dent & Koenka, 2015; Ergen & Kanadli, 2017; Hernández-Barrios & Camargo-Uribe, 2017; Nabizadeh et al., 2019; Sitzmann & Ely, 2011; Vaezi et al., 2018; Yip, 2019).

Considering the above and since self-regulated learning can be trained, academics recommend its promotion using different existing methods such as on-site or online workshops, tutored sessions, and the use of technology (Boekaerts et al., 2000; Broadbent et al., 2020; Núñez et al., 2011; Pintrich et al., 1987; Raaijmakers et al., 2017; Roces-Montero & Sierra, 2017; Tuckman, 2003; Weinstein & Underwood, 1985; Zimmerman, 2015).

It is noteworthy that, no matter what the methodology is, acquiring and modifying learning habits and strategies depends mostly on the student’s deliberate action (Ericsson & Lehman, 1996; Kitsantas et al., 2013). This is the reason for the use of hypermedia-based environments and intelligent tutoring systems: they work as a scaffolding structure; whereas at first those auxiliary means are intensively introduced, they are gradually retired as students become more experts in regulating their competence (Lajoie & Azevedo, 2006; Poitras & Lajoie, 2018). Thus, identifying those strategies that are less developed but potentially useful to succeed in learning activities arises as paramount: such diagnosis will let students start their customized training by themselves given the scaffold strategically organized to do so (Kim et al., 2020).

As a consequence of the above, developing a smartphone app for higher education students to self-assess and train their self-regulated learning strategies is the ultimate goal to achieve. Framed on that plan, this study constitutes such endeavor’s first stage, aimed at developing a valid and reliable scale to measure the dimensions of self-regulated learning. It will be specifically intended to be useful for students attending different majors. So, the objectives are as follows: 1) developing a scale to self-assess regulated learning habits and strategies in Argentinean higher education students, using theoretical and empirical criteria; 2) analyzing content and face validity evidence; 3) examining its factorial structure and internal consistency; 4) estimating the scores’ stability-reliability; and 5) analyzing criterion validity evidence.

Promoting self-regulated learning in higher education students is paramount when diminishing academic delay, failure, and dropout is the goal, given the positive association of the former with academic achievement. This is the reason for developing a system for students to self-assess and train their self-regulated learning strategies in a customized way via a smartphone app. Framed on that project, this study was aimed at developing a scale to assess such strategies generating valid and reliable results. To do so, theoretical and empirical criteria were used to develop 244 items grouped into six dimensions of self-regulated learning: Cognitive-, Metacognitive-, Motivational-, Emotional-, Behavioral-, and Social dimension. Even when different models were considered, special attention was drawn to the information provided by Argentinean and Spanish students attending different majors as well as by teachers, and deans. As a result, a unique version contemplating cultural and linguistic specificities, able to be used in both countries and eventually feasible to be adapted to other populations was obtained. Following this idea, experts from Spain, Argentina, and Uruguay participated in the content validity study.

Students from different majors from Argentina participated in the face validity study since retaining items useful for different fields was the goal at such stage: special emphasis was placed on reducing the error introduced by differences among majors, such as learning materials, practical assignments, or exams (Irwing & Hughes, 2018). To meet that requirement, a thorough item development and depuration process was conducted.

As a result of the above, a second version of the scale retaining 68 items was examined in terms of its internal structure; therefore, 31 items grouped into five out of the six dimensions initially posed were retained via exploratory factor analysis: Cognitive, Metacognitive-, Emotional-, Motivational-, and Social dimension. The Behavioral dimension overlapped the Metacognitive one − 6 metacognitive and 3 behavioral items. After examining the items’ content, that dimension was labeled as Metacognitive since all the behavioral items alluded to the effort allocation, which is a major metacognitive feature among activities planning (A. Brown, 1987; Pintrich, 2004) (e.g., “Generally, I force myself to study instead of postponing it”, “I make an effort to stay up to date with the material and content of each course every week”, “I am determined when it comes to studying”).

Regarding other features of the extracted dimensions, three items of the Emotional dimension are worth mentioning because of their contents related to test anxiety (e.g., “I am afraid to fail exams”, “I feel very anxious in the previous days of an exam”, and “According to the assessment method (for example, oral or written), anxiety can sometimes get the better of me.”). The items corresponding to this dimension are related only to test anxiety and its negative impact on students. These items do not describe possible strategies to cope with that specific type of anxiety, like meditation techniques, positive self-affirmations/positive self-talk, and strategies to boost self-efficacy feelings/beliefs (Jia et al., 2023; Yılmazer et al., 2024). The lack of this type of content in the items corresponding to the Emotional dimension might explain the low interfactor correlations obtained in the factorial analysis. In fact, research has shown a low association between test anxiety and learning self-regulation (Ispinar-Akcayoglu & Ozer, 2024; Wu et al., 2022). Although it is desirable to find low associations to ensure independence among factors, correlations below .30 may represent an issue when a theoretical relation among factors to represent the construct appropriately is expected (Lloret-Segura et al., 2014). Therefore, efforts will be targeted towards adding new items that improve the Emotional dimension, so adequate correlations with the rest of the dimensions can be obtained, without compromising the factors’ independence. Undoubtedly, new items will enhance the dimension’s psychometric quality since unbalanced dimensions may affect both the explained variance percentage and internal consistency (Ferrando et al., 2022; López-Pina, 2008).

The solution obtained via the exploratory factorization reached an FDI value of over .90, marginal reliability over .80, SR over 2, and EPTD over 90%. Such values indicate the adequacy of the scale to be used in higher education when personalized assessment is required (Ferrando & Lorenzo-Seva, 2018). Moreover, H indices over .80 indicate, on the one hand, that the retained items represent the construct properly and, on the other, that the model is adequate and can be replicated in other samples (Hammer, 2016). Internal consistency reached optimal values for all dimensions (ω > .80).

After the above, the solution achieved in the exploratory stage was tested via an exploratory structural equation modeling analysis (ESEM), obtaining optimal fit and parsimony indices, with CFI and TLI over .95, as well as RMSEA and SRMR equal or under .06 (Swami et al., 2023). Not only the estimated parameters were significant, but also the items loadings were the highest for the factors they were originally developed for.

Later, the ESEM model’s factorial invariance was tested by splitting the sample by administration format, fields of study, academic path, and gender; the metric equivalence of the scale’s internal structure was verified among subsamples. That indicates a low magnitude of sample bias (Davidov et al., 2018). In the professional field, such a result guarantees the correct functioning of the scale in different students attending different majors, and different stages of the academic path. Moreover, the additional validation of the online administration format could not be overlooked when considering that the scale will be included in a smartphone app. As per gender, the latent means invariance was not verified; that suggests that male and female students would obtain significantly different scores in the instrument’s dimensions. Therefore, it is advisable to calculate statistical norms per gender.

Regarding the scores’ stability reliability, no significant differences were found after a 30-day interval. That would allow a second application of the scale to evaluate the effect of any possible intervention. Therefore, students could use the app and perform the activities proposed as a result of the assessment as a continuous follow-up of their self-regulated learning strategies.

As for the criterion validity study, statistically significant correlations between the Motivational-, Social-, Cognitive- and Behavioral dimensions of the Self-Regulated Learning Scale and the dimensions of Academic Engagement were verified. This result aligns with both theoretical and empirical evidence stating that motivation, cognition and adaptive behavior precede the students’ engagement (Singh et al., 2022).

It is also important to note that the Emotional dimension of self-regulated learning did not show a significant correlation with the Academic Engagement dimension. This may be because the instrument’s Emotional dimension specifically assesses test anxiety, which occurs in a rather different moment than the actual process of Emotional engagement, which is enabled in classes that are appealing in the students’ eyes (Tejedor, 2024).

Regarding the Procrastination dimension that was used as an external criterion, its low values were associated with a higher regulation of the Motivation-, Cognition-, and Behavior dimensions and with a lower emotional regulation in terms of a low anxiety management related to test anxiety. These results are in line with the ones obtained in another research, which shows that students with self-regulated profiles procrastinate less (Manchado-Porras & Hervías-Ortega, 2021; Rebetez et al., 2015).

Concerning the Social dimension of self-regulation, it did not significantly correlate with Procrastination, which, among social factors in general terms, is more related to the use of social networks, media, internet misuse, or even unexpected events, among others; these are not considered in the developed self-regulation dimension (Redondo-Mendoza, 2022).

As per Learning Strategies, positive and significant correlations with almost all the self-regulated learning dimensions were observed (Duarte-Duarte et al., 2024); the exception was the Emotional dimension, which obtained a negative association with Abilities to Rank Information and a positive one with Resources for Learning. These results seem reasonable since, on the one hand, a low ability to identify main ideas represents a waste of time and a rise in students’ anxiety; on the other hand, higher anxiety may cause students to employ more resources, resulting in an extra-energy drain (Gutiérrez & Avero, 1995; Weinstein et al., 1987).

Academic Achievement showed positive and significant correlations with Motivation-, Cognition-, and Behavior dimensions and negative correlations with the Emotional dimension, as expected. Lastly, the Social dimension did not significantly correlate with academic performance. Although this result was not expected, it does not affect the dimension’s importance since it may act as a mediator in the learning process; such an explanation should be examined in further research.

Mainly, it was observed that the criterion validity study showed values moving from low to moderate, in the theoretically expected direction. They were acceptable assuming that, even when the constructs are theoretically related to self-regulated learning, they are not identical. It is worth mentioning that no association coefficient was over .70, concluding the lack of collinearity between variables (Yoo et al., 2014).

This study entails some limitations that will be addressed shortly in order to improve the scale’s quality: 1) the impossibility of including a Behavioral dimension due to its inclusion into the Metacognitive factor − a proper Behavioral dimension will be developed further; 2) the Emotional dimension is represented by items exclusively referred to test anxiety − items with a more heterogeneous content should be developed; 3) the test-retest study was conducted in a 30-day interval − different intervals between measures will be soon analyzed; 4) the scale’s longitudinal invariance could not be examined due to the sample size; 5) other types of criterion validity were not analyzed, such as predictive or retrospective; 6) the instrument’s incremental validity evidence was not examined to learn about the influence of each dimension in academic performance; and 7) the gender invariance study only included male and female since a large-enough sample of other genders was not available. All the latter will be promptly addressed as well.

To sum up, the study introduces a new scale to assess self-regulated learning dimensions, hosted in a smartphone’s app. It can be downloaded following a link (https://play.google.com/store/apps/details?id=com.learnapp.learnapp&pli=1). For now, the app only allows a self-assessment of a student’s self-regulated learning strategies, with the corresponding feedback and written recommendations; they are targeted to improve those strategies useful to succeed in learning activities that remain underdeveloped. The next endeavor consists of the development of customized interactive activities for the users to train habits and learning strategies. It is expected that this tool set will be useful for students not only to improve their learning quality and therefore academic achievement, but also to diminish academic delay and dropout, especially in freshmen.

This research is the starting point of a tech tool development that, along with the broad use of smartphones, offers an app for students to enable self-training of different learning strategies and self-regulation (Faza & Lestari, 2025). To achieve this goal, different theoretical approaches and information from diverse sources collected in the field led to the development of an instrument that provides an integrated assessment of self-regulatory skills in college students (e.g. Boekaerts, 2011; Pintrich, 2000; Winne & Hadwin, 1998; Zimmerman, 1989).

As it was mentioned, there are different aspects to be improved as the research moves forward. However, it is worth mentioning that the most important step has already been taken with the development of the instrument introduced in this research.