Assessing Psychometric Properties of the Self-Directed Learning Scale in Chinese University Students

Abstract

As one of the 21st-century skills, self-directed learning (SDL) has received widespread attention and has become an essential research topic in education. In this study, we examined the psychometric properties of the SDL scale (SDLS) (Lounsbury et al., 2009) with a sample of 408 Chinese undergraduates. Previous research on Chinese students has validated SDLS by using confirmatory factor analysis directly assuming SDL as a unidimensional factor as it was tested in Western countries. However, considering the rich connotation of SDL, we applied exploratory factor analysis and further used Rasch analysis based on item response theory to enrich researchers’ and practitioners’ understanding of its psychometric properties. We found additional and good psychometric evidence of a two-factor structure of SDLS: students’ initiative and ability of SDL and their self-concept of SDL. The two factors of SDLS have good internal consistency and positive correlation with formative feedback orientation, showing evidence of its criterion validity.

Keywords

self-directed learning self-directed learning scale validation higher education

Introduction

Self-directed learning (SDL), as an essential topic in education, has received widespread attention (e.g., Brockett & Hiemstra, 1991; Costa & Kallick, 2003; Hiemstra, 2013). Among the various definitions of SDL that appeared in the literature, Knowles’s definition used to be the most commonly adopted (O'Shea, 2003). Knowles (1975) described SDL as “a process in which individuals take the initiative, with or without the help of others, in diagnosing their learning needs, formulating learning goals, identifying human and material resources for learning, choosing and implementing appropriate learning strategies, and evaluating learning outcomes (p. 18).” Knowles’s definition seems to be the basis for many other definitions (e.g., Iwasiw, 1987; Spencer & Jordan, 1999).

As such definition concentrates on the skills and abilities that individuals need in the SDL process (Zhoc & Chen, 2016), thus being criticized as lacking a psychological component (e.g., persistence), which is insufficient to ensure lifelong learning (e.g., Macaskill & Denovan, 2013; Oddi, 1987). In addition, taking Lewin’s Field Theory (1951) as a conceptual basis, Leean and Sisco (1981) considered learning from the perspective of chronological progression of time—past, present, and future. They suggested that innate abilities of the rational mind guide SDL. Gradually, the connotation of SDL was integrated with the perspective of personality traits, thus transforming SDL from being a method or form of learning to a concept about developing students as autonomous learners (Macaskill & Denovan, 2013). Based on the perspective of personality traits, SDL is defined as a disposition toward learning activities in which individuals are responsible for developing and undertaking learning efforts autonomously, without being driven or directed by others (e.g., teachers or parents) (Lounsbury et al., 2009). Accordingly, a scale on SDL has been developed by Lounsbury et al. (2009) to measure learners’ SDL activities based on Brockett’s (1983) conceptualization of the SDL. This SDL scale (SDLS) is a unidimensional scale containing 10 items. Its good psychometric properties have been validated in several educational and cultural settings (e.g., Demircioğlu et al., 2018; Lounsbury et al., 2009; Zhoc & Chen, 2016).

However, the conceptualization of SDL as a personality trait has a rich connotation; for example, it reflects a person’s ability to arrange their learning process, learning experience, and disposition to learn by oneself (Brockett, 1983). Although previous studies have attempted to develop multidimensional SDL scales based on dimensions of psychological traits (e.g., Guglielmino, 1977; Oddi, 1986), they are too detailed and overloaded with items, and this attempt to quantify the construct associated with personality dimensions has certain limitations (Brockett & Hiemstra, 1991). Nevertheless, as mentioned above, the connotation of SDL not only includes the component of initiative and ability but also concerns personality characteristics. Furthermore, in Brockett’s (1983) conceptualization of SDL, personal responsibility is the cornerstone since only by taking responsibility for one’s own learning can one adopt a proactive approach to the learning process. In other words, individuals conducting SDL activities rely on the ability to plan, implement or assess their learning process and on their personalities, such as initiative, disposition, and self-concept, to endeavor learning activities autonomously. As Lounsbury et al. (2009) developed the SDLS based on Brockett’s (1983) theoretical framework of the SDL construct, given the rich connotation of SDL, it is reasonable to assume that the SDLS contains more than one dimension. Therefore, it is necessary to explore the potential dimensions of the existing SDLS of Lounsbury et al. (2009) based on its connotations to reveal its richness. In addition, previous studies have typically used structural equation modeling approach like confirmatory factor analysis to provide validity and reliability evidence for the scales (e.g., Lounsbury et al., 2009; Zhoc & Chen, 2016). It is worth noting the use of latent trait theory approaches for testing. For example, item response theory (IRT) is rigorous in providing detailed psychometric properties of each item and personal abilities in responding to these items (Edelen & Reeve, 2007). Therefore, the present study sought to build on the Chinese version of Lounsbury et al.’s (2009) SDLS validated by Zhoc and Chen (2016) and conduct exploratory analysis and IRT analysis to distinguish possible dimensions within the SDLS that reflect key psychological traits.

Method

Participants

408 Chinese undergraduates (male: 282) participated in this study. They were recruited in 2021 from two teachers’ colleges in Guangxi and Anhui provinces, located in South and East China, respectively. The per capita disposable income of the two selected cities in 2021 was about $4644 and $4,200, respectively, which was lower than the national per capita disposable income of residents ($5,445, National Bureau of Statistics of China, 2022). Both teachers’ colleges have been recognized as national comprehensive universities known for their distinctive teacher education. The majority (89.5%) of the participants aimed to pursue a teaching career after graduation. 44.5% majored in mathematics, 40.0% studied chemistry, and the remaining studied environment, computer, geography, and other science subjects. Most of them (89.7%) were in the 19–21 age group. Of the students who participated, 28.7% were first-year students; 62.7% were sophomores; 4.4% were juniors, and the others were seniors. We provided all participants with written informed consent. We first explained the purpose of the study and then guided them to fulfill the questionnaires.

Measures

Self-Directed Learning

We drew this scale from Lounsbury et al. (2009) and Zhoc and Chen (2016), who developed the scale based on Brockett’s (1983) theoretical framework on the SDL construct. Since Zhoc and Chen (2016) have investigated its psychometric properties based on Chinese university students in Hong Kong, we directly used the Chinese version. It included 10 items (see Appendix). One sample item was “I regularly learn things on my own outside of class.” Respondents were instructed to indicate their answers on a four-point scale from 1 = strongly disagree to 4 = strongly agree. Cronbach’s alpha was 0.88.

Formative Feedback Orientation

This scale (12 items) was adapted from the Educational version of Yang et al.’s (2014) formative feedback orientation (FO) scale, which is based on Linderbaum and Levy’s (2010) work in organizational psychology. FO has been proven to be associated with a wide range of learning outcomes (see Yang, 2021, for a systematic review). The FO scale assessed students’ receptivity to feedback to change academic performance and learning outcomes. A sample item was “I can use teachers’ feedback effectively to improve my academic performance.” Students rated on a scale from 1 = strongly disagree to 4 = strongly agree. Cronbach’s alpha was 0.89.

Data Analyses

First, we conducted exploratory factor analysis (EFA) with SPSS 26.0 using the principal axis factoring extraction method and oblimin with Kaiser normalization rotation method to identify potential factor groupings of the items. Then, we further explored the data using a multidimensional Rasch model (Adams et al., 1997). We ran the Rasch analysis with ConQuest software (version 1.0.0.1; Wu et al., 2007). In addition, we computed descriptive statistics and the correlation between SDLS and FO with SPSS 26.0.

Results

Exploratory Factor Analysis

The values for Kaiser–Meyer–Olkin measure of sampling adequacy was 0.82, and Bartlett’s test of sphericity was χ² (45) = 1654.60, p < .001, suggesting that the data were appropriate for factor analysis. Table 1 shows the results of the rotated factor structure of SDLS. We identified two factors with eigenvalues higher than 1. Factor 1 contained 5 items (SDL2, SDL4, SDL6, SDL8, SDL10), and factor 2 had 5 items (SDL1, SDL3, SDL5, SDL7, SDL9). Factor 1 was labeled SDL initiative and ability, which describes individuals’ active pursuit and desire to engage in SDL activities and their capacity for SDL activities. This factor explained 49.6% of the common variance, with an eigenvalue of 4.96. Factor 2 was labeled SDL self-concept, describing how individuals viewed, perceived, and experienced SDL. It interpreted 21.6% of the total variance, with an eigenvalue of 2.16.

Table 1.

Rotated Factor Structure of SDL Scale.

Item	Factor
Item	SDL Initiative and Ability	SDL Self-Concept
SDL1	.102	.716
SDL2	.867	.034
SDL3	.003	.900
SDL4	.933	−.029
SDL5	.054	.789
SDL6	.913	−.029
SDL7	.027	.798
SDL8	.902	−.001
SDL9	−.114	.724
SDL10	.812	.046

Note. SDL = Self-directed learning; each variable in the first column refers to an item of the SDLS, for example, SDL1 = item 1 in the SDLS.

Rasch Analysis

Table 2 displays the item difficulties, standard errors, and item fit statistics of the Rasch model. The results show that the item fit statistics (Unweighted/Weighted MNSQ) of multidimensional Rasch analysis fell within the range of 0.42 to 1.10, which are generally considered to indicate a good fit to the Rasch model (Bond & Fox, 2015; Linacre, 2022; Wilson, 2005).

Table 2.

Item Statistics in Rasch Analysis.

Item	Item difficulty^a	SE	Unweighted (Outfit) MNSQ	Weighted (Infit) MNSQ
SDL1	−4.57	.16	.60	.86
SDL2	.02	.16	.79	.87
SDL3	−4.67	.16	.51	.73
SDL4	.60	.17	.46	.74
SDL5	−4.85	.16	.54	.76
SDL6	−.01	.16	.42	.71
SDL7	−4.47	.16	.59	.83
SDL8	.60	.17	.45	.80
SDL9	−4.20	.16	.81	.94
SDL10	.34	.17	.95	1.10

Note. For items of the scales, please see the Appendix. SDL = Self-directed learning; each variable in the first column refers to an item of the SDLS, for example, SDL1 = item 1 in the SDLS.

All item difficulties are in logits.

Figure 1 shows the Wright map for the multidimensional Rasch model. The continua on the left-hand side show the distribution of students on the SDLS. The distribution of the corresponding items under the two sub-dimensions is on the right. Table 2 and Figure 1 indicate that item difficulties of the multidimensional Rasch model were in the range of −4.85 to 0.60 logits. The Wright map in Figure 1 demonstrates that the range of item difficulty on the scale matches the scope of student ability.

Figure 1.

Wright Map for the Multidimensional Rasch Model. Note. Each “X” represents 6.3 cases.

Internal Consistency and Criterion Validity

The Cronbach’s alpha coefficients of the two components of SDL were 0.87 and 0.95, respectively. The EAP/PV reliabilities generated by ConQuest for the two sub-dimensions of SDLS (Factor 1 and Factor 2) were 0.61 and 0.84, respectively. Both indicate an acceptable level of internal consistency. The overall SDL and FO correlation was 0.76 (p < .001). The correlation between the two components of SDL was 0.29 (p < .001). The two components of SDL were also positively correlated with FO (r = 0.46, p < .001; r = 0.72, p < .001).

Discussion

This study aimed to explore the psychometric properties of the Chinese version of Lounsbury et al.’s (2009) SDLS and distinguish possible dimensions within the SDLS that reflect key psychological traits in a higher education setting in China. We applied EFA and IRT analyses to generate evidence regarding the construct validity and tested its criterion validity by conducting the correlation between SDLS and FO.

The results of EFA and Rasch analysis based on IRT showed evidence of reliability and validity of the two factors of SDLS: SDL initiative and ability and SDL self-concept. This finding is inconsistent with previous studies that have validated the unidimensional characteristics of this scale in both western cultures and a similar Chinese cultural context (Hong Kong) (e.g., Lounsbury et al., 2009; Zhoc & Chen, 2016). The psychometric properties of the SDLS in this study are consistent with the core components of the SDL conceptualization, meanwhile suggesting the construct’s multidimensional properties. The two factors were similar to some of the eight factors of the Self-Directed Learning Readiness Scale (Guglielmino, 1977). The former is considered to be a possible underlying factor of the SDL. Lack of an inherent disposition of self-learning or lack of ability to arrange self-learning independently, SDL activities cannot take place. The other factor, SDL self-concept, is also vital to one’s SDL activities. It is strongly associated with SDL and therefore has been considered by previous researchers such as Brockett and Hiemstra (1991) and Oddi (1986) as one of the two reasons for building their theoretical framework of SDL based on individual personality.

Besides, Cronbach’s alpha coefficients and the EAP/PV reliabilities also revealed that the two factors were internally consistent. The criterion validity results also showed that the two factors of SDLS were positively correlated with FO. Our findings suggested that the two-sub-dimension structure of SDLS could be applied to the sample of higher education from mainland China.

The current study had expected limitations. Our sample was mainly students from teachers’ colleges who majored in science. Although the selected teachers’ colleges were also comprehensive universities in China, the SDLS may have different properties for students of liberal arts and with a non-teacher education comprehensive university background. Thus, researchers and educators must be cautious about generalizing the psychometric properties of SDLS to other Chinese university students. Future studies with samples of more diverse backgrounds are encouraged. Another limitation is that the consequential validity of including learning outcomes (e.g., academic achievement) has not yet been examined. Future studies can explore whether a better understanding of students’ self-directed learning by using the SDLS leads to good academic achievement. Despite these limitations, the present study contributes to the extant literature by validating the psychometric properties of Lounsbury et al.’s (2009) SDLS in a higher education setting in mainland China. Two meaningful factors are derived: SDL initiative and ability and SDL self-concept, which extend the current understanding of SDLS as a unidimensional instrument.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by The Research Cluster Fund from the Education University of Hong Kong (R4203) to Dr. Lan Yang, the corresponding author.

Ethical Approval

Ethical approval was achieved from Human Research Ethics Committee, the Education University of Hong Kong before survey administration.

ORCID iDs

Di Zhang

Lan Yang

Appendix

Table A1.

Items of the Self-Directed Learning Scale.

	Item Description
SDL1	I regularly learn things on my own outside of class.
SDL2	I am very good at finding out answers on my own for things that the teacher does not explain in class.
SDL3	If there is something I don’t understand in a class, I always find a way to learn it on my own.
SDL4	I am good at finding the right resources to help me do well in university.
SDL5	I view self-directed learning based on my own initiative as very important for success in university and in my future career.
SDL6	I set my own goals for what I will learn.
SDL7	I like to be in charge of what I learn and when I learn it.
SDL8	If there is something I need to learn, I find a way to do so right away.
SDL9	I am better at learning things on my own than most students.
SDL10	I am very motivated to learn on my own without having to rely on other people.

Note. SDL = Self-directed learning; each variable in the first column refers to an item of the SDLS, for example, SDL1 = item 1 in the SDLS.

References

Adams

R. J.

Wilson

Wang

W. C.

(1997). The multidimensional random coefficients multinomial logit model. Applied Psychological Measurement, 21(1), 1–23. https://doi.org/10.1177/0146621697211001

Bond

T. G.

Fox

C. M.

(2015). Applying the Rasch model: Fundamental measurement in the human sciences (3rd ed.). Routledge.

Brockett

(1983). Self-directed learning and the hard-to-reach adult. Lifelong Learning: The Adult Years, 6(8), 16–18.

Brockett

R. G.

Hiemstra

(1991). Self-direction in adult learning: Perspectives on theory, research, and practice. Routledge.

Costa

A. L.

Kallick

(2003). Assessment strategies for self-directed learning. Corwin Press.

Demircioğlu

Z. I.

Burak

Ö. G. E.

Fuçular

E. E.

Çevik

Nazligül

M. D.

Özçelik

(2018). Reliability, validity and Turkish adaptation of self-directed learning scale (SDLS). International Journal of Assessment Tools in Education, 5(2), 235–247. https://doi.org/10.21449/ijate.401069

Edelen

M. O.

Reeve

B. B.

(2007). Applying item response theory (IRT) modeling to questionnaire development, evaluation, and refinement. Quality of Life Research, 16(1), 5–18. https://doi.org/10.1007/s11136-007-9198-0

Guglielmino

L. M.

(1977). Development of the self-directed learning readiness scale [Unpublished doctoral dissertation]. University of Georgia.

Hiemstra

(2013). Self-directed learning: Why do most instructors still do it wrong? International Journal of Self-Directed Learning, 10(1), 23–434.

10.

Iwasiw

C. L.

(1987). The role of the teacher in self-directed learning. Nurse Education Today, 7(5), 222–227. https://doi.org/10.1016/0260-6917(87)90005-0

11.

Knowles

M. S.

(1975). Self-directed learning: A guide for learners and teachers. Association Press.

12.

Leean

Sisco

(1981). Learning projects and self-planned learning efforts among undereducated adults in rural Vermont (Final Report No. 9–1051). National Institute of Education.

13.

Lewin

(1951). Field theory in social science: Selected theoretical papers. Harper & Brothers.

14.

Linacre

J. M.

(2022). A user’s guide to winsteps ministep: Rasch-model computer programs: Program manual 5.3.3. Winsteps.

15.

Linderbaum

B. A.

Levy

P. E.

(2010). The development and validation of the feedback orientation scale (FOS). Journal of Management, 36(6), 1372–1405. https://doi.org/10.1177/0149206310373145

16.

Lounsbury

J. W.

Levy

J. J.

Park

S. H.

Gibson

L. W.

Smith

(2009). An investigation of the construct validity of the personality trait of self-directed learning. Learning and Individual Differences, 19(4), 411–418. https://doi.org/10.1016/j.lindif.2009.03.001

17.

Macaskill

Denovan

(2013). Developing autonomous learning in first year university students using perspectives from positive psychology. Studies in Higher Education, 38(1), 124–142. https://doi.org/10.1080/03075079.2011.566325

18.

National Bureau of Statistics of China . (2022). Residents’ income and consumption expenditure in 2021. http://www.stats.gov.cn/tjsj/zxfb/202201/t20220117_1826403.html

19.

Oddi

L. F.

(1986). Development and validation of an instrument to identify self-directed continuing learners. Adult Education Quarterly, 36(2), 97–107. https://doi.org/10.1177/0001848186036002004

20.

Oddi

L. F.

(1987). Perspectives on self-directed learning. Adult Education Quarterly, 38(1), 21–31. https://doi.org/10.1177/0001848187038001003

21.

O'Shea

(2003). Self-directed learning in nurse education: A review of the literature. Journal of Advanced Nursing, 43(1), 62–70. https://doi.org/10.1046/j.1365-2648.2003.02673.x

22.

Spencer

J. A.

Jordan

R. K.

(1999). Learner centred approaches in medical education. British Medical Journal, 318(7193), 1280–1283. https://doi.org/10.1136/bmj.318.7193.1280

23.

Wilson

(2005). Constructing measures: An item response modeling approach. : Erlbaum Associates.

24.

M. L.

Adams

R. J.

Wilson

M. R.

Haldane

S. A.

(2007). ACER ConQuest version 2.0: Generalised item response modelling software [computer program]. Australian Council for Educational Research.

25.

Yang

(2021). The role of feedback orientation in converting external feedback to learning opportunities for implementing assessment-as-learning in the context of feedback. In Yan

Yang

(Eds.), Assessment as learning: Maximising opportunities for student learning and achievement (pp. 53–76). Routledge.

26.

Yang

Sin

K. F.

X. Y.

Guo

J. S.

Lui

(2014). Understanding the power of feedback in education: A validation study of the feedback orientation scale (FOS) in classrooms. The International Journal of Psychological and Educational Assessment, 16(1), 21–46.

27.

Zhoc

K. C.

Chen

(2016). Reliability and validity evidence for the self-directed learning scale (SDLS). Learning and Individual Differences, 49, 245–250. https://doi.org/10.1016/j.lindif.2016.06.013