The Short Version of Approaches and Study Skills Inventory for Students: A Confirmatory Factor Analysis in Vietnamese Nursing Students

,


Introduction
Learning approaches are the diferent methods through which students engage in and execute learning activities.In higher education institutions, three distinct learning approaches, namely deep, surface, and strategic learning, have been identifed early in a conceptual framework of Marton and Saljo (1976), which is one of the most popular concepts for identifying the diferent individual behaviors in learning [1].Specifcally, this framework is a concept about the students' motivation for their learning and their use of appropriate strategies to balance the relationship between themselves, context, and task during the learning process.Firstly, learners with a deep approach are mainly driven by an intention to understand and gain their knowledge through clear evidence and the connecting of ideas in the learning materials, which includes assessment of the individual's own understanding [2].By contrast, surface learners usually have a fear of failure, which is why they attempt to reproduce knowledge and learning materials.Teir predominant strategies are aimed at overcoming the exam by adopting current knowledge and learning the sign rather than exploring what exactly caused the phenomenon [3].Finally, the third group, known as strategic learners, has a desire to achieve the best learning results rather than interests and ideas (deep approach) or a fear of failure (surface approach) [4].Tis kind of student always develops a wide range of strategies that integrate available resources and meet exam criteria to fnish the courses with high scores.
By the literature review, student's study approaches and engagement directly impact their learning performances [5,6].As such, approaches to learning refect individual diferences in their intentions when facing a learning situation, as well as the corresponding strategies by which these intentions are accomplished.In recognition of its importance, studies focusing on learning approaches have been widely conducted among university students with the use of the Marton and Saljo framework (1976), [7,8].Tese studies reported constant evidence that demonstrated individual learning diferences caused by how students' learning approaches exist, such as deep, strategic, and surface.
Students' learning approaches have been assessed as quantitative research by diverse inventories, which are the Approaches to Studying Inventory [9], the Revised Approaches to Studying Inventory [10], and the Approaches and Study Skills Inventory for Students (ASSIST) [11].Te ASSIST, the latest version to measure individual learning approaches, consists of 18 items with the level of students' agreement on a 1-5 scale and has been widely applied because of its brevity in contrast with the full instrument of 52 items.Tree main scales (deep, surface, and strategic approaches to learning) would be measured by the calculation of six items for each element.Tis scale was used in several medical felds, such as medicine [12], psychology [13], and nursing [14].
It is clear that reliability and validity are two important steps for any kind of measurement [15].Previous studies reported that the internal consistency of three dimensions in the ASSIST short version has been variable, ranging from a high, acceptable, or even lower range.Specifcally, Cronbach's alpha ranged from 0.72-0.76,which was satisfactory [16] in the original study and became better in several studies [12,17]; however, this index was less than 0.7 in other studies [16,18,19].Regarding construct validity, three factors of the ASSIST short version were confrmed by using a confrmatory factor analysis (CFA) [17] or a principal component factor analysis of the items [18].However, one item in the deep approach ("When I read, I examine the details carefully to see how they ft in with what's being said") did not ft with this scale [18].
In Vietnam, the reform process in higher education, which was an element of the national strategy for innovation, namely "Doi moi," has a dynamic infuence on nursing education [20].Tis project aimed to strengthen the capacity of nurse training as well as the quality of graduate nursing students' ability to adapt to international professional standards.As such, it required a need to modify current pedagogical approaches with a shift to increase active/experiential learning, refective teaching, and learning strategy application [21].In fact, to the best of our knowledge, there was no validated questionnaire that aimed to assess students' learning approaches and was reported to be compatible with Vietnamese context and culture.Given the pace of educational change in Vietnam, it is necessary to validate a tool to assess nursing students' learning approaches.Te purpose of this study, therefore, was to validate a Vietnamese version of the ASSIST short version by using confrmatory factor analysis.

Te ASSISST Short Version.
Te ASSIST short version consists of 18 items aiming to identify students' approaches to studying [22].Te measurement has three subscales with six items each: deep, surface, and strategic approach.Each item of this questionnaire was rated in the range from 5 � strongly agree, 4 � agree, 3 � neutral or undecided, 2 � disagree, or 1 � strong disagree.Te overall score (possible score: 18-90) and three subscale scores (possible score: 6-30 for each subscale) were calculated, and the higher score indicates the better perception of students regarding learning approaches.

Translation Process.
Te ASSIST short version has been translated and used in several countries which are not English-speaking, such as Norway [18], South Africa [17], and Kuala Lumpur [12].In this study, using a backtranslation process, the questionnaire was translated by two independent bilingual experts [23].Specifcally, this process was performed by a Vietnamese nurse lecturer and nursing student who all have very good skills in using English.After comparing and integrating diferent versions, a team of another Vietnamese nurse lecturer who is competent in using both languages and a native-speaker nursing student were responsible for the back-translation of a revised translation.Finally, the minor in-consistent in wording were adjusted, and no signifcant diference was found.

Validation of the Vietnamese ASSIST Short Version.
A cross-sectional study was conducted on undergraduate nursing students for three months (November 2021 to January 2022).

Participants and Sampling.
Undergraduate nursing students were recruited in this study by using a convenience sampling method.Inclusion criteria were nursing students in the bachelor program and willing to participate in the study.Students in the in-service study mode or those who could not be contacted during the period of data collection were excluded from this study.

Data Collection
Procedure.An online survey was sent to students' nursing groups with the support of the local nurse lecturers.Participants had the fexibility to respond to the questionnaire using any internet-enabled device, including cell phones, tablets, or laptops.It was informed to the students that their participation in the study was entirely voluntary, and they would not receive any fnancial support or face any disadvantages for choosing not to participate.Prior to the completion of the questionnaire, students were 2 Nursing Forum provided with the study's information sheet regarding anonymity during data collection.Tis ensured that they were well-informed about the study's purpose and their rights as participants.Tis survey was also set in "Limit to one response" mode, and all questions were placed in the "required" mode to minimize the duplication rate and missing data, respectively.By submitting the completed online survey, participants indicated their consent to taking part in the study.

Reliability.
Te Vietnamese version of the 18-item ASSIST was tested to establish internal reliability by using Cronbach's alpha among 102 nursing students in a nursing school in Vietnam as the recommendation of at least 100 subjects for a high-reliability estimate [24].

Construct Validity.
Te study invited nursing students from ten nursing schools across Vietnam to participate.Te study checked whether the original three subscales ft our data by using confrmatory factor analysis (CFA).To ensure a stable factor construction, at least 300 participants should be included in the study [25].During this phase, 1340 nursing students completed the online survey and were included.

Data Analysis. Tis study used SPSS version 20 and
AMOS version 20 for data analysis.By using SPSS, descriptive statistics were used to assess the participants' characteristics as well as the level of the students' learning approach.Te internal consistency of the scale was determined by computing Cronbach's alpha, corrected item-total correlation (CITC), and Cronbach's alpha if items deleted values (CAID).For good reliability, the alpha should range from 0.70 to 0.95 [15].And the item was considered to have made a substantial contribution to the measured construct the CITC value exceeded 0.3 and the CAID demonstrated a decrease [26].
Confrmatory factor analysis was used to examine the factorial structure of the ASSIST short version.Te model chi-square goodness-of-ft (set at p < 0.05) and approximate ft indices were used to test the model ft [27].For approximate ft indexes, the root mean squared error of approximation (RMSEA ≤ 0.08), the Tucker-Lewis index (TLI ≥ 0.9), the comparative ft index (CFI ≥ 0.9) were used [27,28].Te signifcance of regression weighted (standardized loading factor) estimates indicates that the indicator variables are signifcant and representative of their latent variables.Modifcation indices (MI) were employed to identify potential correlations between variables.As such, standardized regression weights and MI were used as indicators to select which items ft to remain in the model [28].

Characteristics of Participants.
A total of 1442 nursing students participated in this study (Table 1).Most of them were female (89.39%), second year students (34.26%), and studying in a public school (78.57%).Te average age was 22.16 ± 4.31 years old.

Descriptive Statistics for Each Item in the ASSIST Short
Version.Table 2 shows that the mean scores of the Vietnamese ASSIST short version were 63.94 ± 7.45 out of 90.Te scores for deep, strategic, and surface were 21.95, 20.62, and 21.36, respectively.

Reliability.
Cronbach's alpha of the total scale was 0.89 (Table 3), which indicates the internal consistency of the Vietnamese ASSIST short version was good.Cronbach's alphas for deep, strategic, and surface approaches were 0.82, 0.89, and 0.70, respectively.Te CITC values for the scale ranged between 0.30-0.78which indicates all items highly contributed to the constructs of the ASSIST short version.4, the original model did not report a good ft.After modifying some items with high modifcation index covariance (i.e., 4-Tere is not much of the work here that I fnd interesting or relevant and 8-Much of what I am studying makes little sense: it is like unrelated bits and pieces), 4-and 18-Often worry about whether I'll ever be able to cope with the work properly), the ft of the second model was acceptable.Te CFA for the three-subscale model showed a moderate ft (X 2 /df � 7.097, p < 0.001, CFI � 0.903, TLI � 0.886, and RMSEA � 0.067).

Discussion
Te main purpose of this study was to establish the reliability and validity of the ASSIST tool when translated into Vietnamese.
In this study, Cronbach's alpha scores ranged from 0.72 to 0.82 for subscales and 0.86 for the total scale, satisfying the rule-of-thumb guide of 0.70 and indicating a good internal reliability level.In fact, there have been plenty of studies that focus on the psychometric properties of the English original version of the ASSIST scale, but none have yet report on the Vietnamese version.Te same pattern was seen in other studies, as Cronbach's alpha was generally above 0.80 for the deep and strategic approach scales and above 0.70 for the surface approach scale [18].Sadler-Smith and Tsang (1998) administered a 38-item version of the RASI to 183 Hong Kong students and reported reliability coefcients (Cronbach's α) of 0.80 for the deep, 0.82 for the strategic and 0.79 for the surface scales [29].Another study that took place in six British universities also showed that Cronbach's alpha values for the three subscales ranged between 0.80 and 0.87 [11].In the short version of the ASSIST, it is favorable that each subscale consists of six items, facilitating a fair assessment of reliability across the three subscales.

Nursing Forum
Regarding the validity of the scale, a three-factor solution was used in this study: deep approaches (Factor 1), strategic approaches (Factor 2), and surface approaches (Factor 3), respectively.It has been demonstrated that the items presented in the short ASSIST complied with the theoretically proposed factor structure [4,7,8].To be specifc, the original model reported several items with high levels of modifcation, including items 8 and 4, and 18 and 4. In fact, the main meaning of these three items is the level of interest related to study work.After modifcation, the second model was showed high levels of goodness of ft.Specifcally, the goodness of ft values of the GFI, CFI, and RMSEA also reported a good ft.Te value of 0.067 in this study was acceptable.Te values of CFI and GFI are also both higher than 0.90, indicating an acceptable ft between the model and the data.Tese indices together provide a comprehensive evaluation of the model ft and are strong evidence that there is a favorable ft between the specifed measurement model and the observed data.Similarly, a study conducted among 345 frst and secondary mathematics students in South Africa, reported that three factors of the short version of ASSIST were confrmed by the use of CFA [17].Te fnding showed that the chi-square was statistically signifcant (X 2 � 210.94, df � 132, p < 0.0001).Also the goodness of ft indices were TLI � 0.948, RMSEA � 0.054, and CFI � 0.947 [17].Since then, the results suggest that the Vietnamese short version of ASSIST is a reliable and valid tool that could be used to assess students' learning approaches in Vietnam.Trough the use of this questionnaire, it provides evidence for nurse lecturers and educational settings to implement proper interventions to support nursing students following the current learning approaches.
Nursing Forum However, there were some limitations to our study.Firstly, a cross-sectional study design was used to assess the reliability of the ASSIST short version in this study.Tis fnding is considered as an initial report on the reliability and validity of the Vietnamese ASSIST short version; therefore, further studies are encouraged to test the stability of this instrument such as using a test-retest method.Obviously, test-retest reliability is used to assess how consistent a test is, which identifes the stability and validity of the questionnaire over a period of time [30].Secondly, the study is limited by the use of the self-report methodology.Tis is said to be due to the inability to validate by objective means whether responses comply with social norms or not.Tis may also apply in the case of this study, where students may feel compelled to respond in a certain way if it is socially desired.Terefore, future studies may need to take a more holistic approach by integrating self-reported measures with additional measurement techniques [31].

Conclusions
Our fndings supported the proposed three-factor structure of the short version of ASSIST in the context of Vietnamese nursing students.Tis questionnaire is a useful tool for educators and educational institutions to initially assess students' learning approaches, which will be used for future interventions.

Table 4 :
Te results of confrmatory factor analysis.