Socio-cognitive determinants of plagiarism intentions among university students during emergency online learning: Integrating emotional, motivational, and moral factors into theory of planned behavior

Abstract

The exceptional circumstances surrounding the COVID-19 closures of campuses and emergency online learning have caused challenging circumstances on preserving academic integrity. Still, little is known about how the interplay between diverse contextual and psychological determinants influences beliefs and inclinations to plagiarism during online learning. The current study aims to understand better multiple factors that predict attitudes and intentions to commit plagiarism during and after the pandemic. To that end, an extended model based on the theory of planned behavior (TPB) that examines the impact of socio-psychological, emotional, motivational, and ethical factors explaining plagiarism intentions was tested. The study applied a survey instrument to a sample of 435 undergraduate students from three universities in Oman. Using the Partial Least Squares-Structural Equation Modeling (PLS-SEM), the results showed that fear of COVID-19 significantly and positively impacted the plagiarism attitude. Academic self-efficacy significantly and negatively influenced attitudes to plagiarism. All TPB variables significantly influenced intention to plagiarize, including subjective norms, attitudes perceived behavioral control and past behavior, except moral obligation. The current study’s findings contributed to theory advancement by extending TPB to examining antecedents to subjective norms toward plagiarism and emotional and motivational determinants of attitudes. Finally, the current study recommends practical and research implications for curbing digital plagiarism in higher education post to the pandemic.

Keywords:

• plagiarism
• fear of COVID-19
• theory of planned behavior
• motivation
• academic self-efficacy
• moral obligation

1. Introduction

Plagiarism has a long-standing history as a universally inherent part of any educational system. Students persist in committing plagiarism despite ethics interventions. Such interventions convey deteriorating attitudes to plagiarism but never address nor tackle the psychological factors influencing deliberate or unintentional plagiarism (Moss et al., 2018).

Internet search engines allow immediate and effortless access to large amounts of information. Thus, digital environments embrace new contexts for a wide range of plagiaristic practices committed by those used to plagiarize (Kauffman & Young, 2015). Digital plagiarism is linked to the interaction with contextual incentives for plagiarism in digital settings, such as affordances of copy and paste functions from a webpage (Kauffman & Young, 2015). In such a case, where ICT’s educational affordances can potentially collide against education’s honor codes, the propensity to digital plagiarism is anticipated to increase among students (Johnson, 2009). Recently, plagiarism might have become even more intractable as the world has faced dramatic changes due to the shift to online learning modes during the COVID 19 pandemic. (Nguyen et al., 2020). Online learning environments might have offered students more flexibility, leading them to engage in dishonest behaviors more often than face-to-face modalities (King & Case, 2014; Lanier, 2006).

Empirical literature reveals that various factors influence plagiarism intentions among undergraduates. Such factors can be classified into demographic, sociocultural, socio-psychological, socio-cognitive, contextual, and moral factors. For instance, several related studies have investigated demographic factors such as gender, age, academic major, year of study, socioeconomic status, and ICT expenditure (e.g., Tremayne & Curtis, 2021). Furthermore, previous studies have explored the effects of sociocultural factors, including national cultural dimensions, social norms and socialization processes (Al-Nuaimi et al., 2017; Thomas, 2017). Although previous studies have examined a wide variety of factors, the emotional factors are the least understood (Tindall & Curtis, 2020). Consequently, this study offers sustainable evidence-based practical implications to control plagiarism as an integral part of a sustainable online learning model. The following research question guided the study:

How well does the current study’s integrated model based on the extended TPB account for the students’ intention to commit plagiarism during emergency online learning?

2. Theory and hypotheses development

The current study proposes an extended theoretical model based on the theory of planned behavior (TPB). TPB stipulates that three beliefs guide human behavior: (1) behavioral beliefs (i.e., attitudes) referring to the expected outcomes of a behavior, (2) normative beliefs (i.e., subjective norms) referring to normative expectations of significant referents, and (3) control beliefs (i.e., perceived behavioral control) referring to factors facilitating or impeding the performance of the behavior. Taken together, attitudes (ATT), subjective norms (SN), and perceived behavioral control (PBC) are determined by behavioral, normative, and control beliefs (Ajzen, 1991). The present study expands the TPB by incorporating two additional dimensions into the model: past behavior (PB) and moral obligation (MO), considering their roles on intention addressed by the literature (e.g., Conner & Armitage, 1998; Cronan & Al-Rafee, 2008; Cronan et al., 2018). Hence, the current study deems intention (INT) to plagiarism using digital sources as the function of ATT, SN, PBC, PB, and MO. The model takes further steps forward by integrating new variables functioning as external antecedents of SN and ATT, driven from various theories (e.g., deontological ethics, goal orientation, deterrence, and social learning; Murdock & Anderman, 2006; Yoon, 2011). In the following subsections, components and hypotheses of the extended TPB are explained.

2.1. Attitude

Attitude (ATT) is defined as; “the degree to which a person has a favorable or unfavorable evaluation or appraisal of the behavior in question” (Ajzen, 1991, p. 188). Here, ATT denotes students’ positive or negative evaluation of using others’ digital sources without acknowledgment in their assignment during the COVID19 pandemic. More recent evidence put forth that when students possess more favorable attitudes towards using ICT sources unethically, they will be expected to engage in such behavior more often (Al-Nuaimi et al., 2021; Farooq & Sultana, 2021). Thus, the current study hypothesizes that:

H1: ATT has a significant positive influence on plagiarism intention.

2.2. Subjective norms

Subjective norms (SN) constitute the normative component of TPB and refer to “the perceived social pressure to perform or not to perform the behavior” (Ajzen, 1991, p. 188). Here, SN refers to the opinions of significant others regarding the inappropriate use of digital sources, which are an essential determinant of ethical behavior. For example, D. L. McCabe et al. (2002) have proven that peers’ perceptions are the most influential contextual variables on academic dishonesty. Similarly, D.L. McCabe et al. (2008) have provided evidence that peers’ perceptions substantially affect academic dishonesty behavior, especially in collectivistic cultures. But, in more individualistic societies, the relationship between SN and INT to commit academic dishonesty was found barely significant at the .05 level (Stone et al., 2010). Considering the cultural context of the study, we presume that:

H2: SN towards not engaging in plagiarism has a negative influence on plagiarism intention.

2.3. Perceived behavioral control

Perceived behavioral control (PBC) is defined as the “perceived ease or difficulty to perform the behavior” (Ajzen, 1991, p. 188). Typically, the more resources and opportunities, and the fewer obstacles and impediments are perceived, the greater levels of PBC over behavior are experienced (Ajzen, 1991). Related studies on academic dishonesty have corroborated this postulation showing that a higher degree of PBC is linked with more tendencies to commit a given behavior (e.g., Hendy & Montargot, 2019; Maloshonok & Shmeleva, 2019). Considering that the current study concentrates on plagiarism in online classes, one can argue that higher degrees of abilities, resources, and opportunities readily available for performing the behavior might be connected to higher likeliness to conduct that behavior. Hence, we hypothesize that:

H3: PBC has a significant positive influence on plagiarism intention.

2.4. Past behavior

Past behavior (PB) refers to the frequency of performance of the behavior in question in the past. In this study, PB refers to the severity of plagiarism behaviors using digital sources during the COVID19 pandemic.

The literature produced mixed views regarding the inclusion of PB into the TPB. A research line conceptualizes PB from the lens of the habituation phenomenon. Based on the reasoned action perspective of routinized behavior, Ajzen (2002) maintains that the residual effects of past behavior are “attenuated when measures of “intention and behavior are compatible.” By contrast, the impact of past behavior disappears when attitudes are well-established, intentions are solid and well-planned (Ajzen, 2002, p. 107). On the other hand, in their review study, Conner and Armitage (1998) put forth that differentiation of past behavior from a habit is often unclear, particularly concerning measurement terms. Furthermore, Conner and Armitage (1998) reported empirical evidence that PB accounted for unique variance in behavioral intention.

Aligning with this conceptualization and acknowledging that the COVID19 pandemic has brought along a novel and different context (e-learning) for students to practice plagiarism, one can argue that past behavior (PB) of plagiarism may influence the intention to plagiarize. Relevant empirical evidence suggests that past behavior significantly affects the intention to plagiarize (e.g., Cronan et al., 2018; Uzun & Kilis, 2020). Accordingly, we formulate the following hypothesis:

H4: PB has a significant positive influence on plagiarism intention.

2.5. Moral obligation

From the standpoint of deontological ethics, moral obligation plays crucial role in ethical judgment formation (Yoon, 2011). Moral obligation (MO) is defined as “personal feelings of moral obligation or responsibility to perform, or refuse to perform, a certain behavior” (Beck & Ajzen, 1991, p. 289). MO stands for one’s feeling of guilt or personal obligation concerning carrying out or not carrying out a behavior (Cronan & Al-Rafee, 2008). In the current study, MO refers to students’ feelings of guilt or personal obligation regarding using digital sources without acknowledgment during the pandemic.

A critique of TPB is that it did not consider the effect of moral norms (Conner & Armitage, 1998). When the behavior has a moral dimension, recognition of MO and social pressure is justified (Ajzen, 1991; Beck & Ajzen, 1991). The literature revealed substantial evidence that MO influences intention negatively (Conner & Armitage, 1998; Whitley, 1998). Accordingly, the theory has been extended with the inclusion of MO in ethics-based studies (e.g., Hendy et al., 2021; Koay et al., 2021). Hence, we propose the following hypothesis:

H5: MO has a significant negative influence on plagiarism intention.

The current study has also accounted for the external indicators of Subjective Norms (SN) and Attitude (ATT). SN can be deemed as the functional combination of significant referents’ reactions to performing or not performing a behavior and one’s motivation to comply with referents’ expectations concerning a given behavior (Fishbein & Ajzen, 2015). As one’s obligation to what is right or wrong is related to MO, one can argue that MO contributes to the formation of subjective norms. In other words, as a normative ethical standard, MO can affect social approval of plagiarism negatively. Empirical evidence likewise corroborates this linkage in the context of digital piracy (e.g., Yoon, 2011). Hence, the current study assumes that;

H6: MO has a significant negative influence on subjective norms.

2.6. Justice

Justice (JST) is another deontological concept that reflects equity in similar cases and people according to fair rules (Reidenbach & Robin, 2013). Jung (2009) demonstrated that justice and fairness impacted the students’ judgments of the ICT-related ethical dilemmas. Yoon (2011) put forth that JST negatively influences subjective norms (Yoon, 2011). In the context of academic dishonesty, recent evidence has demonstrated that perceived school justice significantly mediates the relationship between academic performance and cheating behavior (Sabbagh, 2021). Comparably, the current study seeks to examine whether JST influences intention to plagiarism through subjective norms. Hence, the current study presumes that;

H7: JST has a significant negative influence on subjective norms.

2.7. Perceived deterrence

Based on the deterrence theory, perceived deterrence (PD) denotes the perceived certainty and severity of punishment, restraining individuals from doing an illegal behavior (Michaels & Miethe, 1989). Research on ICT ethics proved that perceived prosecution risks negatively affect attitudes to digital piracy (Akbulut & Dönmez, 2018). Besides, previous studies on plagiarism showed that students’ neutralizing attitudes to plagiarism are negatively influenced by the perceived likelihood of being detected (Hughes & McCabe, 2006). Furthermore, Chirikov et al. (2020) found that the perceived severity of penalties inflicted by faculty on plagiarism significantly and negatively influenced approving attitudes toward plagiarism among engineering students. Consequently, the current study presumes that;

H8: PD has a significant negative influence on plagiarism attitudes.

2.8. Goal orientation

In the literature, dishonest academic behaviors were also investigated from the lens of different achievement motivation theories. For example, studies manifested that students’ goal orientations were associated with cheating inclinations. More precisely, mastery-oriented students (i.e., having an intrinsic desire to learn) exhibit less tendency to involve in cheating behavior than those who are primarily performance-driven or ego-oriented, stimulated by external incentives such as earning good grades, escaping failure, and displaying performance (Murdock & Anderman, 2006). In the same vein, an up-to-date meta-analytic review of literature has explicated the significant negative relationships between cognitive determinants of academic dishonesty and personal mastery goal orientations compared to the significant positive relationships between cognitive determinants of academic dishonesty and extrinsic goal orientations (Krou et al., 2021). Thus, acknowledging that different personal motivational orientations such as mastery goal orientation (MGO) and extrinsic goal orientation (EGO) influence the way students approach classroom tasks, and consequently the way they perceive plagiarism, it is feasible to hypothesize that;

H9: MGO has a significant negative influence on plagiarism attitudes.

H10: EGO has a significant positive influence on plagiarism attitudes.

2.9. Academic self-efficacy

Bandura (1997) defines self-efficacy as “the beliefs in one’s capabilities to organize and execute courses of action required to produce given attainments” (p. 3). Academic self-efficacy (ASE) has been operationalized in terms of students’ perceived confidence in their skills and competencies to succeed in their academic performance (Leach et al., 2003). In conjunction with achievement goal orientations and contextual stimulations, ASE levels affect students’ engagement online courses delivered at a distance (Cho & Shen, 2013). Thereby, lower engagement may have implications for committing plagiarism (du Rocher, 2020).

The current study argues that academic self-efficacy influences attitudes to plagiarism, particularly in online learning environments. This is because, emergency online learning is linked with lower self-efficacy and cognitive engagement (Aguilera-Hermida et al., 2021). It was also verified that academic self-efficacy is significantly and positively correlated with negative attitudes towards plagiarism (du Rocher, 2020). Consequently, the current study posits that;

H11: Academic self-efficacy has a significant negative influence on plagiarism attitudes.

2.10. Fear of COVID-19

The pandemic crisis has triggered massive waves of anxiety, fear, depression, and uncertainty, which caused disruptions in behavior and psychological well-being (Ahorsu et al., 2020).Theories of planned behavior and reasoned action were criticized for relying immensely on the rational goal-oriented decision-making processes neglecting the effect of emotions on the behavior and attitude formation (Ajzen & Fishbein, 2005). Emotions have systematic effects on behavioral beliefs (Ajzen & Fishbein, 2005). Although previous studies in the context of academic dishonesty have not explicitly tackled the influence of emotional factors on attitudes and intention, it was found that emotional stability, as a personality trait, is linked with lower involvement in plagiarism (e.g., Karim et al., 2009; Steinberger et al., 2021). Furthermore, a new line of research has distinctly shown that negative emotions are correlated positively with favorable attitudes to plagiarism (Tindall & Curtis, 2020). Notwithstanding, the influence of Fear of COVID-19 (FC) has been rarely investigated in recent academic dishonesty literature. Thus, recognizing the negative emotions induced by the pandemic, the current study formulates the following hypothesis;

H12: FC has a significant positive influence on plagiarism attitudes.

3. Method

This study examines factors influencing plagiarism intentions among university students during the COVID19 pandemic. The current study applies a quantitative research design that employs a cross-sectional survey instrument for data collection and a partial least squares structural equation modeling technique for data analysis. The formulated model is depicted in Figure 1.

Figure 1. Research Model based on Extended Theory of Planned Behavior

3.1. Participants

The instrument was administered to a non-probabilistic convenience sample of undergraduate students. The sample was drawn from three universities in Oman. An e-survey form was sent out to undergraduate students via their institutional emails, and a total of 435 survey forms were received back. After an initial screening, 37 cases were dropped because of having too many missing or identical responses. the final dataset encompassed 398 valid responses. 277 (70 %) participants were females, and 121 (30 %) were males. Most of the students rated themselves as successful (n = 254, 64 %) students. Additionally, most of them (n = 230, 58 %) thought their Grade Point Average (GPA) improved in online courses delivered during the pandemic. Table 1 summarizes the demographic characteristics of the sample.

Table 1. Participants’ demographic profile

Variable	Characteristic	Frequency	Percentage
Gender	Female	277	69.6
	Male	121	30.4
	Total	398	100
Age	Under 18	6	1.5
	18 to 22	276	69.3
	23–27	59	14.8
	28–32	29	7.3
	33 and over	28	7.0
	Total	398	100
Grade level	Freshman	90	22.6
	Sophomore	110	27.6
	Junior	44	11.1
	Senior	62	15.6
	Other	92	23.1
	Total	398	100
Program of study	Diploma	64	16.1
	Bachelor	305	76.6
	Master	20	5.0
	Ph.D.	9	2.3
	Total	398	100
Employment status	Full-time job	75	18.8
	Part-time job	46	11.6
	Seeking opportunities	70	17.6
	Other	207	52.0
	Total	398	100
Performance rating	Not successful	20	5.0
	Successful	254	63.8
	Very successful	124	31.2
	Total	398	100
Grade Point Average (GPA) during the Pandemic	Improved	230	57.8
	Remained the same	83	20.9
	Suffered	85	21.4
	Total	398	100
Frequency of plagiarism during the pandemic	Never	127	31.9
	A few of the times	80	20.1
	About half of the times	116	29.1
	Almost every time	52	13.1
	Every time	23	5.8

3.2. Data collection and instrument measures

Data were collected in the fall semester of the 2020–2021 academic year. Ethical approvals were obtained from the institutions from which the sample was recruited. Afterward, participants gave their informed consent before voluntarily completing the survey.

The survey instrument consists of twelve dimensions, which have been employed from various pertinent instruments used in previous literature. For investigating the TPB variables (i.e., attitude, subjective norms, perceived behavioral control, past behavior, moral obligation, and intention), the current study has benefited from the Academic Integrity (AI) violation survey originally developed by Cronan et al. (2018) and adjusted by Uzun and Kilis (2020) to measure antecedents of intention to plagiarism using digital sources such as the Internet. The attitude was measured on a seven-point-semantic differential scale, whereas the other TPB constructs were measured on seven-point-Likert scales, respectively.

The TPB items adopted from Cronan et al. (2018); Uzun and Kilis (2020) were modified to comply with the current study’s context. More specifically, the intention, attitude, and subjective norms items were modified to reflect the primary focus of this study on plagiaristic behaviors using digital resources such as the Internet, mainly copying and pasting digital content into assignments without proper acknowledgment during online learning that had been predominant over the last couple of years. For example, based on the sources, i.e., Beck and Ajzen (1991) and Uzun and Kilis (2020), the items measuring subjective norms were adapted as follows;

SN1. If I copied others’ works from a digital source (e.g., the Internet) and used them in my assignments without acknowledgment, most people who are important to me would (not care/disapprove).

SN2. No one important to me thinks it is OK to copy others’ work from a digital source (e.g., the Internet) and use them in my assignments (agree/disagree).

SN3. Most people who are important to me will look down on me if I copy others’ works from a digital source (e.g., the Internet) and use them in my assignments without acknowledgment (likely/unlikely).

Justice was measured using two items, scored on a seven-point-Likert scale derived from Reidenbach and Robin (2013). Academic self-efficacy was assessed using eight items, rated on a seven-point-Likert scale, developed by Leach et al. (2003). Besides, the six mastery goal orientation items were adopted from Midgley et al. (1998), whereas the seven extrinsic goal orientation items were drawn from Anderman et al. (1998), each of which were scored on a seven-point-Likert scale. Moreover, perceived deterrence was measured using five items, rated on a five-point-Likert scale, developed by Michaels and Miethe (1989). Finally, the Fear of COVID-19 was measured using seven items scored on a five-point-Likert scale, and which were adopted from the Fear of COVID-19 Scale developed and validated by Ahorsu et al. (2020).

3.3. Data analyses

Prior to conducting the analyses, we utilized the G*Power software to adopt the power of 0.80, median 0.15 for the “intention” construct with five predictors. Accordingly, we obtained that the minimum sample size should be 92 cases for the study. Nevertheless, it is suggested to double (n = 184), or triple (n = 276) the calculated amount to have a more consistent model (Ringle et al., 2014). Given these results, the sample size of the current study was deemed to be sufficient for further analysis. Data were analyzed using partial least squares structural equation modeling (PLS-SEM) by employing a simultaneous two-phase analytical approach. First, we estimated and assessed the reflective measurement model understudy to establish validity and reliability. Second, we assessed the structural model by examining the variation inflation factor (VIF) to check statistical colinearity. We also reported the magnitude (β) and significance (p) of the path coefficients for testing the proposed hypotheses. We referred to the coefficient of determination R² values for explained variance, f² values for effect sizes and Q² values for the predictive relevance of the model (Hair et al., 2017).

4. Results

4.1. Measurement model

This section assessed the measurement model in terms of reliability and validity. Indicator reliability was evaluated by outer loadings. In the current study, the accepted item outer loading threshold value was chosen as 0.50 (Chin, 1998; Hulland, 1999). All outer loadings in the current study are within the allowable range. Nevertheless, four items, namely (MO1, EGO1, EGO2, and FC2), were excluded from further analyses due to having low factor loadings (<0.50).

Internal consistency was estimated through composite reliability (CR), Dijkstra-Henseler’s rho and Cronbach’s Alpha. The values were greater than the benchmark value (.70; Hair et al., 2017); therefore, internal consistency was ensured (Table 2). . For the Past Behavior (PB), it had a relatively low alpha value (0.514). However, Cronbach’s Alpha is regarded as a conservative measure and is likely to underestimate the actual value of internal consistency reliability (Hair et al., 2017). Therefore, the internal consistency of PB and all other constructs were further evaluated by checking Dijkstra-Henseler’s rho (ρa) and CR. Accordingly, all constructs have exhibited sufficient internal consistency as ρa and CR values are above 0.70. Finally, convergent validity was established as reflected by satisfactory Average Variance Extracted (AVE) values >.50 for all constructs (Hair et al., 2017).

Table 2. Results of the measurement model

Construct	Items	Outer loadings	Cronbach’s Alpha ɑ	rho_A (ρa)	CR (ρc)	AVE
Academic Self-Efficacy (ASE)			0.929	0.934	0.941	0.667
	ASE1	0.840
	ASE2	0.814
	ASE3	0.816
	ASE4	0.803
	ASE5	0.825
	ASE6	0.824
	ASE7	0.776
	ASE8	0.832
Attitude (ATT)			0.754	0.821	0.833	0.561
	ATT1	0.833
	ATT2	0.642
	ATT3	0.611
	ATT4	0.874
Perceived Deterrence (PD)			0.837	0.848	0.891	0.671
	PD1	0.762
	PD2	0.810
	PD3	0.853
	PD4	0.849
Extrinsic Goal Orientation (EGO)			0.742	0.827	0.849	0.655
	EGO3	0.895
	EGO4	0.818
	EGO5	0.704
Fear of Covid19 (FC)			0.894	0.989	0.907	0.624
	FC1	0.588
	FC3	0.847
	FC4	0.747
	FC5	0.690
	FC6	0.915
	FC7	0.899
Mastery Goal Orientation (MGO)			0.892	0.903	0.917	0.647
	MGO1	0.795
	MGO2	0.840
	MGO3	0.830
	MGO4	0.830
	MGO5	0.792
	MGO6	0.736
Intention (INT)			0.897	0.900	0.936	0.829
	INT1	0.922
	INT2	0.920
	INT3	0.889
Justice (JST)			0.890	0.891	0.948	0.901
	JST1	0.949
	JST2	0.950
Moral Obligation (MO)			0.823	0.825	0.919	0.850
	MO2	0.926
	MO3	0.918
Past Behavior (PB)			0.514	0.840	0.774	0.642
	PB1	0.953
	PB2	0.614
Perceived Behavioral Control (PBC)			0.891	0.901	0.920	0.696
	PBC1	0.800
	PBC2	0.855
	PBC3	0.862
	PBC4	0.865
	PBC5	0.787
Subjective Norms (SN)			0.730	0.735	0.847	0.648
	SN1	0.810
	SN2	0.785
	SN3	0.819

(AVE: Average Variance Extracted; CR: Composite Reliability; rho_A: Dijkstra-Henseler’s rho).

Discriminant validity was estimated through Fornell and Larcker’s (1981) criteria, cross-loadings, and HTMT values. According to Fornell and Larcker, the square root of an AVE value calculated for a construct should exceed the correlations between that construct and all other latent constructs. As shown in Table 3, this requirement was also satisfied. An examination of cross-loadings pointed out that all items loaded more substantially and significantly onto their respective factors than other factors, as demonstrated in Table 4.

Table 3. Results of discriminant validity

	ASE	ATT	PD	EGO	FC	MGO	INT	JST	MO	PB	PBC	SN
ASE	0.82
ATT	−0.46	0.75
PD	0.38	−0.30	0.82
EGO	−0.09	0.22	0.01	0.81
FC	−0.07	0.21	0.12	0.37	0.79
MGO	0.74	−0.39	0.47	−0.07	−0.04	0.80
INT	−0.53	0.48	−0.34	0.14	0.11	−0.44	0.91
JST	0.08	−0.40	0.10	−0.37	−0.33	0.16	−0.23	0.95
MO	0.13	−0.43	0.15	−0.36	−0.32	0.20	−0.23	0.79	0.92
PB	−0.52	0.48	−0.35	0.30	0.14	−0.48	0.58	−0.32	−0.33	0.80
PBC	−0.42	0.39	−0.38	0.07	−0.12	−0.38	0.47	0.03	0.04	0.44	0.83
SN	0.04	0.22	−0.05	0.23	0.24	−0.03	−0.06	−0.49	−0.49	0.09	−0.25	0.80

Note. Fornell and Larcker criteria. Bold diagonal values are the square root of AVE. Below diagonal values are bivariate correlations. ASE: Academic Self-Efficacy; ATT: Attitude; PD: Deterrence Measures; EGO: Extrinsic Goal Orientation; FC: Fear of COVID-19; MGO: Mastery Goal Orientation; INT: Intention; JST: Justice; MO: Moral Obligation; PB: Past Behavior; PBC: Perceived Behavioral Control; SN: Subjective Norms

Table 4. Results of discriminant validity: Factor Structure Matrix of Loadings and Cross-loadings

	ASE	ATT	DTM	EM	FC	IM	INT	JST	MO	PB	PBC	SN
ASE1	0.840	−0.446	0.325	−0.063	0.033	0.595	−0.456	0.063	0.101	−0.484	−0.390	0.069
ASE2	0.814	−0.379	0.275	−0.059	−0.059	0.583	−0.462	0.076	0.103	−0.393	−0.327	0.043
ASE3	0.816	−0.399	0.277	−0.087	−0.100	0.538	−0.445	0.058	0.109	−0.423	−0.378	0.070
ASE4	0.803	−0.349	0.281	−0.024	0.005	0.559	−0.410	0.012	0.027	−0.365	−0.301	0.016
ASE5	0.825	−0.343	0.272	−0.075	−0.059	0.654	−0.420	0.094	0.112	−0.443	−0.324	0.043
ASE6	0.824	−0.319	0.335	−0.078	−0.068	0.621	−0.382	0.082	0.114	−0.406	−0.281	−0.009
ASE7	0.776	−0.314	0.359	−0.060	−0.059	0.620	−0.409	0.073	0.085	−0.379	−0.360	−0.026
ASE8	0.832	−0.405	0.350	−0.131	−0.126	0.683	−0.468	0.095	0.163	−0.478	−0.385	0.015
ATT1	−0.455	0.833	−0.268	0.120	0.157	−0.353	0.480	−0.182	−0.216	0.393	0.430	0.048
ATT2	−0.211	0.642	−0.184	0.252	0.172	−0.244	0.188	−0.460	−0.491	0.287	0.057	0.393
ATT3	−0.136	0.611	−0.147	0.217	0.189	−0.170	0.149	−0.460	−0.468	0.231	0.054	0.362
ATT4	−0.433	0.874	−0.266	0.160	0.147	−0.354	0.473	−0.291	−0.302	0.473	0.417	0.096
PD1	0.233	−0.199	0.762	−0.027	0.009	0.351	−0.219	0.066	0.109	−0.258	−0.319	−0.053
PD2	0.293	−0.239	0.810	0.021	0.030	0.378	−0.264	0.076	0.096	−0.267	−0.305	−0.015
PD3	0.349	−0.266	0.853	0.011	0.176	0.368	−0.331	0.109	0.159	−0.289	−0.329	−0.051
PD4	0.347	−0.272	0.849	0.028	0.136	0.432	−0.288	0.073	0.119	−0.334	−0.301	−0.034
EGO3	−0.147	0.231	−0.025	0.895	0.353	−0.129	0.194	−0.358	−0.359	0.327	0.090	0.189
EGO4	−0.056	0.149	0.015	0.818	0.260	−0.096	0.058	−0.296	−0.271	0.219	0.058	0.199
EGO5	0.037	0.130	0.068	0.704	0.270	0.102	0.055	−0.216	−0.211	0.122	−0.012	0.189
FC1	0.133	0.014	0.194	0.172	0.588	0.198	−0.033	−0.114	−0.155	−0.093	−0.116	0.111
FC3	−0.043	0.171	0.106	0.312	0.847	−0.006	0.072	−0.282	−0.282	0.111	−0.129	0.234
FC4	0.043	0.104	0.079	0.265	0.747	0.073	0.011	−0.187	−0.199	0.026	−0.096	0.194
FC5	0.102	0.021	0.197	0.225	0.690	0.146	−0.071	−0.180	−0.144	−0.080	−0.179	0.117
FC6	−0.137	0.246	0.064	0.356	0.915	−0.139	0.176	−0.357	−0.344	0.209	−0.075	0.208
FC7	−0.045	0.160	0.138	0.323	0.899	0.000	0.068	−0.268	−0.251	0.096	−0.098	0.182
MGO1	0.594	−0.373	0.392	−0.052	−0.073	0.795	−0.345	0.210	0.218	−0.465	−0.243	−0.104
MGO2	0.627	−0.331	0.413	−0.079	−0.066	0.840	−0.426	0.125	0.140	−0.402	−0.351	0.017
MGO3	0.607	−0.311	0.360	−0.074	0.012	0.830	−0.371	0.110	0.138	−0.383	−0.307	0.006
MGO4	0.629	−0.354	0.354	−0.093	−0.081	0.830	−0.399	0.172	0.223	−0.406	−0.313	−0.027
MGO5	0.571	−0.270	0.398	−0.026	0.043	0.792	−0.283	0.083	0.150	−0.342	−0.307	−0.025
MGO6	0.539	−0.218	0.339	−0.012	0.042	0.736	−0.250	0.029	0.060	−0.282	−0.352	0.017
INT1	−0.473	0.431	−0.307	0.106	0.079	−0.378	0.922	−0.199	−0.206	0.545	0.462	−0.073
INT2	−0.511	0.469	−0.333	0.132	0.118	−0.440	0.920	−0.233	−0.214	0.525	0.431	−0.072
INT3	−0.466	0.417	−0.288	0.160	0.096	−0.377	0.889	−0.201	−0.214	0.504	0.401	−0.007
JST1	0.069	−0.400	0.085	−0.338	−0.315	0.134	−0.214	0.949	0.761	−0.301	0.029	−0.459
JST2	0.092	−0.355	0.104	−0.362	−0.315	0.173	−0.226	0.950	0.747	−0.303	0.021	−0.465
MO2	0.119	−0.370	0.127	−0.299	−0.296	0.166	−0.200	0.692	0.926	−0.287	0.057	−0.470
MO3	0.113	−0.416	0.148	−0.365	−0.301	0.207	−0.227	0.774	0.918	−0.319	0.018	−0.428
PB1	−0.543	0.441	−0.363	0.197	0.066	−0.476	0.597	−0.180	−0.198	0.953	0.489	−0.006
PB2	−0.194	0.347	−0.145	0.400	0.259	−0.251	0.229	−0.517	−0.501	0.614	0.087	0.289
PBC1	−0.291	0.252	−0.273	0.019	−0.164	−0.280	0.354	0.087	0.088	0.323	0.800	−0.332
PBC2	−0.460	0.430	−0.346	0.096	−0.030	−0.406	0.467	−0.025	−0.014	0.423	0.855	−0.155
PBC3	−0.347	0.326	−0.335	0.033	−0.075	−0.301	0.425	0.010	0.006	0.361	0.862	−0.194
PBC4	−0.352	0.338	−0.318	0.079	−0.130	−0.287	0.351	0.052	0.088	0.369	0.865	−0.210
PBC5	−0.288	0.245	−0.308	0.042	−0.114	−0.297	0.358	0.005	0.027	0.344	0.787	−0.188
SN1	−0.006	0.236	−0.031	0.170	0.176	−0.087	−0.044	−0.414	−0.386	0.113	−0.161	0.810
SN2	0.036	0.105	−0.003	0.177	0.198	0.000	−0.050	−0.343	−0.343	0.027	−0.234	0.785
SN3	0.060	0.185	−0.069	0.213	0.196	0.017	−0.044	−0.411	−0.440	0.068	−0.220	0.819

Note. ASE: Academic Self-Efficacy; ATT: Attitude; PD: Deterrence Measures; EGO: Extrinsic Goal Orientation; FC: Fear of COVID-19; MGO: Mastery Goal Orientation; INT: Intention; JST: Justice; MO: Moral Obligation; PB: Past Behavior; PBC: Perceived Behavioral Control; SN: Subjective Norms

Discriminant validity was further evaluated through Heterotrait-Monotrait Ratio (HTMT) by testing the null hypothesis (H0: HTMT ≥ 1) against the alternative hypothesis (H1: HTMT < 1) using a complete bootstrapping procedure with 5000 subsamples. Results demonstrated that confidence intervals were lower than one in all cases, showing that discriminant validity was achieved (Henseler et al., 2015). In addition, hypotheses testing has shown that none of the bias-corrected confidence intervals employed (i.e., 2.5% and 97.5%) incorporate the value 1, which further strengthens the discriminant validity of the measures used in the study (Hair et al., 2017). To conclude, testing the measurement model yielded satisfactory reliability and validity values. Therefore, we proceeded to the assessment of the structural model and hypotheses testing.

4.2. Structural model

A complete bootstrapping method with iterating 5000 subsamples and Bias-Corrected and Accelerated (BCa) Bootstrap confidence interval at 0.05 significance level was applied to test the formulated hypotheses. Furthermore, inner and outer Collinearity Statistics, as manifested by the variance inflation factor values (VIF), were below the critical threshold of 5, indicating that the structural model under study did not have any collinearity issues. Results of testing the structural model are illustrated in Figure 2 and Table 5.

Figure 2. Results of the Structural Model. Note **p < .01; **p < .00

Table 5. Results of the structural model: hypotheses testing

Hypothesis	Path	β	t	P values	ƒ^2	VIF	Confidence Intervals Bias Corrected		Decision
							2.5%	97.5%
H1	ATT ➜ INT	0.219	3.532	***	0.052	1.661	0.103	0.347	Supported
H2	SN ➜ INT	−0.132	2.540	*	0.021	1.449	−0.233	−0.031	Supported
H3	PBC ➜ INT	0.199	3.735	***	0.045	1.584	0.089	0.298	Supported
H4	PB ➜ INT	0.365	6.429	***	0.153	1.544	0.257	0.477	Supported
H5	MO ➜ INT	−0.091	1.807	0.071	0.009	1.623	−0.191	0.005	Rejected
H6	MO ➜ SN	−0.273	3.002	**	0.037	2.707	−0.455	−0.096	Supported
H7	JST ➜ SN	−0.270	2.858	**	0.036	2.707	−0.445	−0.077	Supported
H8	PD ➜ ATT	−0.170	2.785	**	0.031	1.314	−0.287	−0.048	Supported
H9	MGO ➜ ATT	−0.053	0.689	0.491	0.002	2.444	−0.209	0.093	Rejected
H10	EGO ➜ ATT	0.132	2.876	**	0.021	1.165	0.038	0.217	Supported
H11	ASE ➜ ATT	−0.333	4.836	***	0.069	2.237	−0.462	−0.197	Supported
H12	FC ➜ ATT	0.154	3.311	**	0.028	1.185	0.052	0.236	Supported

Note. β: Standardized path coefficient; t: t-statistics; ƒ²: effect size, VIF: variance inflation factor, ASE: Academic Self-Efficacy; ATT: Attitude; PD: Deterrence Measures; EGO: Extrinsic Goal Orientation; FC: Fear of COVID-19; MGO: Mastery Goal Orientation; INT: Intention; JST: Justice; MO: Moral Obligation; PB: Past Behavior; PBC: Perceived Behavioral Control; SN: Subjective Norms; *p < .05; **p < .01; ***p < .001

As Table 5 exhibits, all proposed hypotheses were accepted except for the influence of moral obligation on intention and mastery goal orientation on attitude. Plagiarism intention was significantly explained by attitude (β = 0.22, p < .001), subjective norms (β = −0.13, p < .01), perceived behavioral control (β = 0.20, p < .001), and past behavior (β = 0.37, p < .001). Subjective norms were significantly predicted by moral obligation (β = −0.27, p < .01) and justice (β = −0.27, p < .01). Finally, attitude towards plagiarism was significantly predicted by perceived deterrence (β = −0.17, p < .01), extrinsic goal orientation (β = 0.13, p < .01), academic self-efficacy (β = −0.33, p < .001), and fear of COVID19 (β = 0.15, p < .001). The R² value for the whole model was calculated as 0.439 (see, Table 6) reflecting that the model accounted for 44 % of the variance in the intention. According to Chin (1998), R² values of 0.67, 0.33, and 0.19 stand for substantial, moderate, and weak models. Hence, the explanatory power of the structural model understudy is relatively substantial in predicting variance in plagiarism intention experienced during COVID 19 pandemic.

Table 6. Predictive relevance and explained variance

Constructs	Predictive Relevance Q^2		Coefficient of determination R^2
	Construct Cross validated Communality	Construct Cross validated Redundancy
ASE	0.566
ATT	0.303	0.148	0.285
PD	0.449
EGO	0.326
FC	0.495
MGO	0.505
INT	0.619	0.355	0.439
JST	0.575
MO	0.465
PB	0.100
PBC	0.541
SN	0.297	0.164	0.264

Note. Q²: Predictive relevance; R²: Coefficient of determination ASE: Academic Self-Efficacy; ATT: Attitude; PD: Deterrence Measures; EGO: Extrinsic Goal Orientation; FC: Fear of COVID-19; MGO: Mastery Goal Orientation; INT: Intention; JST: Justice; MO: Moral Obligation; PB: Past Behavior; PBC: Perceived Behavioral Control; SN: Subjective Norms;

In the current study, the indirect effects were also scrutinized. It was found that except for moral obligation and mastery goal orientation, all external antecedents significantly and indirectly predicted plagiarism intention as partially mediated by attitude and subjective norms (see, Table 7).

Table 7. Indirect effects

Path	β	t	p values
ASE-> ATT-> INT	−0.073	2.625	0.009**
PD-> ATT-> INT	−0.037	2.087	0.037*
EGO-> ATT-> INT	0.029	2.201	0.028*
FC-> ATT-> INT	0.034	2.267	0.023*
MGO-> ATT-> INT	−0.012	0.674	0.500
JST-> SN-> INT	0.035	2.062	0.039*
MO-> SN-> INT	0.036	1.673	0.094

Note. β: Standardized path coefficient; t: t-statistics, p*< .05; **p < .01

Finally, relevance of path coefficients was ascertained by investigating the effect sizes (f²) and predictive relevance (Q²) as reported in Table 5 and Table 6, respectively. Effect sizes (f²) values of .02, .15, and .35 denote small, moderate, and strong effects (Cohen, 1988). Accordingly, past behavior had a moderate effect size (f² = .15) on intention, whereas academic self-efficacy had a small to moderate effect size on attitude (f² = .07; see, Table 5). Blindfolding procedure was conducted with an omission distance of 7 to calculate the blindfolding-based cross-validated redundancy measure Q². Hence, the structural model was shown to have an acceptable predictive relevance as it met the criteria of having Q² values greater than 0 (Hair et al., 2017; see, Table 6).

5. Discussion

This study investigated factors determining undergraduate students’ behavioral intentions to commit plagiarism via digital sources during the COVID19 pandemic. To this end, a research model was constructed based on the extended TPB, which was reliable and valid. Furthermore, the majority of the hypotheses were supported.

As far as TPB variables are concerned, the attitude was a significant determinant of intention to execute plagiarism. Consequently, the more favorable attitudes toward plagiarism students have, the more inclined they will be to plagiarize. Such a result comes in line with the mainstream relevant literature (e.g., Parks-Leduc et al., 2021; Tremayne & Curtis, 2021). Attitudes are shaped by the underlying salient beliefs (Conner & Armitage, 1998). Hence, lenient attitudes toward plagiarism reveal that undergraduate students might neutralize the offense of plagiarism, believing that plagiarism is permissible (Abbasi et al., 2021; Al-Nuaimi et al., 2021), especially under compelling circumstances, just as the social isolation restrictions during the COVID lockdown.

Subjective norms negatively and significantly determined intention to plagiarize, suggesting that normative beliefs of significant referents, including parents and peers, constitute a decisive determinant of plagiarism. Although this finding contradicts some studies (e.g., Maloshonok & Shmeleva, 2019; Pham et al., 2021), it is compatible with others (e.g., Eret & Ok, 2014). Furthermore, it is also parallel that Oman is a collectivistic culture where conformity of norms is essential. On top of that, plagiarists practice social projection of their attitudes onto their peers, leading them to accept plagiarism (Moss et al., 2018).

As for the effect of perceived behavioral control on intention, a significant and positive influence was proved, consistent with previous research (e.g., Hendy & Montargot, 2019; Yang et al., 2021). In other words, the perceived ease of performing digital plagiarism increases the likelihood of committing it in the future. The COVID-19 pandemic has widened digital divides among young individuals regarding access and use of the Internet and other technologies. While some students have already had digital resources or opportunities to pursue online education, some of them have not (Williamson et al., 2020). Given that having digital resources and abilities might provoke plagiarism (Jereb et al., 2018), it is reasonable to attain such a result.

Contrary to expectations, moral obligation did not significantly affect intention, implying that students are prone to plagiarize irrespective of whether they consider plagiarism immoral. Conversely, previous studies revealed a significant effect of moral obligation on the intention to plagiarize (e.g., Lin & Clark, 2021). The non-significant effect of moral obligation contradicts the proponents’ evidence for incorporating moral obligation into the TPB for explaining unethical behavior. This unexpected result can be attributed to the function of presumably mediating variables (e.g., moral disengagement) that interfere with the association between moral obligation and intention to commit plagiarism. For example, it has been found that moral disengagement mechanisms mediate the relationship between moral judgment and decision-making regarding academic dishonesty (Stephens, 2018). To further illustrate this, moral disengagement mechanisms mitigate the influence of moral obligation on dishonest academic actions, suggesting a judgment-action gap concerning academic dishonesty. Stephens (2018) contends that moral obligation is only one component among other multiple components of moral functioning.

The past behavior was the most conclusive determinant of intention to plagiarize, which is congruent with the previous literature on academic dishonesty (e.g., Cronan et al., 2018; Uzun & Kilis, 2020). This result insinuates that individuals who had previously plagiarized are at a high risk of perpetrating plagiarism in the future. The current study does not endorse past behavior’s habitual perspective, which argues that automatic behavior production relies on well-established routines. Instead, we evaluate past behavior from the reasoned action perspective (Ajzen, 2002). We believe that students are mainly involved in plagiarism with conscious deliberation in a new unstable emergency online learning context dictated by the COVID lockdown. Therefore, we postulate that the role of past behavior in predicting later behavior is not independent of the intention and other cognitive TPB variables for this context.

Furthermore, successful past behavior repetition intensifies perceived behavioral control over the behavior and reinforces favorable attitudes toward plagiarism. Furthermore, plagiarists might fallaciously assume that their peers and significant referents approve plagiarism, jeopardizing the impact of injunctive norms. Moreover, concurrently, past behavior might diminish the influence of moral judgment on intention (Ajzen, 1991; Conner & Armitage, 1998; Moss et al., 2018; Whitley, 1998). Above all, we argue that the strength of past behavior implies that students can execute plagiarism once similar conditions occur.

This study has also investigated external predictors of the critical TPB variables (i.e., attitude and subjective norms). When analyzing the antecedents of attitude towards plagiarism, it was found that extrinsic goal orientation significantly and positively impacted attitude towards plagiarism, which is in good agreement with previous research on extrinsic goal orientations and academic dishonesty beliefs and behaviors (e.g Jordan, 2001; Krou et al., 2021). When students pursue or are under pressure to get high grades by any possible means neglecting the costs, they foster a positive attitude to plagiarism. Therefore, adopting extrinsic goal orientations would drive students to perceive plagiarism as a beneficial tool to maximize their scores (Anderman & Midgley, 2004).

The current study showed that mastery goal orientations were not significantly associated with the attitude, albeit negative. This result is harmonious with previous studies demonstrating mastery goal orientations either negatively or not associated with academic dishonesty (e.g., Anderman et al., 1998; Jordan, 2001). Unlike students with extrinsic goal orientations, those with mastery goal orientations would have no motivation to cheat because they believe cheating would not help them master learning or acquire and construct knowledge (Stephens & Gehlbach, 2007).

Academic self-efficacy was proven to be the most robust antecedent of attitudes than the other variables examined in this study. Specifically, it negatively influenced plagiarism attitude, which corresponds well with the previous literature on plagiarism (du Rocher, 2020) and cheating (Murdock & Anderman, 2006) That is because academic self-efficacy is a significant predictor of cognitive engagement in online learning environments, as recently confirmed (Aguilera-Hermida et al., 2021). In its turn, cognitive engagement negatively predicts favorable attitudes to plagiarism (du Rocher, 2020). Besides, it is reasonable to assume that the COVID lockdown atmosphere might have served as a source of low self-efficacy cues for the students. According to Bandura (1986), negative emotional arousal weakens performance by diminishing success expectations and constraining individuals’ efforts to cope with difficult situations. By contrast, according to the Neuropsychological Theory of Positive Affect, positive emotions are linked with higher brain dopamine levels, which increase cognitive flexibility and yield performance improvement on various cognitive tasks (Ashby et al., 1999). Concordantly, students with increased cognitive flexibility might have utilized more active learning strategies and thereby possessed less favorable attitudes towards plagiarism (du Rocher, 2020).

One of the essential contributions of the current study is examining the effect of COVID19 fear on the attitudes towards plagiarism. Our results reveal that COVID19 fear relates significantly and positively to attitudes towards plagiarism. Such results indicate that students would be inclined to hold positive plagiarism convictions when feeling intense levels of COVID19 anxiety, which fits well with the limited research on the topic (e.g., Tindall & Curtis, 2020).

Perceived deterrence negatively and significantly influenced attitude towards plagiarism, implying that the more students perceive the probability of being detected and punished and the perceived punishment severity, the less they accept plagiarism. This finding is in line with the literature on academic dishonesty, proposing that cheating costs, perceived seriousness of plagiarism are inversely related to cheating propensity (Michaels & Miethe, 1989). Similarly, recent studies have revealed that students who do not sufficiently understand their university plagiarism policies tend to exhibit positive attitudes toward plagiarism and, hence, are more willing to plagiarize (e.g., Amida et al., 2021).

A further contribution of this study is investigating normative ethics antecedents of subjective norms in the context of academic dishonesty. Based on justice theory and deontological ethics, the current study’s findings explicate that both justice and moral obligation significantly and negatively influenced subjective norms towards not engaging in plagiarism. The indirect influence of justice on intention confirms Jung’s (2009) findings, indicating that justice affects students’ ethical judgments in ICT-related ethical dilemmas.

Our study has found a significant negative influence of moral obligation on subjective norms, in line with Fishbein and Ajzen (2015), asserting that moral obligation contributes to forming normative standards. Alternatively put, students with a high moral obligation to avoid plagiarism project their moral convictions on plagiarism-related subjective normative beliefs. Taken together, the effects of justice and moral obligation on intentions imply that the more students deem plagiarism as an unfair and unethical behavior, the less they will believe that their significant referents approve of engaging in plagiarism, which coincides with the integrated ethical model proposed by Yoon (2011).

6. The theoretical contributions of the study

The current study contributes to the advancement of theory in ICT ethics, particularly concerning academic dishonesty in two ways. First, the current study incorporated an emotional variable (i.e., COVID-19 fear) to the integrated TPB model by exploring its effect on plagiarism attitude. Second, the current study has provided a comprehensive view of the impact of motivational variables on attitudes to plagiarism. In this regard, the current study incorporated an academic motivational framework into an extended TPB model similar to Murdock and Anderman (2006).

7. Implications for practice and pedagogy

The current study proposes practical implications for addressing academic integrity issues arising in online learning mediums during and after the pandemic.

Various technological solutions can be applied in this regard. Higher education institutions may employ interventions like proctoring, plagiarism detection software, and the use of time-stamps on assignment submissions (Sullivan, 2016). However, technological solutions alone may not remediate academic dishonesty’s social responsibility-related problems (Reedy et al., 2021). Hence, it is suggested to integrate psychosocial methods, ethical education, and technological solutions to combat plagiarism during and post the pandemic.

Based on the current study results, past behavior has been the most influential factor on intention. Since students may have engaged in plagiarism during high school, it is essential to minimize the effects of past plagiarism behavior starting from their foundational years at college. For example, one way to achieve this goal is to provide students with explicit information literacy instructional components within university academic compulsory courses to promote their information literacy skills needed to avoid plagiarism, e.g., standard referencing and proper paraphrasing skills. Such explicit information literacy instruction may assist students to form firm attitudes against plagiarism, and strong intentions to avoid plagiarism. As a result, the residual effects of past plagiarism behaviors students may have committed during their pre-college education will vanish. On the other hand, the heavy reliance on educational platforms, learning management systems (LMS), and videoconferencing applications during and post the pandemic would make it easier for faculty to integrate information literacy instruction using interactive tools such as forums, wikis, surveys, and announcements.

Since students’ access to digital sources cannot be restricted, students should be taught how to make informed use of digital information and sources ethically and critically. In the meantime, assessments should endorse high-order thinking skills, higher self-regulation, and metacognitive skills (Reedy et al., 2021).

A wide variety of pedagogical practices are recommended to enhance and maintain students’ mastery goal orientation, including online collaborative learning experiences, flipped online classrooms, cultivating students’ autonomy in the learning process, providing students with constant sufficient feedback, and communicating explicit guidelines, deadlines, rules of academic conduct, ethical and achievement expectations (Zaccoletti et al., 2020).

8. Recommendations for future research

Multiple directions for future research emanating from this study may enrich the academic dishonesty research. For example, comparative cross-national studies may examine the impact of perceived safety against the pandemic on plagiarism attitudes compared to the influence of corona-fear. Since the pandemic containment measures have immensely altered assessment and instructional methods, future studies may investigate the impact of these new contextual variables on plagiarism cognitions. For example, cognitive engagement in online learning environments, awareness of technical solutions detecting plagiarism, proctoring measures, and ethical conduct codes are potential variables for further examination.

9. The study limitations

The current study exhibits a few methodological limitations that might constrain the generalizability of its findings. For example, the present study has relied mainly on a survey instrument for data collection. However, as asserted by Cronan et al. (2018), given that the current study’s findings are consistent with the previous related empirical literature, self-reporting and social desirability biases associated with survey data are not problematic. Furthermore, the current study has recruited a non-probability convenience sample, which might not represent the target population thoroughly compared to random probability methods. Nevertheless, based on the non-parametric assumptions on sampling approved by the Partial Least Squares (PLS) approach to SEM employed in the current study, the shortcomings of convenience sampling are controlled.

Compliance with Ethical Standards

Ethical approvals were obtained from the institutions in which the sample was recruited. Informed consent was obtained from all participating students.

Availability of data and materials

The datasets analyzed in the current study are available from the corresponding author upon reasonable request.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The authors have no funding to report.