Curbing Bribe-Giving in Malaysia: The Role of Attitudes and Parents

Abstract

One of the main challenges developing nations face is curbing bribery. While there are many efforts to curb bribery, most focused at macro level, such as law, while little has been examined at the micro level, e.g., individual behavior and intention. Those who did investigate at the micro level tend to focus on the recipients rather than the ones giving the bribe. We explored eight factors that influence Malaysian young adults’ bribe giving intention based on the Reasoned Action Approach (RAA). A total of 345 respondents (M_age = 20.68, SD = 2.01, 189 are females) completed questionnaires about all RAA variables. Partial Least Squares Structural Equation Modelling (PLS-SEM) was carried out using smartPLS3.0 to analyze the result. The result revealed that out of the eight variables, four variables—Instrumental attitudes, Experiential Attitudes, Parents’ descriptive norms and Capacity—explain 74% of the variance in bribe giving intention. An important take-away is that young adult’s perception of whether their parents gave or did not give bribes in a given situation is important in influencing their bribe giving intention. Bribe giving prevention messages must be targeted explicitly toward parents, where they play a crucial role in curbing this dishonesty.

Keywords:

• Attitudes
• bribe-giving
• corruption
• dishonesty
• parents
• young adult

Bribery is a type of corruption that many developing nations face and this dishonest act prevents all sorts of healthy developments (Hübler et al., 2021; Mengzhen et al., 2021). Bribery is estimated to be five times more likely in low-income countries (United Nation, 2021) compared to high-income countries. The economic inequalities and the need to outperform and remain competitive with limited resources created opportunities for people to use bribery as an easy and effective solution to the problem (Alatas, 1991).

The most commonly known bribe scenario would be a car driver giving money to the traffic police to overlook their unlawful behavior (Sekaran, 2020). Bribery is defined as the abuse or misuse of power or trust in a quid pro quo exchange (Robertson, 2017). Using the example of the scenario, the quid pro quo exchange happened as there are two parties that work together; the giver, driver who pays the bribe and the receiver, traffic police who receives/asks for the bribe. It is not uncommon for news and research in ASEAN to focus heavily on bribe receivers (Case, 2008; Kapeli & Mohamed, 2015, 2019). For example, the news reports often focus on corrupt politicians (Ngui & Wright, 2016), and other authority figures despite the smaller amounts, such as traffic police (Sekaran, 2020). Yet, the givers are often left out in the reports and it is their action too that contributes to this social problem. Ariely et al. (2013) found that the minor dishonest acts done by the thousands were found to be substantially more detrimental to society collectively. This implies that the effect of giving bribes, regardless of its amount, can be detrimental to society especially if it was done by thousands.

It is a common perception that penalizing corrupt individuals is the most effective way to curb corruption. For example, if found guilty, a person could be imprisoned for a term not exceeding 20 years in Malaysia (Satar, 2021). However, such macro-level effort has proven to be ineffective (Ariely et al., 2013; Muhamad & Gani, 2020). To complement the effort from marco-level, past studies (Othman et al., 2014, Budiman et al., 2013, Durairaja et al., 2019; Zaloznaya, 2014) have suggested looking into the micro-level determinants of corruption behavior. The present study intends to add to the lacuna in literature focusing on exploration of micro-level of bribe giving intentions—in a developing nation by bridging two important gaps by focusing on (i) bribe-giver and (ii) the micro-level factors. This research is also an effort to work toward the United Nation Sustainable Development Goals (SDG) 16.5, which aims to substantially reduce corruption and bribery in all their forms (UN, 2021).

Theoretical background and hypotheses development

According to Fishbein and Ajzen (2010), we must first form an intention before carrying out any action. Regardless of any types of behavior, there are three fundamental factors that would influence the formation of intention which are (i) attitudes, (ii) perceived norms, (iii) perceived behavioral control. This model is named the Theory of Planned Behavior (TPB). Although Pozsgai-Alvarez (2015) and Rabl and Kühlmann (2008) have used the same model to predict corrupt action and bribery, we would like to propose three reasons why this study will be carried out using the similar theory. Firstly, we must acknowledge the possibility of cultural differences between Germany, Latin America and South East Asia (ASEAN), the two studies cannot be fully generalized in the context of ASEAN. Secondly, this study focuses specifically on bribe giving intention. Thirdly, it is about the updated version of the theory which will be explained in the following paragraph.

However, Fishbein and Ajzen (2010) again proposed the updated version of TPB, the Reason Action Approach (RAA) model. The RAA model separated each of the main variables in TPB into two sub variables. Compared to TPB, this separation in the RAA model enhances our micro level understanding toward each variable (Fishbein & Ajzen, 2010) and allows us to predict actual behavior more accurately (McEachan et al., 2016). Hence, this study incorporates RAA as its framework. There are six factors that can influence intention according to RAA, (i) instrumental attitudes, (ii) experiential attitudes, (iii) injunctive norms, (iv) descriptive norms, (v) autonomy, and (vi) capacity (Fishbein & Ajzen, 2010).

Attitudes

The RAA model suggests that individuals would always consider their own attitudes toward a certain behavior before carrying it out. There are two types of attitudes; instrumental attitudes (IA) and experiential attitudes (EA). IA is formed when one considers the consequences of performing the behavior. When one thinks that the behavior will be followed by positive evaluation, a positive IA is formed and vice versa. EA refers to the emotions that the behavior will elicit. When one thinks that the behavior will make them feel good, a positive EA is formed (Fishbein & Ajzen, 2010).

IA toward corruption were found to be a significant predictor of corrupt behavior (Simpser, 2020). Previous research findings also show that individuals with favorable attitudes toward corruption were less likely to vote corrupt politicians out of their position (Chang & Kerr, 2017). Moldovan and Van de Walle (2013) found that although individuals hold unfavorable attitudes toward bribery in the healthcare setting would still offer gifts to the medical staff for fear that their non-conforming acts would result in losing the service that they deserved. However, the people who offer gifts may not see such acts as bribery. Therefore, it is possible for them to hold unfavorable attitudes toward bribery but still perform gift-giving as a way to conform to the norm in the health care setting. EA were also found to strongly correlate with intentions and behaviors. For instance, intentions against sexual assaults significantly increases as one’s EA toward it grows more positive (Reynolds-Tylus et al., 2019). EA was also an important predictor of one’ support toward physical distancing during pandemic (Wang, 2021). Hence, the proposed hypotheses are:

H1. Instrumental attitudes (IA) toward bribery is positively related to young adult’s bribe giving intention

H2. Experiential attitudes (EA) toward bribery is positively related to young adult’s bribe giving intention

Perceived norms

It is common for individuals to consider approval from people deemed to be important, or the likelihood of them acting the same in similar situations. For young adults, close friends and parents are considered as important people who could influence their intentions. Close friends have been shown to greatly influence student’s decision to be involved in corrupt behaviors in the academic settings (Tagoe, 2017; Kam et al., 2018). Parents were also found to be a significant predictor of their children’s corrupt intentions and behaviors (Amelia et al., 2020). In this study, we would like to identify which sources of perceived norms have influences over young adult’s bribe giving intention. Therefore, the hypotheses formed will focus on the effect close friends and parents have on bribe giving intention separately.

Perceived norms can further be categorized into two; descriptive norms (DN) and injunctive norms (IN). DN are the perceptions that their important ones are or are not giving bribes in a given bribe situation (Reno et al., 1993), whereas IN are the individuals’ perceptions concerning acceptability of the important ones with respect to giving bribes (Cialdini et al., 1990). Research shows that DN plays an important role in predicting corrupt behaviors. For instance, corrupt intention among Chinese students was found to be significantly influenced by the DN in the university (Zhao et al., 2019). In addition, in an experiment where DN were designed to reflect a reduction in the corrupted practices of a South Africa town, participants’ willingness to accept and give bribes was reduced (Kobis et al., 2019). Studies that focus on developing nations also found that DN does not only have high correlation with corrupt behavior (Abbink et al., 2018; Greppin et al., 2017), it also has a causal effect on anti-corruption behaviors (Köbis et al., 2015). This may be because people tend to act according to the descriptive norms in the society to not feel deviated (Crozier & Taylor, 2019). However, it is important to note that Lan and Hong (2017) found that the effect of descriptive norms on corrupt behavior is more normative in some societies than others. Hence, the proposed hypotheses are:

H3. Descriptive norms toward bribery for close friends (DNF) is positively related to bribe giving intention

H4. Descriptive norms toward bribery for parents (DNP) is positively related to bribe giving intention

IN were found to significantly influence one’s decision to be involved in corruption practices (Altenburger, 2017; Greppin et al., 2017). For instance, in the organizational setting, when one thinks that their colleagues would not approve of their decision to be involved in bribery, they are less likely to do so (Gorsira et al., 2018). Besides, IN were found to be effective in phrasing anti-corruption messages to the public, as it was able to provide people with new perspectives regarding their existing beliefs, and shift their beliefs toward anti-corruption (Agerberg, 2019). Legros and Cislaghi (2020) suggested IN were effective in curbing corruption in Mexico. Hence, based on the previous literature regarding IN, the proposed hypotheses are:

H5. Injunctive norms toward bribery for close friends (INF) is positively related to bribe giving intention

H6. Injunctive norms toward bribery for parents (INP) is positively related to bribe giving intention

Perceived behavioral control (PBC)

RAA model further suggests that individuals would consider their ability to carry out the behavior before performing it, which is PBC. Studies show that when one has a higher sense of PBC, they are more likely to perform the behavior (Rabl & Kühlmann, 2008). PBC can be further divided into two categories, namely capacity and autonomy (Fishbein & Ajzen, 2010). Capacity can be defined as the reflection of internal factors that one has control over, such as individuals’ perceived level of confidence in performing the behavior (Yzer, 2012). Capacity was found to significantly influence one’s intention in many RAA related studies, such as the intention to underreport actual income to avoid tax (Rosid et al., 2018), as well as the intention to consume more vegetables among young adults (Menozzi et al., 2015). When one is perceived to have capacity to perform the behavior, their confidence in doing so will increase (Fishbein & Ajzen, 2010). Bribe giver’s socioeconomic background, such as their wealth, power, position, and opportunities are the factors that influence their confidence (Othman et al., 2014). However, research shows that having a low perceived ability to offer bribes may not necessarily reduce bribing behavior. Peiffer and Rose (2018) found that individuals who come from disadvantaged backgrounds are more likely to bribe for better offers and experience. They are also more likely to take bribes (Ariyabuddhiphongs & Hongladarom, 2014).

Autonomy can be defined as external control factors that individuals’ perceived to have over their behavior (Yzer, 2012). In other words, individuals that have a higher sense of autonomy when it comes to bribing think that they themselves have complete control over making that decision. Previous study related to traffic violations found that when individual have higher perceived autonomy over traffic violations (e.g., drink-driving, ignoring road signs), the less likely they would exhibit the intention or behavior to do so even under peer pressure (Castanier et al., 2013). Jacob and Ouimet (2017) also demonstrated that when civil servants in Canada have higher perceived autonomy, the more trust they have about their ability to resolve workplace ethical dilemmas. However, the extent of autonomy we can have over our behavior may be limited by environmental factors. For example, you rely on government’s funding to pay for your children’s education. However, tomorrow is the last day of the grace period for paying the school fees and you have yet to receive legal documents that verify your situation. What if the only way to resolve this is to bribe or your child may have to withdraw from the program? Thus, your sense of autonomy toward the decision to bribe or not is well limited by the situation. This may explain why poor people are more likely to bribe (Peiffer & Rose, 2018), as the consequences of not bribing may be more expensive and difficult to deal with.

According to all the reviewed literature, the study proposed the hypotheses to be:

H7. Capacity is positively related to bribe giving intention

H8. Autonomy is negatively related to bribe giving intention

Figure 1 shows this study’s research model that is formed from the above hypotheses and reviewed literature.

Figure 1. The research model.

Methodology

Ethical consideration

This study was approved by the university’s ethics committee. Participants also provided informed consent prior to participating in the survey.

Study design

This correlational cross-sectional study uses an online survey method to explore all RAA factors related to bribe giving intention. Five different bribe-giving scenarios were presented to the respondents in this study using the situational judgement technique (SJT) (Stevens, 2018, Tripathi, 2019). For each scenario, we included questionnaires measuring every RAA factor.

1. [Large Money] In Malaysia, a person offers a traffic police a large amount of money to overlook their unlawful behavior.

2. [Small Money]In Malaysia, a person offers a traffic police a small amount of money to overlook their unlawful behavior.

3. [Small Gift] In Malaysia, a person offers a traffic police a small gift to overlook their unlawfulness.

4. [Illicit Giver] In Malaysia, a person gives a government employee a small gift in order to obtain a passport without proper documentation.

5. [Deserved Giver] In Malaysia, because of a delay, a person gives an immigration officer a small gift in order to make sure that his passport gets processed.

Respondents

A total of 600 Malaysian respondents initially participated. To ensure the reliability and validity of the data collected, we performed the following manipulation checks;

1. Attention check. Two checks were done for each scenario. The checks were administered before and after the respondent selected the corresponding bribe-giving scenario that was presented to them. This is to ensure the respondents fully understand the context of the scenario where they need to indicate their opinion. A total of 249 respondents’ data were excluded because they failed the attention checks.

2. Social desirability. There is a possibility that respondents might have a strong social desirability due to the nature of this topic. We included the English version of Kurzskala Soziale Erwünschtheit-Gamma (KSE-G) by Nießen et al. (2019) to measure one’s tendency to show socially appropriate response behavior. The scale is deemed to have high reliability and validity score (Münscher et al., 2020). Data from two respondents were further excluded as their mean scores deviated more than two standard deviations from the total mean.

3. Outliers. We then reviewed the raw data for outliers using standardized residual values that are above 3.3 or < –3.3 according to Pallant (2020). Four respondents were excluded.

4. Common Method Bias. Common Method Bias was accessed using the Harmon Single-Factor Test (Podsakoff et al., 2003). Results of the exploratory factor analysis found that the first factor explains 46% of the variance, thus indicating the absence of common method bias in this study (Babin et al., 2016).

Data from 345 respondents were used for final analysis after the manipulation check. 156 (45.20%) were males and 189 (54.80%) were females. The age range of the participants was from 18 to 30 years with the mean age of 20.68 years and SD = 2.01. n = 309 (90.60%) of the respondents are currently university students.

Measurement

The survey questionnaires consist of all RAA variables (IA, EA, DNF, DNP, INF, INP, capacity, autonomy) and also the dependent variable, which is the bribe giving intention. It was developed using Fishbein and Ajzen (2010)’s RAA survey development technique. This study also incorporated the situational judgement technique (SJT) method by modifying five bribe-giving scenarios from Truex’s (2011) study. SJT is a type of psychological aptitude test (Stevens, 2018) where respondents are presented with hypothetical scenarios that contain specific problems or dilemmas and provided with a set of response alternatives (Mussel et al., 2018). The five scenarios were chosen after consultation with the experts from The Malaysian Anti-Corruption Commission (MACC). The scenarios selected for this study are deemed relatable within the local social context and there was prevalent of such incidents happening among young adults.

Scales to measure RAA variables

To measure IA, the respondents were told to judge the five different scenarios presented to them with the question “The person’s action is acceptable” using 7-points Likert scale with 1 indicating Strongly Disagree to 7 Strongly Agree. The possible minimum score is 5 and maximum score is 35. Same method was used to measure all other RAA variables. Table 1 presents the question to measure each variable and the meaning of the score.

Table 1. Survey questions developed to measure all RAA variables.

RAA variable	Survey question	Meaning of higher score
Instrumental attitudes toward bribery (IA)	The person’s action is acceptable.	Positive instrumental attitudes toward bribery
Experiential attitudes toward bribery (EA)	If I am the person as described in the scenario, it is a pleasant experience to act in the same manner.	Positive experiential attitudes toward bribery
Close friends’ descriptive norms toward bribery (DNF)	If my close friend is the person as described in the scenario, s/he would act in the same manner.	Higher perception that close friends would offer bribe
Parent’s descriptive norms toward bribery (DNP	If my parents are the person as described in the scenario, they would act in the same manner.	Higher perception that parents would offer bribe
Close friend’s injunctive norms toward bribery (INF)	My close friends will agree with me if I act like the person in the scenario.	Higher perception that close friends would agree if bribe is offered
Parents’ injunctive norms toward bribery (INP)	My parents will agree with me if I act like the person in the scenario.	Higher perception that parents would agree if bribe is offered
Capacity	I am confident that I can act like the person in the scenario if I want to.	High capacity to bribe
Autonomy**	If I am the person as described in the scenario, I know that the action is beyond my control.	High autonomy to decide to offer bribe or not
Bribe giving intention	If I am the person as described in the scenario, I intend to do the same.	High in bribe giving intention

Note. **Reverse scoring. The measurements were adopted and modified based on Truex (2011).

Procedure

Advertisements about this study were posted on social media (i.e., Twitter, Facebook, Instagram, and YouTube) and the university’s public notice board. Interested respondents could scan the QR Code available on the advertisement to visit the research site. Upon agreeing to participate, respondents answered all the questions presented to them. Data collected were analyzed with the help from SmartPLS 3.0 software.

Results

The model was examined using structural equation model (SEM) to assess overall model fit (Bollen, 1989). SEM is considered a confirmatory factor analysis (CFA) (Schreiber et al., 2006; as cited in Yusif et al., 2020). The partial least squares structural equation modelling (PLS-SEM) was chosen for analysis to test the predictive variables based on the theoretical framework and to explore theoretical extensions.

The measurement model

The results of internal reliability and convergent validity for constructs were presented in Table 2. Convergent validity was assessed by examining (i) factor loading, (ii) CR, and (iii) AVE (Fornell & Larcker, 1981). The lowest factor loading for all items in this study was 0.72, which was above the recommended 0.60 (Chin et al., 1997). For CR, the liberal measure of internal consistency reliability, the score for all variables ranged from 0.90 to 0.95, which was also above recommended level (Gefen et al., 2000). For AVE, the value was acceptable as it ranged from 0.64 to 0.74 (Hair et al., 2010).

Table 2. Construct reliability and validity.

Constructs	Items	Loadings^a	AVE construct^b	CR^c
Instrumental attitudes toward bribery (IA)	IA1	0.79	0.64	0.90
	IA2	0.84
	IA3	0.88
	IA4	0.73
	IA5	0.76
Experiential attitudes toward bribery (EA)	EA1	0.84	0.70	0.92
	EA2	0.89
	EA3	0.89
	EA4	0.78
	EA5	0.77
Close friends’ descriptive norms toward bribery (DNF)	DNF1	0.86	0.69	0.92
	DNF2	0.88
	DNF3	0.85
	DNF4	0.79
	DNF5	0.77
Parent’s descriptive norms toward bribery (DNP)	DNP1	0.87	0.70	0.92
	DNP2	0.89
	DNP3	0.85
	DNP4	0.80
	DNP5	0.78
Close friend’s injunctive norms toward bribery (INF)	INF1	0.88	0.70	0.92
	INF2	0.89
	INF3	0.86
	INF4	0.78
	INF5	0.77
Parents’ injunctive norms toward bribery (INP)	INP1	0.88	0.71	0.93
	INP2	0.90
	INP3	0.88
	INP4	0.76
	INP5	0.78
Capacity	CAP1	0.87	0.77	0.95
	CAP2	0.90
	CAP3	0.90
	CAP4	0.81
	CAP5	0.82
Autonomy	AUT1	0.88	0.74	0.93
	AUT2	0.91
	AUT3	0.93
	AUT4	0.87
	AUT5	0.80
Bribe giving intention	BRIBE1	0.84	0.66	0.91
	BRIBE2	0.91
	BRIBE3	0.87
	BRIBE4	0.72
	BRIBE5	0.72

^aAll item loadings >0.6 indicates indicator reliability (Chin et al., 1997).

^bAll average variance extracted (AVE) >0.5 indicates convergent reliability (Hair et al., 2010).

^cAll composite reliability (CR) >0.7 indicates internal consistency (Gefen et al., 2000).

Discriminant validity is established when a measurement is not correlated with constructs to which it is assumed to be dissimilar (Ruel et al., 2016). It is also referred to as divergent validity (DeVellis, 2016). It can be measured using the heterotrait-monotrait ratio (HTMT) (Hair et al., 2019). To establish discriminant validity, the HTMT values should be <1.00 (Franke & Sarstedt, 2019). In this study, the HTMT values for all pairs of reflective constructs ranged from 0.40 to 0.96; hence, it can be concluded that discriminant validity has been established. In summary, the valid measurement model was established (see Table 3).

Table 3. Heterotrait-monotrait ratio (HTMT).

Latent variables	HTMT value
IA -> BRIBE	0.83
IA -> AUTONOMY	0.54
IA -> CAPACITY	0.57
IA -> DNF	0.76
IA -> DNP	0.73
IA -> EA	0.87
EA -> BRIBE	0.85
EA -> AUTONOMY	0.53
EA -> CAPACITY	0.62
EA -> DNF	0.78
EA -> DNP	0.73
DNF -> BRIBE	0.79
DNF -> AUTONOMY	0.53
DNF -> CAPACITY	0.71
DNP -> BRIBE	0.77
DNP -> AUTONOMY	0.49
DNP -> CAPACITY	0.63
DNP -> DNF	0.80
INF -> BRIBE	0.80
INF -> AUTONOMY	0.52
INF -> CAPACITY	0.73
INF -> DNF	0.94
INF -> DNP	0.78
INF -> EA	0.78
INF -> IA	0.79
INP -> BRIBE	0.76
INP -> AUTONOMY	0.45
INP -> CAPACITY	0.64
INP -> DNF	0.76
INP -> DNP	0.96
INP -> EA	0.73
INP -> IA	0.72
INP -> INF	0.78
CAPACITY -> BRIBE	0.71
CAPACITY -> AUTONOMY	0.40
AUTONOMY -> BRIBE	0.56

IA: instrumental attitudes toward bribery; EA: experiential attitudes toward bribery; DNF: close friends’ descriptive norms toward bribery; DNP: parents’ descriptive norms toward bribery; INF: close friends’ injunctive norms toward bribery; INP: parents’ injunctive norms toward bribery, CAPACITY: capacity; AUTONOMY: autonomy; BRIBE: bribe giving intention.

The structural model

The structural model represents the theory under study through an empirical approach to predict a model that consists of different variables (de Anda, 2018). The criteria for evaluation of structural models in PLS-SEM includes (i) t-statistics for significance, (ii) Coefficient of Determination R², and (iii) Predictive Relevance.

The direction and significance level of path coefficient help us to understand if the hypothesis can be supported by the proposed model (Yusif et al., 2020), and the magnitude provides insight about the strength of the relationship between two latent variables (Urbach & Ahlemann, 2010). The bootstrap test procedure with 5000 samples was carried out using SmartPLS. After that, t-statistics value (Möller, 2018) was carried out to confirm if path coefficients of the inner model are significant or not. The result (see Table 4) indicated four paths to be statistically significant and another four indicated otherwise.

Table 4. Summary of significant result testing of the structural model path coefficient.

Hypothesis	The path for RAA variable and bribe giving intention	Path coefficient/Std β	Mean (M)	Standard deviation	t-Statistics (|O/STDEV|)	p-Values	BCI LL	BCI UL	Hypothesis supported
H1	IA -> BRIBE GIVING**	0.27	0.27	0.09	3.10	0.00	0.10	0.44	Yes
H2	EA -> BRIBE GIVING**	0.26	0.26	0.10	2.78	0.01	0.08	0.43	Yes
H3	DNF -> BRIBE GIVING	0.01	0.01	0.11	0.11	0.92	−0.21	0.21	No
H4	DNP -> BRIBE GIVING*	0.24	0.24	0.11	2.24	0.03	0.05	0.46	Yes
H5	INF -> BRIBE GIVING	0.07	0.07	0.12	0.64	0.55	−0.15	0.29	No
H6	INP -> BRIBE GIVING	−0.08	−0.08	0.10	0.74	0.45	−0.30	0.12	No
H7	CAPACITY -> BRIBE GIVING**	0.20	0.20	0.05	3.96	0.00	0.11	0.31	Yes
H8	AUTONOMY -> BRIBE GIVING	−0.07	−0.07	0.05	1.64	0.12	−0.17	0.00	No

Note. *p < .05; **p < .01.

The path is significant if the critical value for t-statistics is larger than 1.96 and p-values is lower than .05.

R² value = 0.74.

A significant positive relationship was found between instrumental attitudes toward bribery (IA) (β = 0.27, t = 3.10, p < 0.01); experiential attitudes toward bribery (EA) (β = 0.26, t = 2.78, p < 0.01); descriptive norms toward bribery for parents (DNP) (β = 0.24, t = 2.24, p < 0.05); capacity to perform bribery (β = 0.20, t = 3.96, p < 0.01) and bribe giving intention. Hence, Hypothesis 1, 2, 4, and 7 were supported.

To explore the explanatory power of the model, R² value of 0.74 indicates a moderate coefficient of determination. This means that four out of eight latent variables included in this study moderately explain 74% of the variance in bribe giving intention. Besides that, the Stone-Geisser test is used to explore the study model’s predictive relevance (Q²) (Yusif et al., 2020). Q² value explains the model’s predictive power or predictive relevance. The Q² value for bribe giving intention is 0.48. The predictive relevance for the model is supported as Q² value is above zero.

Discussion

In this study, we explored bribe giving intention, one of the most prevalent forms of corruption. With the RAA model (Fishbein & Ajzen, 2010), we examined the micro factors related to bribe giving intention among young adults. The results showed that both instrumental and experiential attitudes, parents’ descriptive norms and capacity to perform bribery are the factors influencing the development of young adult’s bribe giving intention.

The results for attitudes toward bribery are consistent with those of other studies and suggest that attitudes are the main predictors of intention (Chang & Kerr, 2017; Gutiérrez et al., 2017; Lim, 2017; Simpser, 2020). This means, in order for young adults’ to form a bribe giving intention, they must think that it is an acceptable behavior and that the act itself should be perceived as a pleasant act. Although this result contradicts with study by Isa and Abdullah (2021) that indicates attitudes have no significant relationship with intention to bribe, their study measured bribe attitudes in general while the current study presents respondents with specific scenarios that are familiar to the respondents, such as giving bribes to police officers in order to measure their attitudes. This provides an ecologically-valid context to the nature of bribery.

Around 35.8% of university students think that acceptance of gifts in the form of money, goods or services in exchange for services given was not a corrupt act (Malay Mail, 2017). This poll reflects university students’ attitudes toward bribery. As attitudes can be accessed ahead of time and used to predict intention (Ajzen, 2005) we could work on modifying attitudes. Based on Newcomb’s (1967) longitudinal study, the college environment is important in shaping attitudes. Attitudes that a person holds when they are in their 20s could have a long lasting impact. In a practical way, this result encourages focusing on shaping negative IA and EA toward bribery among our young adults.

In this study, for perceived norms, there are four separate variables, close friends’ descriptive norms toward bribery (DNF), parent’s descriptive norms toward bribery (DNP), close friend’s injunctive norms toward bribery (INF) and parents’ injunctive norms toward bribery (INP). Prior studies have noted the importance of norms influence behavior (Crozier & Taylor, 2019; Kobis et al., 2019), however the current study has been unable to demonstrate the theoretical path for perceived norms suggested by Fishbein and Ajzen (2010). Surprisingly, only DNP has a significant positive relationship with bribe giving intention. This means the perception that their parents would or would not give bribes would change young adults’ intentions. On the contrary, the perception that their close friend would offer or would not offer a bribe has no influence in the bribe giving intention. Not only that, concern if parents (INP) or close friends (INF) would approve their decision to offer a bribe have no relationship as well. This result may be explained by the fact that the young adults are aware that; as compared to close friends (DNF and INF), parents have more opportunity to experience situations where they could offer bribes to police officers or immigration officers. For example, bribing traffic police indicating that one must have a driving licence, experiencing violating the rules and getting caught. It is logical to assume that SJT scenarios presented are hardly faced by the respondents due to the young age. Malaysians could legally obtain a driving license when they turn 17, thus for the majority of the respondents have only around 3 years of driving experience. Study found that there is a significant association between young adults and bribing teachers or professors (Mangafic & Veselinovic, 2020). Perhaps, if the SJT scenario also measures bribing teachers or professors in which they have more potential experience, we may find INF and DNF to have a significant relationship.

This finding has important implications for us to re-think about norms when it comes to bribe giving behavior and its stakeholders. Different stakeholders influence young adults in different ways. Based on the scenarios presented in this study, when dealing with traffic police, government officers and immigration officers, parents are more influential than close friends when it comes to bribing police or government officers. Not only that, perception about what parents would do in the bribe giving scenarios have more influences on our young adults as compared to if they would approve the behavior or not (INP). In a practical setting, to decrease or curb bribe giving intention among young adults, the focus should be on parent’s descriptive norms. Government or non-governmental organizations could use this study result to educate parents that if their children think that they would bribe to resolve a problem, their children would form a high bribe giving intention.

For PBC variables, only capacity had a significant relationship with bribe giving intention but not the autonomy. This path is also contradictory to what Fishbein and Ajzen (2010) have suggested. People who have high confidence that they can offer a bribe would have a higher bribe giving intention. This finding is aligned with Ariyabuddhiphongs and Hongladarom (2014), Benk et al. (2017), and Sanyal (2005) studies. It is important to note that capacity is also known as the internal control (I am confident that I can give bribe) factor while autonomy is regarded as external control factor (I know that the action is beyond my control) (Fishbein & Ajzen, 2010). The result suggests that internal control is more influential when guiding one’s intention. Although past literature suggests variables, such as money or power that would influence one’s confidence level to bribe or not (Othman et al., 2014), future studies could explore in detail the meaning behind confidence in bribing. This is because, as a young adult who is still studying, their confidence level to bribe might be motivated by different factors other than money and power. The speculation posits the confidence to bribe is built through personal observation of inconsequential bribing activities. Hence, it would be interesting to study how personal experience is related to the confidence in the act of bribing (Medley-Rath & Morgan, 2022).

From a practical perspective, to curb bribe-giving behavior in developing societies, effort must be directed in shaping negative instrumental and experiential attitudes toward bribery, reducing young adult’s belief that their parents would pay bribes in the given situation and reducing one’s confidence to give bribes.

Ethical approval

This research received approval from the University’s Ethics Committee.

Author contributions

Lim Mengzhen: project administration, writing—original draft, methodology, software, validation, formal analysis, and investigation. Yongchy Sin: writing—review and editing. Wan Munira Wan Jaafar, Azlina Mohd Khir, Hanina H. Hamsan, and Min Hooi Yong: supervision and conceptualization. Shin Ling Wu, Pei Boon Ooi, Derek Lai Teik Ong, and Chu Sun Ong: software and formal analysis.

Acknowledgments

We would like to extend appreciation toward Malaysia Anti-Corruption Commission (MACC) for the input while planning for this research. Special thanks to Professor T. Ramayah for advice on the PLS-SEM method. The research team would like to thank all the research assistants for their assistance in one way or another in completing this research. They are: Syahrul Zharif bin Saidina Othman, Kovithaa Selvam, Aisyah Mohd Farid, Tan Lee Shee, Nava Waheed, Che Wan Danish Hilman Bin Che Wan Khairul Anuar, Amanda See Pei Yee, Yungqi Khoo, Bilvashri Seyon, and Mariyam Aroofa Arif.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This research was self-funded by the principal investigator.