Adaptation of the Penal Attitudes Scale for use in Turkey

Abstract

Penal attitudes represent how people justify punishment assigned to offenders or what they perceive punishment’s function to be. The purpose of this study is to adapt the Penal Attitudes Scale (PENAS) for use in Turkey in Turkish and to test the resultant psychometric properties of the translated scale. For adaptation, a translation/back-translation method was applied. Respondents (N = 389) voluntarily participated in this study and completed the PENAS, Moral Foundations Questionnaire and Perceptions Toward Criminals Scale. The results of exploratory and confirmatory factor analyses showed that the six-structure PENAS is reliable and valid in the Turkish sample. Finally, the penal attitudes scale demonstrated good construct validity, showing statistically significant correlations with moral foundations and perceptions about the morality and social networks of criminals. Ultimately, the PENAS is a reliable, valid and highly useful instrument for the Turkish population.

Punishment is undoubtedly an indispensable tool for dealing with crime in societies. Punishment is defined as ‘a negative sanction intentionally applied to someone perceived to have violated a law, rule, norm, or expectation’ (Vidmar & Miller, 1980). Throughout evolutionary history, punishment has served to prevent group members from deviating from social norms or rules or re-offending by ostracizing or incarcerating them (Gollwitzer & van Prooijen, 2016; Wenzel & Okimoto, 2016). Since laypeople’s beliefs and perceptions of punishment have the potential to influence policymakers in societies, it is crucial to understand why ordinary people call for harsh punishment or what the rationale behind the punishment is. Although punitive reactions to lawbreakers (e.g. recommended years in jail, angry feelings toward them, perceived intent, responsibility and foreseeability) and social psychological mechanisms underlying public support for the death penalty have been scrutinized empirically in an array of disciplines (e.g. sociology, criminology and psychology) since the 1950s (see Carroll et al., 1987; Tyler & Weber, 1982), less attention has been devoted to exploring penal attitudes underlying desires to penalise wrongdoers in the social psychology literature until the 2000s.

Penal attitudes, also called punishment motivations, goals, or functions, are justifications of why and how transgressors should be penalised. With different justice concepts, punishment officially assigned to offenders might be perceived as just or legitimate (Colémont et al., 2011; Keijser et al., 2002). Penal attitudes have been classified in different ways so far (Carlsmith et al., 2002; Darley, 2009; Gollwitzer & van Prooijen, 2016). In the past, utilitarian (behavioural control, prospective, instrumental, or consequentialist perspective) and retributive (retrospective or just deserts perspective) approaches were distinguished (Carroll et al., 1987; Vidmar, 2000; Vidmar & Miller, 1980). The main difference between them is whether the motivation for administering punishment is future-oriented or past-oriented.

The utilitarian approach focuses on calculations of benefits over costs of punishment for the future, such as decreasing crime rates, by discouraging potential harm-doers (general or specific deterrence), incapacitating past offenders to prevent re-offending (incapacitation), or rehabilitating past offenders (rehabilitation) (Carlsmith et al., 2002; Carlsmith & Darley, 2008; Keijser et al., 2002). Although theoretically distinguished, specific and general deterrence approaches are quite different (Keijser et al., 2002). According to the general and specific deterrence notions, punishment acts as a deterrent to all members of a community and to a specific person, respectively. Specific deterrence, in general, overlaps with the incapacitation approach, since the primary goal of both is to prevent specific wrongdoers from recidivism by either discouraging or incarcerating them (Carlsmith & Darley, 2008). The rehabilitative approach focuses on past offenders’ reintegration into society, regaining moral identity and improving their behavioural repertoires by changing their maladaptive thoughts (Wenzel & Okimoto, 2016). Past research suggests that rehabilitation is negatively associated with incapacitation, retribution, deterrence and support for capital punishment (Feather & Souter, 2002).

The retributive approach is typically described as an ‘eye for an eye’ or a ‘tit-for-tat’ perspective and represents the most common and traditional justification for punishment (Darley & Pittman, 2003). More specifically, according to the retributive justice notion, offenders deserve punishment in proportion to the damage inflicted by the crime committed (Carlsmith et al., 2002; Carlsmith & Darley, 2008; Darley & Pittman, 2003; Keijser et al., 2002; Vidmar, 2000). That is, rather than being justified by future-forward and behavioural control purposes, punishment is considered important in itself as a reaction to past law violations. Previous research has demonstrated that the primary punishment motive is retribution (Carlsmith et al., 2002; Darley & Pittman, 2003; Keller et al., 2010). Subsequently, the retribution approach was divided into two categories: just deserts and moral balance (Keijser et al., 2002). The just deserts, also called the ‘get even’ perspective, focuses on the desire for retribution, revenge, vengeance and retaliation for offenses, while the main emphasis of the moral balance perspective is the restoration of moral balance in a society disrupted by crime (Gerber & Jackson, 2013; Keijser et al., 2002).

Finally, restorative justice (compensatory justice) has been gaining increasing acceptance in modern democracies. It is considered an alternative to the traditional criminal justice system (Gollwitzer & van Prooijen, 2016). The central focus of this approach is restitution and compensation for damage caused by criminal acts instead of punishment (Carlsmith & Darley, 2008; Darley & Pittman, 2003; Keijser et al., 2002). In restorative justice interventions, affected parties, both victims and offenders, take part in discussions so that offenders can express their remorse and apologize to victims. This process also enables both parties to express their feelings and discuss compensation (Gollwitzer & van Prooijen, 2016).

It is worth noting that none of punishment orientations is mutually exclusive: penalties or fines might be justified with more than one justice concept. Although deterrence, incapacitation and rehabilitation are theoretically categorized in the utilitarian perspective, rehabilitation differs greatly from the two (Keijser et al., 2002), and they are even reported to be negatively related (Feather & Souter, 2002). Hence, scholars developed several multidimensional penal attitudes scales to examine the motivations driving people to call for strict or lenient punishment for offenders (see Carroll et al., 1987; Gerber & Jackson, 2013; McKee & Feather, 2008; Okimoto & Wenzel, 2011; Yamamoto & Maeder, 2019). Some focused on satisfactory performance of criminal justice system, punishment and rehabilitation with the three-dimensional sentencing goal scale (Carroll et al., 1987). A multidimensional scale was developed by Keijser et al. (2002) to investigate Dutch judges’ and law students’ attitudes to punishment, with more precise conceptualization and operationalization of punishment approaches derived from moral theories. Following this, the 20-item sentencing goals scale was developed to assess people’s deterrence, retribution, incapacitation and rehabilitation-oriented attitudes (McKee & Feather, 2008). In addition, some focused on symbolic motives of punishment (value and status/power concerns) (Okimoto & Wenzel, 2011; Wenzel et al., 2012). By reviewing items in the current scales, Gerber and Jackson (2013) developed the retribution scale comprised of two subscales (retribution as just deserts and retribution as revenge) to provide more nuanced understanding of retributivism.

The measurement tools to investigate perceptions and attitudes about crime and punishment are scant in Turkish literature. There are a few developed and adapted scales in Turkish. For example, Gonultas et al. (2019) developed the Perceptions toward Criminals Scale to investigate perceptions about moral identity and social networks of criminals. Also, the Scale of Altruistic Fear of Crime was developed by Talu and Avci (2019) to examine perceived fear of crime of mothers about their children. The Scale of Attitudes on Violence developed by Velicer et al. (1989) and revised by Anderson et al. (2006) was adapted to Turkish (Ozkan et al., 2018), but it measures various types of violence (war-coded, intimate, physical, criminal-coded violence) and only the criminal coded subscale is related to perceptions and attitudes about crime or punishment. None have measured the rationale behind punishment from the perspective of ordinary people, to our knowledge.

There is a need for adaptation or development of scales or questionnaires in Turkey to achieve far-reaching understanding of penal attitudes, in particular moral justifications for punishment. Literature on penal attitudes is derived mainly from the European context (see Rade et al., 2016). Although Turkey is tied to Europe at cultural, economic and historical levels and embraces European legal and democratic principles, Turkish culture is very different from the individualistic culture in Western countries. Compared to Europe, Turkey has collectivistic values, where members are likely to display strong loyalty to familial and social groups, as well as to authority and, more broadly, the nation (Kağıtçıbaşı, 1997; Oyserman et al., 2002). Therefore, to achieve a far-reaching understanding of penal attitudes, in particular moral justifications for punishment, there is a need to adapt the scales in culturally, linguistically and religiously distant non-Western cultures.

The present study aimed to adapt the Penal Attitudes Scale (PENAS) for use in Turkey in Turkish and to test the resultant psychometric properties of the translated scale. To this end, we ran exploratory and confirmatory factor analyses of the version of the PENAS translated into Turkish on the Turkish population, to examine whether the six-factor model is validated. The lack of a reliable and valid penal attitudes scale in Turkish limits the evaluation of concurrent validity in this study. Therefore, the associations between the PENAS subscales and justice-related concepts are examined to assess criterion validity. Since moral justifications for punishment and justice attitudes are related to moral values (Silver, 2017), the psychological underpinnings behind penal attitudes might be related to moral foundations. Previous research has shown that utilitarian motivations for punishment are related more closely to the binding foundations (purity, authority and ingroup) than to the individualizing foundations (harm and fairness) (Giacomantonio et al., 2017). Furthermore, perceptions about the morality and social networks of criminals may insidiously reflect laypeople’s attitudes toward criminal justice. Because intentionality and impulsivity of wrongdoers are, respectively, taken into consideration in the retributive and utilitarian approaches (Carlsmith et al., 2002; Carlsmith & Darley, 2008; Keller et al., 2010), we expect that as morality is perceived more negatively, people endorse more moral balance, just deserts, deterrence and incapacitation. We did not draw specific hypotheses regarding the perceived social networks of criminals.

Method

Participants

Data were collected via a convenience sampling method. A total of 424 participants from Turkey took part in this study voluntarily. Since the data of 35 respondents who failed attention-check or left more than 5% of scale items missing were excluded from analyses, the final sample consisted of 389 participants. Of these, 279 (71.7%) were women, 108 (27.8%) were men, and one (0.3%) was other. The age range of participants was 18–66, and the mean age was 24.22 (SD = 7.5). Participants in the 18–25 age range, emerging adults, were prevalent, at 81.7%. They were followed by the 26–39 age range, young adults (10.8%), and those over 40 (7.5%), adults. The majority of the sample consisted of university students (81.7%); others held high school (8.7%), post-graduate (6.2%), secondary school (2.3%) and elementary school (0.8%) qualifications, and one (0.3%) is missing.

Materials

Penal Attitudes Scale (PENAS)

The 26-item PENAS was developed by Keijser et al. (2002). It consists of six dimensions: deterrence (4 items), just deserts (5 items), moral balance (3 items), incapacitation (6 items), rehabilitation (4 items) and restorative justice (4 items). The items are answered on a five-point anchored disagree/agree response set. The Cronbach’s α values in the original scale are .81 for deterrence, .82 for just deserts, .68 for restorative justice, .70 for moral balance, .68 for rehabilitation and .79 for incapacitation (Keijser et al., 2002).

Moral foundations questionnaire (MFQ)

The Turkish version (Yilmaz et al., 2016) of the Moral Foundations Questionnaire (Graham et al., 2011) was used to assess five moral dimensions (harm/care, fairness/reciprocity, authority/respect, purity/sanctity and loyalty/ingroup), each measured by six items. This scale is comprised of two sections: relevance and judgments. In the relevance section, participants were asked to indicate to what degree they agreed with a presented item (e.g. ‘whether or not someone did something to betray his or her group’) while making moral judgments. The responses were placed on a 6-point scale, with anchors from not at all relevant to extremely relevant. In the judgment section, participants were asked to what extent they agree with given sentences (e.g. ‘It can never be right to kill a human being.’) on a 6-point scale, anchored at 0 = strongly disagree to 5 = strongly agree. The harm/care and fairness foundations reflect individualizing foundations, whereas the authority, purity and ingroup foundations reflect binding foundations. It has been previously demonstrated that it is a reliable and valid instrument in Turkey to examine five moral foundations (Yilmaz et al., 2016). The Cronbach’s αs for the sub-dimensions in this study were as follows: .74 for purity, .70 for authority, .63 for ingroup, .63 for fairness and .60 for harm.

Perceptions toward criminals scale (PCS)

The PCS was developed by Gonultas et al. (2019). It comprises two sub-dimensions: perceptions about personality traits and morality of criminals (8 items, ‘Criminals are dangerous people.’) and perceptions about social networks of criminals (4 items, ‘Criminals come from families involved in crime.’). Respondents rated on a 7-point scale, ranging from strongly disagree to strongly agree. It yielded satisfactory internal consistency in our sample, α = .88 and .87, respectively, for perceived social networks and morality of criminals.

Procedure

Initially, the translation, synthesis and back-translation steps were conducted. In the first stage, 26 items from the original scale were translated into Turkish by researchers. The translated version was checked by a linguist, and the items were back-translated. In the final step, the items translated back to English were compared with the items in the original scale.

This study was approved by the Research Ethics Committee of the university of researchers in December 2021 (Approval No: 9EA47HC). The participants were approached through personal networks, convenience sampling and social media. Only adult Turkish native speakers were able to participate in the study. The participants completed the questionnaires either online via Google Forms or paper-based. All participants were informed of the purpose of the study. After obtaining consent, the data were collected anonymously on a voluntary basis. Respondents completed the demographic questions, PENAS, PCS and MFQ, respectively. One attention check was embedded in the survey to ensure that they completed the survey by reading each item. In the attention-check question, participants were asked to circle point 4 to indicate they read the questions carefully. Participants were fully debriefed and thanked for their participation at the end of the study.

Analysis

Data entry and preliminary analysis of the study were conducted via the IBM Statistical Package for the Social Sciences (SPSS) software version 24. In the case of the presence of missing items of less than 5% of the total in data, subjects’ intra-subscale mean replaced missing values. EFA and CFA were conducted consecutively to examine and confirm the factorial structure of the scale. Cronbach’s alpha values were calculated to examine the internal consistency of each dimension. Finally, for the criterion validity of the PENAS, its subscales’ correlations with those of both the MFQ and PCS were checked.

Results

Exploratory factor analysis (EFA)

To examine construct validity, a preliminary principal component analysis with oblique rotation technique was performed on 26 items. Factor loadings varied from .325 to .926. The acceptable factor loading is expected to be higher than .40 (Stevens, 2009). One item (‘It is not the function of the criminal justice system to reform criminals’) did not meet this standard. Therefore, the single item was removed from the scale.

Exploratory factor analysis with the extraction of six factors on the remaining 25 items was performed and re-analysed. A principal component analysis with oblique rotation technique was used. The Kaiser-Meyer-Olkin (KMO), the measure of sample adequacy, was .889, and Bartlett’s test of sphericity was significant, χ²(300) = 4677.072, p < .001. The results of EFA revealed six factors with eigenvalues higher than one: incapacitation (Items 1–5), just deserts (Items 6–10), deterrence (Items 11–14), moral balance (Items 15–18), rehabilitation (Items 19–21) and restorative justice (Items 22–25). The first factor explained 30.64% of the variance (eigenvalue = 7.66), the second factor 10.75% of the variance (eigenvalue = 2.69), the third factor 9.99% of the variance (eigenvalue = 2.5), the fourth factor 5.07% of the variance (eigenvalue = 1.27), the fifth factor 4.82% of the variance (eigenvalue = 1.21) and the sixth factor 4.14% of the variance (eigenvalue = 1.04). Factor loads of All items were higher than .40, varying from .426 (Item 15) to .929 (Item 9) (Table 1).

Table 1.

Factor loadings of items.

Items	M	SD	Components
			1	2	3	4	5	6
1. To ensure the safety of citizens, perpetrators of serious crimes should be incarcerated for as long as possible.	4.06	1.01	.79
2. Unless the perpetrator of a serious crime receives an unconditional prison sentence, he will continue to pose a threat to society.	4.2	0.93	.85
3. For a great many offenders, it is safer for society to have them locked up rather than walking around freely.	4.14	0.96	.80
4. In punishing serious crimes of violence, the safety of citizens is of greater importance than the needs of the offender.	4.1	0.99	.76
5. It is better to incarcerate known (regular) offenders for longer periods since this will prevent many crimes from taking place.	3.72	1.07	.66
6. Most people who advocate resocialization measures for perpetrators of offences attach little importance to the seriousness of the crimes committed.	2.99	1.18		.50
7. Punishment is deserved suffering.	3.38	1.16		.72
8. Even for alternative sanctions the infliction of suffering should be a prominent feature.	2.62	1.1		.86
9. Infliction of suffering should be an explicit element in every sanction.	2.54	1.1		.93
10. Punishment without an element of suffering is no punishment.	2.48	1.15		.87
11. In order to deter a larger number of potential offenders, heavier sentences should be imposed in (The Netherlands).	4.09	1.02					.81
12. If heavier sentences were to be meted out, more potential offenders would be deterred than at present.	3.95	1.02					.85
13. When criminal statistics show the prevalence of a particular offence to be increasing, the severity of punishment for that offence also should increase.	4.03	0.99					.72
14. Heavy sentences increase the credibility of the criminal justice system.	3.46	1.14					.78
15. Punishment restores the legal order in society disrupted by an act of crime.	3.83	0.95				.43
16. Punishment restores the ‘moral balance’ in society disrupted by a crime.	3.35	1.14				.74
17. The meting out of punishment to perpetrators of offences is a moral duty.	3.73	1.04				.70
18. By undergoing punishment, a criminal pays off his debt to society.	3.12	1.19				.76
19. The central focus of the criminal justice should be on the principle of correction.	3.86	1.0						.75
20. Officials in the criminal justice system have the moral duty to help criminals to get back on the right track.	3.55	1.12						.79
21. In establishing type and severity of punishment, the possibilities for resocialization should play a dominant role.	3.63	1.12						.70
22. Criminal prosecution is superfluous in situations where offender and victim have, as a result of mutual consultation, arrived at a solution to the conflict.	2.67	1.12			.69
23. A criminal process can only be qualified as a success when both offender and victim are satisfied with the outcome.	2.82	1.25			.79
24. The best form of punishment is one which, given the harm caused by the crime, maximizes the possibilities for restitution and compensation.	3.23	1.15			.60
25. The resolution of conflict is a neglected goal within our criminal justice system.	3.61	1.01			.47
Eigenvalues			7.66	2.69	2.5	1.27	1.21	1.04
Explained variance (%)			30.64	10.75	9.99	5.07	4.82	4.14
Reliabilities (Cronbach’s α)			.892	.867	.611	.766	.866	.684

Note. Component 1: Incapacitation, Component 2: Deserts, Component 3: Restorative Justice.

Component 4: Moral Balance, Component 5: Deterrence, Component 6: Rehabilitation.

Confirmatory factor analysis (CFA)

A confirmatory factor analysis with the final 25 items retained in EFA was conducted in IBM AMOS in order to ascertain whether the six-factor structure was the best fit for the data. The CFA analysis yielded adequate fit for the data (χ²/SD = 2.48 p < .001). The fit indices are as follows: root mean square error of approximation (RMSEA) = 0.062, normed fit index (NFI) = .866, comparative fit index (CFI) = .915, incremental fit index (IFI) = .915, root-mean-square residual (RMR) = .085 and goodness-of-fit index (GFI) = 0.879. Figure 1 illustrates the six-factor confirmatory analysis model.

Figure 1.

The six-factor confirmatory analysis model.

Reliabilities

To examine internal consistency, Cronbach’s alpha values were calculated for each dimension. The internal consistency coefficient was .89 for incapacitation, .88 for just deserts, .87 for deterrence, .77 for moral balance, .68 for rehabilitation and .61 for restorative justice. The alpha coefficient for the complete 25-item scale was .88. This scale displayed satisfactory internal consistencies in the Turkish sample. These reliabilities were comparable to those reported in the original scale (from α = .68 to .82) (Keijser et al., 2002).

Criterion validity

For the criterion validity of the PENAS, we examined its subscales’ correlations with those of both the MFQ and the PCS. Results yielded significant correlations between the subscales of the PENAS. The correlation matrix in Table 2 demonstrated that deterrence was positively correlated with other dimensions (r = .61 for incapacitation, r = .52 for moral balance and for r =. 49 for just deserts, for all p < .01), except restorative justice and rehabilitation (p > .05). Just deserts was positively associated with other subscales (r = .39 for incapacitation, r = .38 for moral balance and r = .23 for restorative justice, for all p < .01), except rehabilitation (p  > .05). Incapacitation showed a positive correlation with moral balance (r = .54, p < .01) and a weak positive correlation with rehabilitation (r = .11, p < .05). Restorative justice was significantly correlated with both moral balance (r = .19, p < .01) and rehabilitation (r = .31, p < .01). Also, there was a positive association between moral balance and rehabilitation (r = .27, p < .01).

Table 2.

Correlations between subscales of PENAS, MFQ and PCS.

	Deter.	Deserts	Inca.	R. justice	M. balance	Rehab.	P. morality	P. social networks	Harm	Fairness	Ingroup	Authority	Purity
Deterrence	1
Deserts	.486**	1
Incapacitation	.611**	.393**	1
R. justice	.089	.230**	.046	1
M. balance	.519**	.383**	.536**	.187**	1
Rehabilitation	.076	−.074	.111*	.309**	.268**	1
P. morality	.375**	.343**	.356**	.057	.389**	−.047	1
P. social networks	.179**	.149**	.094	.070	.216**	.033	.383**	1
Harm	.233**	−.004	.263**	.080	.197**	.220**	.147**	.124**	1
Fairness	.165**	−.074	.263**	.056	.112*	.256**	.053	.115*	.701**	1
Ingroup	.360**	.217**	.196**	.159**	.330**	.126*	.262**	.214**	.434**	.298**	1
Authority	.322**	.299**	.213**	.135**	.283**	.033	.297**	.186**	.272**	.134**	.693**	1
Purity	.265**	.271**	.151**	.134**	.281**	.060	.295**	.203**	.351**	.185**	.642**	.728**	1
M	15.53	14.02	20.22	12.33	14.03	11.04	24.54	12.38	20.96	22.2	15.94	14.9	15.65
SD	3.53	4.6	4.16	3.08	3.33	2.54	5.57	3.4	4.54	4.36	5.18	5.54	6.08
Cronbach’s α	.866	.867	.892	.611	.766	.684	.869	.880	.599	.626	.635	.701	.740

Note. Deserts: just deserts; R. Justice: restorative justice; M. Balance: moral balance; P. Morality: perceived morality; P. Social networks: perceived social networks.

*p < .05; **p < .01.

We obtained significant positive correlations between perceived morality of offenders and deterrence (r = .37, p < .01). Furthermore, perceived morality showed a positive relationship with just deserts (r = .34, p < .05), incapacitation (r = .36, p <. 05) and moral balance (r = .39, p <.05). Perceived social networks of offenders were significantly correlated with deterrence (r = .18, p < .01), just deserts (r = .15, p < .01) and moral balance (r = .22, p < .01). They were also positively correlated with perceived morality (r = .38, p < .01).

With regard to moral foundations, harm showed a weak but positive correlation with deterrence (r = .23), incapacitation (r = .26), moral balance (r = .20), rehabilitation (r = .22), perceived morality (r = .15) and social networks (r = .12), all p <.01. Similarly, fairness was positively correlated with deterrence (r = .16), incapacitation (r = .26), moral balance (r = .11) and rehabilitation (r = .26), all p < .05. It was positively related to perceived social networks (r = .11, p < .05), yet did not show a positive correlation with perceived morality (p > .05). Ingroup was positively related to all penal attitudes (r values varied from .13 to .36), perceived morality (r = .26, p < .05) and social networks (r = .21, p < .05). Both authority and purity showed positive correlation with all penal attitudes (r values varied from .13 to .32, for all p < .05), except rehabilitation (p > .05), as well as both perceived criminals’ morality and social networks (r values varied from .14 to .39, for all p < .01).

Discussion

The purpose of this study was to evaluate the psychometric properties of the PENAS in the Turkish sample. Despite the considerable interest in moral judgments, the in-depth understanding of the nature of punishment motivations has been relatively neglected in literature. Therefore, there was a need for a scale in order to examine a broad array of penal attitudes. It is the first penal attitude scale in Turkish, to our knowledge.

We first ran exploratory factor analysis on the 26 items. Of the 26 items, 25 had values higher than .40, which showed the strong representativeness of the items. However, one item (‘It is not the function of the criminal justice system to reform criminals.’) with a low factor loading was excluded from the adapted scale. Since it is a reverse-scored item, the reason why it is not a function of criminal justice system might be attributed to other concepts of justice. That is, because it might be perceived that the primary function of criminal justice is deterrence or retribution, reforming criminals might be perceived as nonessential. This might result in loading positively on other factors to some extent, which in turn decreases the distinctiveness of the item, although it was expected to load negatively on the rehabilitation factor.

After conducting a new EFA with 25 items, we still reached the same six-factor structure of the scale, consistent with the original scale. However, one item, Item 6 (‘Most people who advocate resocialization measures for perpetrators of offences attach little importance to the seriousness of the crimes committed.’), loaded on the Just Deserts factor in our study, differing from the original version. It seems that Turkish population interpreted the sentence more focusing on the severity of the crime and the necessity of punishment rather than resocialization acts.

Furthermore, confirmatory factor analysis with the final 25 items retained in EFA yielded adequate fit when evaluated for fit indices (the items can be seen in the Appendix). To start with, the value of χ²/SD was found to be below 3 (χ²/SD = 2.48), indicating a good fit for the data (Kline, 2005). Comparative fit index value was over .90 (CFI = .915), also showing a good fit (Hu & Bentler, 1999). Similarly, the incremental fit index value indicated an acceptable fit, as it was over .90 (IFI = 0.915) (Kline, 2005). Moreover, root mean square error of approximation value was below .07 (RMSEA = 0.062), revealing a good fit (Steiger, 2007). However, obtained normed fit index (NFI) value was .866, and goodness-of-fit index (GFI) value was .879 for this sample, which were slightly less than cut off value of .90 for the good fit (Marsh & Grayson, 1995; Schumacker & Lomax, 1996). Last, the root-mean-square residual (RMR) value was .085 and was a little over the criterion of .08 for an acceptable fit (Hu & Bentler, 1999).

As with previous research (Feather & Souter, 2002; Giacomantonio et al., 2017; Keijser et al., 2002), our results revealed that the deterrence, just deserts, incapacitation and moral balance approaches were highly correlated with each other. In general, all of these are justifications for strict punishment desires at large and were taken as a single superordinate factor in past research (Colémont et al., 2011). The rehabilitation approach, on the other hand, was related to concepts of restorative justice and moral balance. In the original work, it was correlated with restorative justice, and both were claimed to represent social constructivist views despite the different theoretical backgrounds, but they did not show a link with moral balance (Keijser et al., 2002). In addition, rehabilitation is correlated with the incapacitation concept, albeit relatively weakly. Even though there was a negative correlation between rehabilitation and incapacitation in the original scale (Keijser et al., 2002), the positive correlation in our study is compatible with the theoretical framework, as both are utilitarian concepts.

The relationships between the subscales of penal attitudes and moral foundations were also assessed to get a more detailed understanding of which moral foundations might be behind punishment orientations. The higher harm and fairness, the greater incapacitation, deterrence, rehabilitation and moral balance were endorsed. In particular, the theoretically utilitarian approaches seem to be linked with individualizing values. Because utilitarian motives are related to concerns about the protection of individuals from potential harm in the future (Carlsmith & Darley, 2008), it is reasonable that people who prioritize individualizing values, characterized by concerns for others’ well-being, empathy for others, altruism, etc., aim to minimize the damage inflicted by crime, with utilitarian goals. Previous works had already established the relationship between individualizing foundations and endorsement of utilitarian and retributive standpoints (Giacomantonio et al., 2017), particularly in crimes victimizing individuals (i.e. theft, armed robbery, domestic violence, etc.), but not in victimless crimes, namely collective (i.e. funding terrorist organizations, revealing government’s secrets publicly, etc.) or crimes against the divine (i.e. consensual incest sexual intercourse, drug abuse, prostitution, etc.) (Silver, 2017). Although the correlations in our study seem to be relatively modest, the values are comparable to those in past research (Giacomantonio et al., 2017).

Regarding the binding foundations (ingroup, authority and purity), nearly all of them are associated with harsh sentencing goals, namely deterrence, incapacitation, moral balance and just deserts. As Giacomantonio et al. (2017) noted, collective interest outweighs the utilitarian justifications for punishment, even at the expense of proportionality, to encourage individuals to abide by legal rules and social norms. The links between the binding foundations and offender-centred retribution were also established (Giacomantonio et al., 2017; Silver, 2017). Therefore, that binding foundations are linked with utilitarian and retributive motives comes as no surprise. However, rehabilitation seems to be unrelated to the binding foundations, except ingroup. It is reasonable that people who are loyal to their in-group support reintegration and resocialization of ingroup lawbreakers. That, however, leaves the question of whether ingroup and rehabilitation are still correlated in the case of out-group violators. Future research might test if the rehabilitation approach is selective in intergroup settings. Another social constructivist approach, restorative justice, is related to the binding foundations to a small extent, though. It might be rooted in the notion that the primary goal of restorative justice is to re-establish and re-validate shared rules and norms that are expected to be adhered to by society through social consensus (Wenzel et al., 2008).

Both retributive and utilitarian motives are related to perceptions about criminals. The results revealed that the higher the just deserts, incapacitation, deterrence and moral balance attitudes, the more negative the perceived morality of criminals. It is consistent with previous research having demonstrated the positive correlation between stricter punishment and dispositional attribution to crime, such as negative personality traits (Carroll et al., 1987; Tetlock et al., 2007). This is because criminals’ morality denotes whether they intentionally cause harm or not, according to the retributive philosophical standpoint (Carlsmith et al., 2002; Carlsmith & Darley, 2008; Keller et al., 2010) and impulsivity, which implies the probability of committing the same offense in the future, according to the utilitarian perspective (Keller et al., 2010). Furthermore, all harsh punishment goals were related to offenders’ perceived social networks, except for incapacitation. However, the rehabilitation and restorative justice approaches were unrelated to both the perceived morality and social networks of criminals.

We are aware that our study is not without limitations. The penal attitudes scale (Keijser et al., 2002) was initially developed to examine moral justifications for punishment among judges and law students; therefore, it may have been difficult to understand for laypeople who are not familiar with criminal justice terminology. It is also noteworthy to highlight the limitation that using convenience sampling probably led us to have a less homogenous sample. As our sample consisted mostly of emerging adults (81.7%) and larger number of women (72%),the results may not be generalizable to the whole Turkish community. However, we believe that having a more educated sample tended to have positively impacted the results, as the punitive attitudes scale consisted of items of complex sentences and therefore required a higher level of cognitive capacity. Despite the inconsistent findings (see Feather & Souter, 2002; Giacomantonio et al., 2017; Shariff et al., 2014), penal attitudes might change with age. Penal attitudes in a more representative general population samples might be discussed in future research.

Another limitation is that behavioural and self-reported attitude measures might yield different results (Keller et al., 2010). Although seeking information related to retributive philosophy in order to recommend sentences, people may report as endorsing special deterrence attitudes. Penal attitudes might also differ depending on the type and severity of crime, or whether the crime is victimless or not (Silver, 2017).

This study has important theoretical and practical implications. Given that the criminal justice system is not independent of ordinary people’s attitudes about crime and punishment, demand for more lenient or stricter punishment might affect policymakers and the current criminal justice system, or vice versa. For example, opinion research in Turkey revealed that the majority of people are dissatisfied with the criminal justice system and do not find rehabilitation in prisons adequate (İnsan Hakları Derneği, 2021). By examining perceptions about why and how criminals should be penalised from the perspective of naive people, we can obtain more detailed information about the public demands regarding punishment.

Cross-cultural studies are needed to explore how penal attitudes change in different cultural settings. It is our hope that this scale adaptation may help to explain cultural differences in a broad array of penal attitudes in further research.

Declaration of conflicts of interest

Selen Göksal has declared no conflicts of interest.

Demet Vural Yüzbaşı has declared no conflicts of interest.

Filiz Kunuroglu has declared no conflicts of interest.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee [approved by Izmir Katip Çelebi University, Social Research Ethics Committee, decision no: 2021/23-10 and the date: 28.12.2021] and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Appendix.

The adapted form of PENAS in Turkish

Hapsetme

Madde 1. Vatandaşların güvenliğini sağlamak için, ciddi suçların failleri mümkün olan en uzun süre hapsedilmelidir.

Madde 2. Ciddi bir suçun faili şartsız hapis cezası almadıkça toplum için tehdit oluşturmaya devam edecektir.

Madde 3. Pek çok suçlunun, serbestçe dolaşmaları yerine hapsedilmeleri toplum için daha güvenlidir.

Madde 4. Ağır şiddet suçlarının cezalandırılmasında, vatandaşların güvenliği suçlunun ihtiyaçlarından daha önemlidir.

Madde 5. Birçok suçun işlenmesini önleyeceği için, bilinen (düzenli) suçluları daha uzun süre hapsetmek daha iyidir.

Hak Etme

Madde 6. Suç faillerinin yeniden topluma kazandırılmasına yönelik tedbirleri savunanların çoğu, işlenen suçların ciddiyetine pek önem vermiyor.

Madde 7. Ceza hak edilmiş acıdır.

Madde 8. Alternatif yaptırımlar için dahi, acı çektirme öne çıkan bir özellik olmalıdır.

Madde 9. Acı çektirme, her yaptırımda belirgin bir unsur olmalıdır.

Madde 10. Acı çekme unsuru içermeyen ceza, ceza değildir.

Caydırma

Madde 11. Suç işlemeye meyilli daha fazla sayıda kişiyi caydırmak için Türkiye’de daha ağır cezalar verilmelidir.

Madde 12. Geçmişte daha ağır cezalar verilmiş olsaydı, günümüzdekinden daha fazla suç işlemeye meyilli kişi caydırılmış olacaktı.

Madde 13. Suç istatistikleri belirli bir suçun yaygınlığının arttığını gösterdiğinde, o suç için verilen cezanın ağırlığı da artmalıdır.

Madde 14. Ağır cezalar ceza adalet sisteminin güvenilirliğini artırır.

Ahlaki Denge

Madde 15. Ceza, suç eylemiyle bozulan toplumdaki yasal düzeni yeniden sağlar.

Madde 16. Ceza, suç eylemiyle bozulan toplumdaki ahlaki dengeyi yeniden sağlar.

Madde 17. Suçluları cezalandırmak ahlaki bir görevdir.

Madde 18. Bir suçlu ceza çekerek topluma olan borcunu öder.

Rehabilitasyon

Madde 19. Ceza adaletinin temel odak noktası ıslah/düzeltme ilkesi olmalıdır.

Madde 20. Ceza adalet sistemindeki yetkililerin, suçluların doğru yola dönmesine yardımcı olma gibi bir ahlaki görevi vardır.

Madde 21. Cezanın türünü ve şiddetini belirlemede, yeniden topluma kazandırma olasılıkları baskın bir rol oynamalıdır.

Onarıcı Adalet

Madde 22. Suçlu ve mağdurun karşılıklı istişare sonucunda çatışmanın bir çözüme ulaştığı durumlarda cezai kovuşturma gereksizdir.

Madde 23. Bir ceza süreci ancak hem fail hem de mağdur sonuçtan memnun olduğunda başarılı olarak nitelendirilebilir.

Madde 24. En iyi ceza şekli, suçun neden olduğu zarar göz önünde bulundurulduğunda, zararın karşılandığı ve tazminat olanaklarının en üst düzeye çıkarıldığı cezadır.

Madde 25. Anlaşmazlıkların çözümü, ceza adalet sistemimizde ihmal edilen bir hedeftir.

Correction Statement

This article was originally published with errors, which have now been corrected in the online version. Please see Correction (http://dx.doi.org/10.1080/13218719.2023.2182115)