English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
Rye and Pargament (2002) have defined forgiveness as letting go of negative affect, cognition, and behavior in response to considerable injustice. In their view, forgiveness may also involve responding positively toward the offender, through compassion for example (see also, Worthington, 2005). Since about thirty years (Enright, Santos, & Al-Mabuk, 1989), psychologists have empirically examined the way people conceptualize forgiveness and the determinants of forgiveness.
Two kinds of determinants have been investigated: Personal and situational. Studies on personal determinants have shown that forgivingness - the general disposition to forgive (Roberts, 1995) - considerably vary from one person to another, and is associated with forgiveness culture in the family during childhood, religious affiliation, personal development through the lifespan, broad personality traits (e.g., emotional stability), more specific personality traits (e.g., perspective-taking abilities), and psychopathology, to quote a few. It has been shown that: (a) people who have experienced high levels of forgivingness from parents were themselves more inclined to forgive than people whose parents were reluctant to do it (Mullet, Rivière, & Munoz-Sastre, 2006), (b) Buddhists were slightly more resentful and less forgiving than Christians from the same cultural background (Paz, Neto, & Mullet , 2007), (c) people with an extensive time perspective (usually younger people) were less forgiving than people with a limited time perspective, usually older people (Allemand, 2008; Maganto & Garaigordobil, 2010), (d) neurotic people were less willing to forgive in general than emotionally stable people (Mullet, Neto, & Rivière, 2005), (e) people who were capable of perspective-taking were more forgiving than people unable to do so (Takaku, 2001), and (f) people with autism had trouble using information regarding the harmful act before forgiving (Rogé & Mullet, 2011).
Studies on situational determinants have shown that offender's intent to harm, type of offense, severity and permanence/cancellation of consequences, presence of apologies, and close other's attitudes impact on willingness to forgive in concrete situations (e.g., Gauché & Mullet, 2005). When harm was collective, these situational determinants -acknowledgement of responsibility in particular - were also shown to play a significant role (Etxebarria, Páez, Valencia, Bilbao, & Zubieta, 2010). Fehr, Gelfand and Nag (2010) me-ta-analyzed results from 175 studies reporting correlates of interpersonal forgiveness and concluded that these situational determinants accounted for much greater variance in willingness to forgiveness than people's dispositions.
Most of the studies on forgiveness included in Fehr et al. (2010) meta-analysis have been conducted in North America or in Western Europe. The few studies on willingness to forgive that have been conducted in Latin America (López-López et al., 2013, 2018) showed, however, a different picture than the one suggested by these authors. In these studies a huge majority of Latin American participants seemed to be totally insensitive to circumstances.
López-López et al. (2013) examined Colombian people's willingness to forgive persons who have been more or less actively involved in the violence that ravaged the country during the past 60 years. They used a scenario technique and presented their 400 adult participants (aged 18-55 years) with 48 realistic cases in which a former perpetrator of violence (a member of the guerillas, the paramilitary, the military or a drug cartel) asked (or does not asked) for forgiveness from a victim's family. These cases were constructed using a three-factor orthogonal design: Degree of Responsibility (organizer, mere agent, or passive bystander) x Severity of the Negative Acts Committed (murder, kidnapping, destruction of property, or theft) x Apologies (no apology at all, acknowledgment of responsibility, begging forgiveness, and begging forgiveness and offering reparation).
Participants were asked to judge the degree to which they would be willing to forgive if they were a family member. Through cluster analysis, three basic attitudes were found. The most common one, which was shared by 67 % of the sample, was termed "no forgiveness under any condition." Fifteen percent of participants expressed the opposite attitude; that is, they considered that forgiveness should be systematically granted, irrespective of circumstances. Only a small minority (18%) of participants, mostly from the poorest segments of society, took into account the circumstances described in the scenarios. They considered that forgiveness could be granted each time the former perpetrators expressed true repentance (and, in the case of former organizers, if they have offered adequate compensation and had not committed very severe crimes). López-López et al. 2018, using the same methodology and a larger sample (N = 550, aged 18-67 years) replicated these findings. The respective percentages were 52% (never), 18% (always), and 13% (depending on circumstances). In addition, 17% of the sample was completely undetermined; that is, their ratings were always located in the middle of the response scale.
The Present Study
The fact that only a small minority of participants in the studies by López-López et al. (2013), 2018 took into account the circumstances of the offense before judging of their level of willingness to forgive perpetrators of violence may be attributed either to the particular situation examined in these studies (and, more generally, to the political context in Colombia) or to some Colombian people's peculiarities regarding the granting of forgiveness. The first reason seems, however, much more likely than the second reason. As shown by Bagnulo, Muñoz Sastre and Mullet (2009), Latin American people do not conceptualize forgiveness in a way that is fundamentally different from the one found in Western Europe (see also, Mullet & Neto, 2014).
The present study examined the determinants of Colombian people's willingness to forgive in a more neutral context than the one studied by López-López et al. (2013). The material was adapted from Mullet et al. (2007) who presented their participants with scenarios depicting a situation in which a medical error had been committed by a physician. Four factors were selected: (a) the severity of consequences of the error (severe vs. extremely severe, including the risk of death), (b) the degree of negligence associated with the physician's act (clear negligence vs. no negligence), (c) apologies or contrition for the act from the physician (direct apologies at the bed of the patient, indirect apologies to family members, or no apologies), and the patient's current health status (consequences still affecting the patient vs. consequences fully canceled).
Our hypothesis was that (a) a majority of Colombian participants would take into account the circumstances of the offense when judging of their level of willingness to forgive the physician, and (b) only a minority of participants would be insensitive to these factors and would judge in an all or none way; some of them would, as in López-López et al. (2013), never be willing to forgive and others would always be willing to forgive. This hypothesis was based (a) on the findings by Bagnulo et al. (2009) reported above, (b) on findings by Neto, Pinto, Suwartono,Chiaramello, and Mullet (2013) showing that Latin Americans did not differ from Western Europeans in their attitude towards seeking forgiveness, (c) on findings by Etxebarria et al. (2010) showing that Spanish people's attitudes to reconciliation were sensitive to conciliatory message from responsible authorities, and (d) on findings by Guédez and Mullet (2014) who showed that a substantial minority of Venezuelan people (43%) were, to a large extent, able to take into account the circumstances of an offense before forgiving themselves.
Method
Participants
Participants were 104 adults aged 18-76 years (M = 38.89, SD = 14.37) who lived in Bogotá. Their demographic characteristics are shown in Table 1. They were approached by one of the research assistants while they were walking along the main sidewalks of the city, usually in areas close to commercial centers and public buildings (e.g., post offices). They were not paid. The participation rate was 52% (in total, 200 people were contacted). The main motive given for not participating was lack of time. The study conformed to the ethical recommendations of the Colombian Society of Psychology; that is, full anonymity was respected and informed consent was obtained from all participants.
Material
The material consisted of 24 cards describing situations in which a doctor committed a medical error. Each scenario contained four items of information, in the following order: (a) the severity of consequences of the error, (b) the degree of negligence associated with the act, (c) apologies or contrition for the act, and (d) the patient's current health status. The scenarios were obtained by orthogonally crossing these four factors. The design was Severity x Negligence x Apologies x Current health status, 2 x 2 x 3 x 2.
An example of a scenario is the following: "Dr. Vasquez is Mr. Marco's family doctor. Dr. Vasquez prescribed a new treatment for Mr. Marco. Mr. Marco had a strong allergic reaction to this treatment. He was hospitalized for several months, during which he was between life and death. Mr. Marco's allergic reaction was not predictable from a medical point of view. While Mr. Marco was hospitalized, he was visited several times by Dr. Vasquez, who has expressed time and again his sincere apologies. Mr. Marco is now dismissed from the hospital and is leading a normal life. He won't suffer from any subsequent medical consequences. If you were Mr. Marco, to what extent would you be willing to forgive Dr. Vasquez?"
Procedure
The data collection procedure took place in 2016 and 2017. It was completed in a quiet room, usually in the participant's home. Each person was tested individually. Participants responded using a 15-point response scale ranging from definitely not (left anchor) to definitely yes (right anchor). As recommended by Anderson (2008), 2018, the experimentation included two phases. In the familiarization phase, the experimenter explained what was expected and presented each participant with 12 vignettes taken from the complete set. For each vignette, the participant read it out loud, was reminded by the experimenter of the items of information in the story, and then made an acceptability rating by marking the response scale. After completing the 12 ratings, the participant was allowed to review his/her responses and compare and change them if needed. In the experimental phase, each participant gave ratings for the whole set of 24 vignettes. No time limit was imposed, but the participant was no longer allowed to review and change previous responses. In both phases, the experimenter made certain that each participant, regardless of age or educational level, was able to grasp all the necessary information before making a rating.
The participants took 25 to 35 minutes to complete both phases. The experimental phase was shorter since the participants were already familiar with the task and the material. The participants were told in advance of the approximate length of the experiment. No participant voiced any complaint about the number of vignettes or about the credibility of the proposed situations.
Table 1 Demographic Characteristics of the Sample and of Each Cluster. Values in parentheses are percentages.
| Characteristic | Almost Never Forgive | Depends on Circumstances | Almost Always Forgive | Undeterm | Total |
|---|---|---|---|---|---|
| Age | |||||
| 18-29 Years | 5 (14) | 20 (57) | 9 (26) | 1 (3) | 35 |
| 30-45 Years | 5 (15) | 16 (49) | 9 (27) | 3 (9) | 33 |
| 48+ Years | 6 (17) | 21 (58) | 7 (19) | 2 (6) | 36 |
| Gender | |||||
| Male | 10 (22) | 25 (54) | 11 (24) | 0 (0)a | 46 |
| Female | 6 (10) | 32 (55) | 14 (24) | 6 (10)a | 58 |
| Education | |||||
| Primary and Secondary | 10 (23)a | 24 (55) | 9 (20) | 1 (2) | 44 |
| Tertiary | 6 (10)a | 33 (55) | 16 (27) | 5 (8) | 60 |
| Social Class | |||||
| Lower Class | 3 (19) | 5 (31)a | 7 (44) a | 1 (6) | 16 |
| Middle Class | 8 (13) | 38 (62)a | 12 (20)a | 3 (5) | 61 |
| Upper Class | 5 (19) | 14 (52) | 6 (22) | 2 (7) | 27 |
| Religious Involvement | |||||
| Atheist | 4 (50)a | 2 (25) | 1 (12) | 1 (12) | 8 |
| Regular Attendee | 12 (13)a | 55 (57) | 24 (25) | 5 (5) | 96 |
| Victim of Error | |||||
| No | 10 (14) | 41 (56) | 17 (23) | 5 (7) | 73 |
| Yes | 6 (19) | 16 (52) | 8 (26) | 1 (3) | 31 |
| Total | 16 (15) | 57 (55) | 25 (24) | 6 (6) | 104 |
Note. Percentages with the same lowercase letter are significantly different.
*p < .05
Results
A cluster analysis, using the K-means procedure (Hofmans & Mullet, 2013), was first applied in order to detect qualitatively different patterns of ratings. A four-cluster solution was retained (Schepers & Hofmans, 2009). An overall ANOVA was conducted with a design of Cluster x Negligence x Severity x Apologies x Current health status, 4 x 2 x 2 x 3 x 2. Owing to the great number of comparisons, the significance threshold was set at .001. As the cluster effect and three of the four two-way interactions involving cluster were significant, separate ANOVAs were conducted at the cluster level (excepted for the smallest one). Results are shown in Table 2.
The first cluster (N = 16, 15% of the sample) was the expected almost never forgive cluster. As shown in Figure 1, both curves were located at the bottom of the graph; that is, most ratings were low (M = 3.54). They were slightly higher (a) when the consequences were fully cancelled (M = 4.56) than when they were not (M = 2.52), and (b) when the physician directly apologized to the patient (M = 4.22) than when the physician apologized to the family (M = 3.63) or when no apologies were offered (M = 2.77). As shown in Table 2, participants considering themselves as atheist and participants with primary or secondary education were more often members of this cluster than regular attendees to the church and participants with tertiary education.
The second cluster (N = 57, 55% of the sample) was the expected depends on circumstances cluster. As shown in Figure 1, curves were clearly ascending and separated. Ratings were clearly higher (a) when the consequences were fully cancelled (M = 9.04) than when they were not (M = 5.50), (b) when the physician directly apologized to the patient (M = 9.75) than when the physician apologized to the family (M = 6.90) or when no apologies were offered (M = 5.16), (c) when the allergic reaction could not have been anticipated (M = 8.17) than when it could have been anticipated (M = 6.38), and (d) when the reaction was not too severe (M = 7.53) than when it was very severe (M = 7.02). Participants from lower social class were less often members of this cluster than other participants.
Figure 1 Patterns of results corresponding to the four clusters. In each panel, the mean willingness to forgive ratings are on the y-axis, the apologies levels are on the x-axis, and the two curves correspond to the two levels of the cancellation of consequences factor.
The third cluster (N = 25, 24% of the sample) was the expected almost always forgive cluster. As shown in Figure 1, both curves were located at the top of the graph; that is, most ratings were high (M = 12.56). Ratings were slightly higher when the physician directly apologized to the patient (M = 13.34) than when the physician apologized to the family (M = 12.72) or when no apologies were offered (M = 11.62). Participants from lower social class were more often members of this cluster than other participants.
Finally, the fourth cluster (N = 6, 5% of the sample) was called undetermined. Ratings were always in the middle range of the scale (M = 7.60). Female participants were more often members of this cluster than male participants.
Two additional ANOVAs were conducted on the whole set of data, with gender or age as between-subject factors. They showed that, overall, the effects of these factors on willingness to forgive were not significant.
Discussion
As expected, a majority of participants took into account the circumstances of the offense for judging of their level of willingness to forgive in each case. For them, three situational factors were particularly important. Ratings were high when the error was accidental, when the physician directly apologized to the patient, and when the patient's current health state was good. Ratings were low when the error was the result of neglect from the physician's part, when no apologies were offered, and when the patient's current health state was still affected by the treatment. These results are fully consistent with findings reported by Mullet, Rivière and Muñoz-Sastre (2007). Compared to these factors, severity of reaction played a minor role; this result was also consistent with findings reported by these authors.
As expected, a minority of participants rated all scenarios either low or high; that is, they did not give much importance to the circumstances of the offense. They were, however, not completely blind to them. In both cases, the current state factor, for example, had an effect. As a result, it can be stated that the decision to rate all scenarios more or less in the same way is a voluntary decision, possibly based on personal conceptualizations regarding forgiveness: Atheists were, more (less) often than regular attendees members of the never (always) forgive cluster. This result was consistent with findings by Mullet et al. (2003) who reported a strong effect of religious involvement (namely, regular attendance to church) on willingness to forgive.
Finally, a small group of participants did not express any clear view regarding forgiveness in this kind of situation. This finding is important from a methodological view. If the participants in this cluster had been asked to give only one response-to a generic question or to a single scenario- their responses would have been interpreted as medium level of willingness to forgive in the particular situation described in the question or in the scenario. Having them respond to multiple scenarios permitted us to distinguish an absence of opinion from medium level of willingness to forgive, i.e., it showed, or at least suggested, that they did not actually make judgments but merely put marks more or less at the same spot for each scenario. Such a group of participants without any definite views had also been found in López-López et al. 2018.
