When participants returned to the lab for Session Three, they were asked to
respond to 54 rating questions about each of their own moral narratives. (The methods for this are described in Chapter 2; the questionnaire administered is found in Appendix C). These ratings allowed us to investigate a number of open questions about moral
perceptions of self, including looking at accuracy of self-perception and the effects of demographic and neuropsychological variables. These results are most easily interpreted
when they are framed by the questions that initiated them. To that end, these questions will be used as headings throughout this section.
Among the ratings collected in Session Three were three questions in which participants rated (on a scale from 1 to 10) the morality of their own narratives. They were asked to reflect on the morality of their behavior in the narrative at two time points: first, thinking back to their feelings at the time and second, looking back on the situation now (i.e., on the day of testing)235. For the third rating, the participants were asked to judge how others reading the narrative would rate the behavior in the narrative. These three ratings allow us to investigate whether participants’ ratings of themselves change over time and how critical participants are of their own behavior. By comparing these ratings to those generated by the raters who read these narratives we can also examine whether the participants’ own ratings are similar to outsiders’ ratings as well as how accurate participants are in assessing how others judge them236.
Does the length of time since the memory occurred affect how moral the participant feels the action was?
The saying that ‘time heals all wounds’ alludes to the idea that over time painful memories can become less painful. Chapter 4, Section 6 discusses the effects of the age
235 The specific questions given were: “At the time, how wrong or right did your actions in this episode seem to you? Use a scale from 1 to 10. If, at the time you did it, you felt this was a very wrong thing to do, you would give it a 1. If, at the time you did it, you felt this was a very right thing to do, you would give it a 10.” and “Reading this today, how wrong or right do your past actions in this episode seem to you? Use a scale from 1 to 10. If now you feel this was a very wrong thing to do, you would give it a 1. If now you feel this was a very right thing to do, you would give it a 10.”
236 It is very important to note that these data can only be considered preliminary findings. As was discussed in Chapter 3, the raters were chosen to be a homogenous group (with respect to education). All other demographic information about the raters is unknown. Thus, these data are biased at least by
education. However, the raters are still diverse in the ratings they gave and we feel confident using them for these preliminary analyses.
of a memory on its recollection. Knowing that a memory’s remoteness can affect its recall prompted us to explore whether the age of a memory can change how the
participant feels about their actions, as well. As shown in Figure 40, 45% of participants did not change their feelings about how right or wrong their actions were between the time of the narrative and the time of the rating. Those who did change their feelings did so in a normally distributed way across the entire range of possible changes. The mean change was to lower the rating by 0.5 points, but the variance in this was quite large (SD = ± 2.93 points). When we calculate the difference between the ratings of how
participants felt at the time of the event and how participants felt at the time of the rating, averaged over 5 year periods, we can see that the difference in ratings changes very little for different time frames (Figure 41). This leads us to conclude that time for reflection affects participants’ moral ratings about their actions very little.
Do subjects judge themselves differently than they think others will? How accurate are subjects at predicting what others think of them?
Finding out whether we rate ourselves differently than others rate us in any situation is fundamental to many social behaviors237. We were able to compare our subjects’ own ratings of their narratives with ratings from other people, as well as with their own perceived ratings of others’ ratings. We find that subjects’ ratings of
themselves are highly correlated with the rating they think others would give them (R2 =
0.64, p < 0.001) as well as with the actual ratings that raters give them (R2 = 0.63, p <
237 Gossip is an excellent example of this.
0.001). The ratings they think others would give them and the ratings that they actually receive are also highly correlated (R2 = 0.57, p < 0.001).
While these ratings are all highly correlated, the correlation does not reveal the constant difference in rating scores. We find that subjects rate themselves as significantly more moral than they are rated by others, rating themselves an average of 0.78 ± 1.47 points higher on a scale from one to ten (t[757] = 7.16, p < 0.0001). (Similarly, subjects rate themselves significantly higher than they think others will rate them (mean = 0.77 ± 1.25 points, t[757] = 6.51, p < 0.0001).) Interestingly, despite rating themselves as more moral, subjects are very accurate at predicting the rating that others will give them. There is no significant difference between the prediction the subjects gave and the actual moral score given by the raters (mean = 0.07 ± 1.51 points, t[757] = 0.47, p = 0.64).
These findings suggest that, in general, subjects have an excellent sense of their ‘moral standing’ in the world but still feel that they are more moral than that standing would suggest. There are, however, demographic and neuropsychological factors that can impact how morally a participant perceives him or herself.
Does one gender rate themselves as more moral than the other?
When asked to rate how ethical they were on a scale from one to ten, all subjects rated themselves as quite ethical (mean = 8.4 ± 1.65 points for women, mean = 7.7 ± 1.70 points for men). However, the women’s self-ratings were significantly higher than the men’s (t[745] = 6.18, p < 0.0001). Ratings of the narratives by outside raters found no significant difference between the behaviors of men and women (t[756] = 1.87, p = 0.06), which is to be expected since the raters had no specific knowledge about the gender of
the narrator (only that that they assumed from the narrative) and, as reported in Section 5, raters were not particularly good at intuiting the gender of the rater from the narrative.
Are certain personality types more likely to be self-critical?
One might expect that personality would affect how a person views their morality, especially if, as shown above, we are able to construct two views of morality – one that judges how we feel about ourselves and one that judges how the world views us. It appears that three personality factors can affect how critical a person is about their own moral position. Subjects who score high on the Neuroticism factor on the NEO-FFI personality measure are significantly more critical of their own morality than subjects who score low on this dimension (t[654] = 4.31, p < 0.0001), rating themselves 0.6 points lower on a 10 point scale. One can imagine that subjects who worry more might judge themselves more critically than subjects who are less prone to worry. However, these subjects are still able to make accurate assessments of how others will actually rate them. (Subjects who score high on N and subjects who score low on N are not significantly different when comparing their own judgments of how others will rate them to actual ratings by others; t[84]= 0.39, p = 0.70). This suggests that they are able to create two different assessments of their moral value, a personal (internal) one and an objective (external) one. Interestingly, subjects who score high on Openness and Agreeableness also are more critical of themselves than their low scoring counterparts (t[622] = 2.85, p < 0.01 for O; t[631] = 5.18, p < 0.0001 for A), rating themselves 0.4 points lower (for O) and 0.7 points lower (for A). While subjects with high O scores retain the ability to accurately judge how others will rate their behavior (t[92] = 0.18, p = 0.86 for O),
subjects with high A scores do have difficulty with this assessment when compared to their low A peers (t[85] = 2.58, p = 0.01). While the interpretation for this finding is less straightforward than for the N dimension, it may be that subjects who are more open and more agreeable are more empathetic and therefore less critical of other people, rating themselves lower by comparison to the rest of the social group. Subjects who are less open and less agreeable may be more judgmental of others and therefore rate themselves higher by comparison. More research would be needed to investigate the ramifications of personality difference on self ratings; including other measures of self-worth could expand this finding significantly.
Does IQ affect a participant’s personal moral rating?
Since we have shown that subjects are capable of maintaining two separate constructs about moral value, it seems important to measure whether a participant’s intelligence has any impact on their ability to generate these two values. A linear
regression shows that IQ has no significant effect on a participant’s personal moral rating (R2 = 0.03, p = 0.29). Similarly, we find that IQ also does not affect a participant’s ability
to predict others’ ratings of his or her own behavior (R2 = 0.00004, p = 0.99). Together, the findings demonstrate that IQ does not affect a participant’s ability to generate separate moral ratings for their own perception of moral value and the perception of others about them.