Motivation and Emotion

, Volume 32, Issue 4, pp 270–277

The effect of infant-like characteristics on empathic concern for adults in need

Authors

    • Department of PsychologyUniversity of Wisconsin Oshkosh
  • Luis V. Oceja
    • Despacho 83, Módulo 4, Departamento de Psicología Social y MetodologíaUniversidad Autonoma de Madrid
  • E. L. Stocks
    • Department of PsychologyUniversity of Texas at Tyler
  • Kirstin Zaspel
    • Department of PsychologyUniversity of Wisconsin Oshkosh
Original Paper

DOI: 10.1007/s11031-008-9101-5

Cite this article as:
Lishner, D.A., Oceja, L.V., Stocks, E.L. et al. Motiv Emot (2008) 32: 270. doi:10.1007/s11031-008-9101-5

Abstract

Three experiments tested the hypothesis that empathic concern for adults in need is enhanced by the degree of target infant-like characteristics. Participants reported feeling more empathic concern for an adult target with a more infant-like face than for an adult with a more adult-like face in a Spanish sample (Experiment 1) and in an American sample (Experiment 2). A similar effect was found when participants were presented with either an adult with a more infant-like voice or an adult with a more adult-like voice in a second American sample (Experiment 3). Additional analyses suggest that the infant-like characteristic effect on empathic concern is not mediated by observer perceptions of target attractiveness, target age or youthfulness, target vulnerability, or observer similarity to the target. These results support the proposition that infant-like cues enhance empathic concern in human observers and that the phenomenon generalizes across stimulus modality, gender, and nationality.

Keywords

EmpathyEmpathic concernInfant-like characteristicsBabyfacednessSimilarityVulnerabilityAttractivenessAge

Infant-like characteristics are linked to a variety of perceptual phenomena (for a comprehensive review see Berry and McArthur 1986). For example, research suggests that observers attribute higher levels of honesty, kindness, warmth, naiveté, social submissiveness, and physical weakness to babyfaced targets relative to non-babyfaced targets (McArthur and Apatow 1983/1984; Berry and McArthur 1985, 1986). Although interest in the effects of infant-like characteristics has focused primarily on perceptual responses of observers, ethologists such as Lorenz (1971) and Eibl-Eibesfeldt (1971) have long held that the key function of such features in humans and other organisms is to mobilize parental nurturant responses toward those who possess them.

McDougall (1908) offers an early psychological account of the link between the perception of infant-like characteristics and nurturant responses. He suggests that infant-like characteristics signal a target’s vulnerability, the perception of which activates the parental instinct in humans. According to McDougall, the parental instinct is psychologically complex and consists of cognitive, affective, and connotative (motivational) components. Whereas the cognitive and connotative components are highly flexible and influenced by an individual’s learning history, the affective component is invariant. With regard to the parental instinct, McDougall suggests that perceptions of vulnerability (cognitive) elicit altruistic caregiving motives and tendencies (connotative) via activation of the tender emotion (affective).

McDougall’s (1908) account is intriguing because he proposes that infant-like characteristics promote parental care by initiating an emotional-motivational response. Such a possibility appears to share some conceptual overlap with Panksepp’s (1998) model of mammalian parental care, which he believes is guided by an innate neural substrate that organizes and guides emotional and motivational responses to young targets (for a similar perspective, see MacLean 1990). According to Pankepp, this system is sensitive to various sensory cues found commonly in infant members of the species. Furthermore, this “basic” affect system, which Panksepp (1998) refers to as the CARE system, generates discrete feelings, which in combination with higher-order cognitive input, may form the basis for the capacity of humans to experience empathy and concern for those in need.

Consideration of this parental emotional-motivational perspective yields the hypothesis that those in need who possess relatively high amounts of infant-like characteristics will elicit stronger feelings of parental care in observers. Somewhat consistent with this hypothesis, Lorberbaum et al. (2002) found that human mothers who listened to infant vocal cries showed increased activity in brain regions that Panksepp (1998) and MacLean (1990) link to parental care in rats, such as the cingulate cortex, thalamus, midbrain, hypothalamus, dorsal and ventral striatum, and the lateral septal region. Although Lorberbaum et al.’s research suggests that exposure to infant-like sensory cues increases activity in a number of neural regions that comprise Panksepp’s CARE affect system, no subjective reports of participants’ emotional experience were acquired in their research.

Research by Keating et al. (2003) also provides some indirect support for the hypothesis that those in need who possess greater infant-like characteristics elicit higher parental affect in observers. They found people were more likely to return job applications left in public locations if the applications contained a picture that depicted a babyfaced (large eye sized) applicant than if the picture depicted a non-babyfaced (small eye sized) applicant. However, this effect only emerged for white male and black female applicants; pictures of white female and black male applicants failed to yield the effect. Examination of Keating et al.’s facial stimuli suggests that the babyface effect may have failed for the latter two types of applicants because the eye size manipulation resulted in highly exaggerated and caricatured looking applicant faces. In addition, as with Lorberbaum et al.’s (2002) study, it is unclear whether Keating et al.’s manipulations affected parental affect because they acquired no subjective reports of affective experience from participants.

Researchers working from a social-psychological perspective have reported findings that may speak to the tenability of the infant-like characteristic-parental affect hypothesis. Batson et al. (2005, Experiment 2) found that undergraduate participants asked to read about a dissimilar vulnerable target (young child, puppy, or dog) recovering from a broken leg reported feeling more sympathetic, compassionate, tender, softhearted, warm, and moved than participants asked to read about a similar, less vulnerable target (adult student). Similarly, Dijker (2001) found that participants reported higher levels of pity for a person who had age, gender, and posture characteristics that suggested the person was vulnerable. The feelings measured in these studies are conceptualized as comprising an affective response commonly referred to as empathy (Batson 1991) or empathic concern (Davis 1996) in the prosocial motivation literature. The affective response measured by Batson et al. (2005) appears particularly relevant as it seems to capture the quality of feeling many people experience when encountering infants and children in a wide variety of situations.

Although Batson et al. (2005) and Dijker (2001) show that relatively vulnerable targets do indeed elicit enhanced “tender” feelings in observers relative to less vulnerable targets, the ambiguity of their manipulations raises questions about what produced the effect. Specifically, it is unclear whether the effect results from the manipulation of observed or imagined infant-like characteristics of the targets or the manipulation of more complex cognitive perceptions that potentially varied with their experimental stimuli.

The present research

Despite the aforementioned evidence that targets in need who possess a high degree of infant-like characteristics evoke stronger parental-like affect in observers, this hypothesis has received weak experimental scrutiny. In the present research, this hypothesis was tested by examining how varying the degree of infant-like characteristics influenced observers’ experience of empathic concern using two distinct sensory manipulations of infant-like characteristics (faces and voices) across three experiments. The present research also examined the cross-national generalizability of the infant-like face effect using Spanish and American samples.1

In Experiment 1, Spanish participants were presented with a story about a person in need and asked to indicate the extent to which they felt empathic concern for the person. The story contained a picture of the person, whose features were altered to appear relatively infant-like or adult-like. In Experiment 2, American college students were presented with the same story and pictures and asked to rate the extent to which they felt empathic concern. Experiment 2 also included an item that assessed the extent to which participants felt protective toward the person in need to determine if feeling empathic concern was linked to a parental-like orientation toward the person. In Experiment 3, the effect of infant-like characteristics on empathic concern was examined by presenting American college students with an audio interview of a person in need in which the person’s voice was digitally altered to sound relatively adult-like or infant-like. Finally, in Experiments 2 and 3, participants were presented with various items designed to measure perceived attractiveness (Experiment 2), perceived youthfulness (Experiment 2) or age (Experiment 3), perceived vulnerability (Experiments 2 and 3), and perceived observer similarity to the person in need (Experiments 2 and 3) to determine if such perceptions accounted for any potential infant-like characteristic effect on empathic concern.

Method

Experiment 1

Participants

Participants were 40 non-student, Spanish adults (19 women, 21 men) who volunteered to participate in psychological research. Eleven women and 12 men were randomly assigned to the infant-like face condition and 8 women and 9 men were randomly assigned to the adult face condition.

Procedure

Participants were run individually. Research assistants visited participants’ homes and asked them to complete a questionnaire packet alone in a separate room. The purpose of the study was described as an examination of people’s reactions to potential news columns in a local newspaper.

All participants received the same news article about a female student whose parents recently died in an automobile accident. According to the article, the student was struggling to finish school while taking care of her surviving siblings. One of two pictures of the young woman was presented within the article. Both pictures were selected from a larger stimulus set created by Gründl (2006) to explore the effect of babyfacedness on observers’ perceptions of attractiveness. The first picture was of a young adult German woman. The second picture was an image of the same woman that was altered to make her appear more infant-like. The infant-like version of the picture was created by morphing the original adult face with a computer generated babyface template. The babyface template was based on the relative spacing of facial characteristics derived from averaging four actual babies’ faces. The morphed face was generated by setting the contribution of the adult face and babyface template each at 50%. The resulting adult and infant-like faces are presented in Fig. 1. Research assistants were unaware of which face participants received.
https://static-content.springer.com/image/art%3A10.1007%2Fs11031-008-9101-5/MediaObjects/11031_2008_9101_Fig1_HTML.jpg
Fig. 1

Grayscale images of the adult face (left) and infant-like face (right) used in Experiments 1 and 2. Images were presented in color to participants. Images reprinted with permission from Gründl (2006)

After reading the article, participants rated how much they experienced a variety of emotional reactions while reading the article using a 7-point scale (1 = Not at all, 4 = Moderately, 7 = Extremely). Included among these reactions were five adjectives used to assess empathic concern in previous research (Batson et al. 1987, 2005). These adjectives were compassionate, softhearted, tender, moved, and warm.

Results

Participants’ ratings of the five empathic concern adjectives were averaged to form an empathic concern index (α = .90). A 2 (Gender: male vs. female) × 2 (Characteristic: infant-like vs. adult) ANOVA revealed a statistically significant Characteristic main effect on empathic concern: participants presented with the infant-like face reported higher levels of empathic concern for the person in need than did participants presented with the adult face, F(1, 36) = 14.31, p = .001, \( \eta _{\text{p}}^{2} = .28 \) (see Table 1). Neither the Gender main effect nor the Gender × Characteristic interaction achieved statistical significance, both Fs(1,36) < 1.18, ps > .18.
Table 1

Means and standard deviations of measures by infant-like characteristic condition and experiment

Measure

Characteristic condition

Adult

Infant-like

Experiment 1 (Spain)

    Empathic concerna

3.20 (1.02)

4.50 (1.17)

    N

17

23

Experiment 2 (US)

    Empathic concerna

4.05 (0.97)

4.77 (1.02)

    Protectivenessa

3.15 (1.66)

4.15 (1.46)

    Perceived attractiveness

4.25 (1.07)

4.70 (1.22)

    Perceived youthfulness

4.00 (0.97)

4.65 (1.18)

    Perceived vulnerability

5.28 (1.09)

5.60 (0.75)

    Perceived similarity

3.23 (1.00)

3.54 (1.39)

    N

20

20

Experiment 3 (US)

    Empathic concerna

4.42 (1.19)

5.20 (0.97)

    Perceived age

20.16 (1.34)

19.21 (2.02)

    Perceived vulnerability

6.44 (1.84)

6.42 (1.54)

    Perceived similarity

4.13 (1.48)

3.88 (1.27)

    N

18

18

Notes: For each measure, aindicates that the Adult and Infant-Like means differ at α = .05, two-tailed. All scores can range from 1 to 7, except for ratings of perceived age and scores on the perceived vulnerability and perceived similarity measures in Experiment 3. The latter two measures can range from 1 to 9 and 1 to 8.60, respectively. Standard deviations are in parentheses

Experiment 2

Participants

Participants were 40 female undergraduate students enrolled in introductory psychology courses at the University of Wisconsin Oshkosh. Participants received course extra credit in exchange for their participation. Participants were assigned to one of two face conditions (infant-like vs. adult) using a randomized block procedure.

Procedure

Participants were run individually. The procedure was similar to the one used in Experiment 1 except that participants were told that the articles were being presented to examine student’s reactions to news columns under consideration for the student newspaper. Participants were asked to select a number between 1 and 8 to determine which of eight articles they would read. After participants chose a number, the research assistant brought them a folder with the corresponding number. This procedural element was included to reduce the likelihood that participants would suspect that the specific content of this single article was the focus of the study.

Inside the folder was one of the two versions of the articles (translated into English) used in Experiment 1. The research assistant was unaware of which picture participants received in the article. After they finished reading the article, participants rated a variety of emotional reactions, including the empathic concern items used in Experiment 1 (translated into English) plus the term sympathetic.2

Participants also completed a second questionnaire that included items designed to assess various reactions and perceptions. One question, which was designed to assess participants’ feelings of protectiveness toward the person in the article, asked participants “How protective did you feel toward the person in the article?” (1 = Not at all, 7 = Extremely). Also included in this questionnaire was an item designed to measure perceived attractiveness: “How attractive did the person in the article appear?” (1 = Not at all, 7 = Extremely) and an item designed to measure perceived youthful appearance: “How young did the person in the article appear to you?” (1 = Not at all, 7 = Very young).

Three additional items assessed perceptions of observer similarity to and self-other overlap with the target. Such perceptions have been linked to manipulations that affect the experience of empathic concern in other research (Cialdini et al. 1997; Maner et al. 2002). Two of these items asked participants “How similar to yourself did you perceive the person in the article to be?” (1 = Not at all, 7 = Extremely), and “To what extent do you see yourself and the person in the article as part of the same group?” (1 = Not at all, 7 = Very much). The third item was Aron et al.’s (1991) Inclusion of the Other in the Self measure, which asks participants to circle which of 7 pairs of circles that vary in their degree of overlap best reflects “the extent to which you and the person in the article are connected.” Finally, four items were included to assess perceived vulnerability, which Dijker (2001) claims is an important perceptual antecedent of empathic concern (what he calls pity): “How vulnerable did the person in the article seem to you?” (1 = Not at all, 7 = Extremely vulnerable); “How innocent do you perceive the person in the article to be?”(1 = Not at all, 7 = Extremely innocent); “How helpless was the person in the article?” (1 = Not at all, 7 = Extremely helpless); and “How great is the need of the person in the article?” (1 = Not at all, 7 = Very great).

Results

Participants’ ratings of the six empathic concern adjectives were averaged to form an empathic concern index (α = .72). The items designed to measure perceived similarity and self-other overlap as well as the items designed to measure perceived vulnerability were averaged to form a perceived similarity index (α = .79) and a perceived vulnerability index (α = .72). The means on these indexes along with the means on the measures of protectiveness, perceived attractiveness, and perceived youthfulness by experimental condition are listed in Table 1.

A t-test revealed that participants in the infant-like face condition reported higher levels of empathic concern for the person in need than did participants in the adult face condition, t(38) = 2.28, p = .028, \( \eta _{\text{p}}^{2} = .12 \). Participants in the infant-like face condition also reported feeling more protective toward the person than did participants in the adult face condition, t(38) = 2.02, p = .050, \( \eta _{\text{p}}^{2} = .097 \). Furthermore, the association between the characteristic manipulation and the measure of protectiveness, r(38) = .31, p = .05, is substantially reduced if one controls statistically for the influence of empathic concern, rpartial(37) = .16, p = .33, Δr = .15.

Additional t-tests conducted on the perceptual measures revealed that participants in the infant-like face condition indicated that the face appeared younger than did participants in the adult face condition, t(38) = 1.90, p = .065, \( \eta _{\text{p}}^{2} = .087 \). However, no statistically significant differences in perceived attractiveness, perceived vulnerability or perceived similarity were found between the two characteristic conditions, all ts(38) < 1.24, ps > .22. Finally, Table 2 presents the correlations among all the variables as well as the partial correlations between the characteristic manipulation and the empathic concern index while controlling for each of the perceptual measures. As shown in the table, controlling for each of the perceptual measures has little influence on the strength of association between the characteristic manipulation and empathic concern, .01 < Δr < .05.
Table 2

Correlations among measures by experiment

Measure

Measure

Partial r

Δr

1

2

3

4

5

Experiment 2 (N = 40)

    1. Characteristic

  

    2. Empathic concern

.35*

  

    3. Perceived attractiveness

.20

.25

.31

−.04

    4. Perceived youthfulness

.29

.21

.12

.31

−.04

    5. Perceived vulnerability

.18

.29

.34*

.36*

.32

−.03

    6. Perceived similarity

.13

.40*

.49*

.03

.48*

.33*

−.02

Experiment 3 (N = 38)

    1. Characteristic

  

    2. Empathic concern

.33*

  

    3. Perceived age

−.27

−.18

.29

−.04

    4. Perceived vulnerability

−.01

.03

−.04

.33*

.00

    5. Perceived similarity

−.09

.25

.00

.18

.37*

+.04

Notes: *p < .05, two-tailed. Partial r is the correlation between the Characteristic manipulation (0 = adult, 1 = infant-like) and Empathic Concern measures after controlling for the corresponding perceptual measure. Δr is the change in direction and magnitude of the correlation between the Characteristic manipulation and Empathic Concern measures produced by controlling for the corresponding perceptual measure

Experiment 3

Participants

Participants were 38 undergraduate students (24 women, 14 men) enrolled in various psychology courses at the University of Texas at Tyler who received course extra credit for study participation. Thirteen women and 6 men were assigned to the infant-like voice condition and 11 women and 8 men were assigned to the adult voice condition using randomized blocks.

Procedure

The procedure of Experiment 3 was similar to that of Experiment 2, except that stimuli were presented as pilot broadcasts created by the university student radio station, the need situation was different, a different presentation modality was used to manipulate the target’s infant-like characteristics, and male participants were included in the sample.

All participants received a broadcast describing the same target in the same need situation. The broadcast described a female student who was suffering from a genetic heart defect. The broadcast detailed her struggle living with the condition and her attempt to acquire appropriate medical treatment. Information about the student’s experience was communicated via a male news broadcaster and a brief segment in which the student personally described her situation. Participants listened to a broadcast in which the student’s voice was either unaltered or digitally altered to a higher pitch in order to create a younger sounding voice. This manipulation was accomplished using Audacity recording software to increase the pitch of the target’s voice by 12% in the infant-like voice condition. This pitch increase was selected because it resulted in a believable younger sounding voice. All the other aspects of the altered and unaltered audio files (e.g., dialogue, length, tempo, the announcer’s voice, etc.) were identical. The version of the audio file that participants heard was randomly assigned by the computer so that the research assistant was blind to condition.

After listening to the broadcast, participants were presented with a list of feelings adjectives that included five of the empathic concern adjectives used in Experiment 2.3 Participants also responded to five items designed to assess perceptions of observer similarity to and self-other overlap with the target. Three of the similarity items were those used in Experiment 2. Two additional items were included as well. One of these items asked participants “To what extent would you use the term we to describe yourself and the other participant?” (1 = Not at all, 9 = Extremely), and “How much do you think that you and the person(s) being interviewed have in common?” (1 = Nothing at all, 9 = A great deal). The three items included to measure the perceived vulnerability of the person in need were the vulnerability and need items from Experiment 2 and a third item that asked “How dependent did the person in the broadcast seem to you?” (1 = Not at all, 9 = Extremely dependent). All the perceived similarity and vulnerability items were rated using 9-point scales except for the Inclusion of the Other Scale, which was coded using a 7-point scale. Finally, participants were asked to estimate the age of the person in the broadcast.

Results

Participants’ ratings of the five empathic concern adjectives were averaged to form an index (α = .80). The items designed to measure perceived similarity and self-other overlap as well as those items designed to measure perceived vulnerability were averaged to form a perceived similarity index (α = .83)4 and a perceived vulnerability index (α = .68). The means on these indexes by condition are listed in Table 1.

A 2 (Gender: male vs. female) × 2 (Characteristic: infant-like vs. adult) ANOVA revealed a statistically significant Characteristic main effect on empathic concern: participants presented with the infant-like voice reported higher levels of empathic concern for the person in need than did participants presented with the adult voice, F(1, 34) = 4.33, p = .045, \( \eta _{\text{p}}^{2} = .11 \). Neither the Gender main effect nor the Gender × Characteristic interaction achieved statistical significance, both Fs(1, 34) = .34, ps = .57.

Although a statistically significant characteristic effect on empathic concern emerged, 2 (Gender: male vs. female) × 2 (Characteristic: infant-like vs. adult) ANOVAs conducted on the perceived vulnerability index, perceived similarity index, and age measures revealed no statistically significant main effects or interactions, all Fs(1, 34) < 1.64, ps > .20. Finally, Table 2 presents the correlations among all the variables as well as the partial correlations between the characteristic manipulation and the empathic concern index while controlling for each of the perceptual measures. As shown in the table, controlling for each of the perceptual measures has little influence on the strength of association between the characteristic manipulation and empathic concern, −.04 < Δr < .04.

Discussion

The results of these three experiments provide the first clear experimental evidence that individuals in need who vary in the degree of infant-like characteristics elicit different degrees of parental-like emotional experience in observers. Specifically, across the three experiments, both male and female participants reported higher levels of empathic concern for targets in need who possessed more infant-like characteristics than for targets in need who possessed more adult-like characteristics. In line with the proposition that empathic concern reflects a parental emotion, participants in Experiment 2 who viewed the target with more infant-like facial features reported feeling more protective toward the person than did those who viewed the target with more adult-like facial features. Furthermore, a partial correlation analysis suggested that this effect was mediated by feelings of empathic concern.

Additional results from Experiments 2 and 3 suggest that the infant-like characteristics effect on empathic concern may be relatively direct. Across both experiments, partial correlation analyses revealed that controlling for ratings of perceived target attractiveness, perceived youthfulness and age, perceived target vulnerability, or perceived observer similarity to the target had little influence on the strength of the association between the manipulation of infant-like characteristics and self-reported empathic concern. The infant-like characteristic effect on empathic concern is particularly impressive because it was found using highly controlled stimuli from two different presentation modalities (visual and auditory) in two different American samples. Furthermore, the infant-like face effect was replicated in samples from two different nationalities (America and Spain) using faces from a third (German).

Limitations

In addition to its strengths, it also is important to consider the potential limitations of the present research. One limitation is that only one set of facial stimuli was used in Experiments 1 and 2 and only one set of vocal stimuli was used in Experiment 3. Thus, although the infant-like characteristics effect on empathic concern was replicated across presentation modalities, it is possible that the effect does not generalize within presentation modality. In research concerned with detailing the effects of facial features on social perception, this limitation typically is addressed by utilizing within-subject designs in which participants are presented with multiple faces that vary in the feature or set of features of interest (see Berry and McArthur 1986 for a review). Use of such a design was avoided in the present research because it likely would undermine experimental realism and create high reactivity in participants if used to examine realistic emotional reactions to those in need. Furthermore, a between-subjects manipulation of stimulus sets was avoided to achieve methodological efficiency.

It is important to emphasize that from the outset the purpose of the present research never was to illuminate the minimal set of facial or vocal features necessary to affect empathic concern in observers. Rather, the purpose was to examine the tenability of a causal association between broad conceptual variables (i.e., infant-like characteristics and empathic concern). Specifically, the present research sought to establish a strong test of the association by demonstrating consistency (or a lack thereof) in empirical findings across multiple operationalizations of the independent conceptual variable (faces and voices) rather than demonstrating consistency across multiple examples of a single operationalization of the independent conceptual variable (various sets of faces or voices).

One also may question the extent to which the findings of Experiments 2 and 3 rule out certain perceptual variables as mediating the effect of infant-like characteristics on empathic concern. Controlling for the various perceptual measures produced relatively small changes in the correlation between the infant-like characteristic manipulation and self-reported empathic concern in both Experiments 2 and 3 (−.04 < Δr < .05). This approach to determining the presence of mediation is admittedly somewhat subjective. However, formal tests of mediation using the four-step Baron and Kenny approach or a Sobel’s test (MacKinnnon et al. 2002) also suggest that the infant-like characteristics effect on empathic concern is not mediated by perceived attractiveness, perceived age, perceived similarity, or perceived vulnerability.5 Given the magnitude of the correlations and small changes in correlations depicted in Table 2, it is unlikely that increasing the sample size of Experiments 2 or 3 would yield statistical findings that run contrary to the mediation conclusions drawn from the present analyses.

Implications and future directions

The present research points toward a number of implications and avenues for future research. First, numerous social-psychological studies demonstrate that empathic concern is associated with increased helping of those who elicit it (Batson 1998; Penner et al. 2005). Consideration of these previous findings along with the present ones suggests the hypothesis that the perception of infant-like characteristics among those in need will enhance helping of such individuals, and such helping will be mediated, at least partially, by feelings of empathic concern. Therefore, it may prove worthwhile for future research to test this hypothesis.

Second, not only does previous research demonstrate that empathic concern increases helping behavior, but it also suggests that the motivation responsible for such helping is altruistic (i.e., the goal of the motivational state is to increase the welfare of the other as an end in itself) (Batson 1991; Batson and Shaw 1991). If one assumes that the experience of empathic concern reflects activation of a parental emotional-motivational system that is sensitive to infant-like cues, and one considers previous findings that empathic concern promotes altruistic motivation, then one may draw the following implication: the capacity for human altruism arises from mechanisms originally evolved to promote care of young (Batson et al. 2005; Panksepp 1998; MacLean 1990). This theoretical perspective on the source of empathic concern diverges from many popular cognitive perspectives that claim it arises from perceptual identification, perceived similarity, or cognitive-perceptual merging of the self with the person in need (for reviews of these perspectives, see Batson et al. 1997, 2005). Contrary to these cognitive perspectives, results from Experiments 2 and 3 yield no evidence that various measures of perceived observer similarity to the target in need mediate the infant-like characteristics effect on empathic concern. Consequently, these findings are consistent with Batson et al.’s (2005) conclusion that perceived observer similarity to a target in need is not necessary for the experience of empathic concern.

The present findings also are consistent with ethological claims that infant-like cues act as innate-releasing mechanisms that automatically elicit caregiving responses in adult members of a species (Lorenz 1971). The present results are consistent with this possibility because they reveal that the infant-like characteristics effect on empathic concern is uninfluenced by higher-order cognitive-perceptual processes associated with empathic concern in other research. However, it is possible that infant-like cues enhance empathic concern because situations that elicit this affective experience often involve infant and child targets, and the characteristics of these targets may come to elicit empathic concern as a result of associative learning. Thus, future research will be necessary to discriminate between an innate-releasing account and an associative learning account of the infant-like characteristic effect on empathic concern.

Footnotes
1

The three samples used in these studies were convenience samples and their selection was not based on any theoretical rationale. They were selected to determine if any effects would generalize and to provide increased external validity for any potential findings.

 
2

This term was not used in Experiment 1 because research by Oceja and Jiménez (2007) suggests that the term cannot be translated into Spanish adequately.

 
3

The term softhearted was excluded because the need situation involved a heart condition.

 
4

In Experiment 3, the Inclusion of the Other in the Self Scale was coded using a 7-point scale, whereas the other four perceived similarity items were rated using 9-point scales. Consequently, participants’ scores on the averaged perceived similarity index could range from 1 to 8.60.

 
5

Examination of Table 2 reveals that the data fail to meet Step 2 of the Baron and Kenny approach to testing mediation (i.e., the predictor and potential mediator show a statistically significant association). The results of the non-statistically significant Sobel’s tests can be obtained from the first author.

 

Acknowledgments

The authors extend their gratitude to Martin Gründl for granting permission to use the facial stimuli presented to participants in Experiments 1 and 2. Also, the authors thank Belén López, Tamara Ambrona, Irene Fernández, Mark Scurlock, Eric Peyton, and Chris Buchannon for their assistance in collecting data as well as Paul Silvia and an anonymous reviewer for providing helpful comments on an earlier draft. Data collection for Experiment 1 and a portion of the preparation of this manuscript were supported by Ministry of Education and Science grant SEJ2005-06307/PSIC, which was awarded to Luis Oceja.

Copyright information

© Springer Science+Business Media, LLC 2008