Are Empathy and Concern Psychologically Distinct?

Supplemental Materials

Are Empathy and Concern Psychologically Distinct?

by M. R. Jordan et al., 2016, Emotion

Supplementary Online Materials (SOM-R)

Section 1: Factor Analysis

Scale Items

Interpersonal Reactivity Index

Novel subscales

Factor analysis robustness

Replication of Previous IRI Factor Solution

Section 2: Public Goods Game

Instructions

Comprehension Questions

Further Analyses

Section 3: Identifiable Victim

Stimuli

Identifiable Victim Condition

Statistics Condition

Further Analyses

Interaction Interpretation

Regressions

Mediation Analyses

Path Models

Section 4: The Subscales and Demographics

Subscale relationships

Correlation matrix

Mutlticollinearity diagnostics

Section 1: Factor Analysis

Scale Items

Below are the six subscales used in each of the studies presented in the main text. For each study, we fully randomized the presentation of these items. Items were presented with six on each page for a total of seven pages of questions and participants were asked to rate the degree to which they agreed with each item on a five-point scale from “strongly disagree” to “strongly agree”. Items were no more likely to appear with other items from their subscale than with items from any other subscale.

Interpersonal Reactivity Index

Perspective Taking subscale

I sometimes find it difficult to see things from the "other guy's" point of view.

I try to look at everybody's side of a disagreement before I make a decision.

I sometimes try to understand my friends better by imagining how things look from their perspective.

If I'm sure I'm right about something, I don't waste much time listening to other people’s arguments.

I believe that there are two sides to every question and try to look at them both.

When I'm upset at someone, I usually try to "put myself in his shoes" for a while.

Before criticizing somebody, I try to imagine how I would feel if I were in their place.

Fantasy subscale

I really get involved with the feelings of the characters in a novel.

I am usually objective when I watch a movie or play, and I don't often get completely caught up in it.

Becoming extremely involved in a good book or movie is somewhat rare for me.

After seeing a play or movie, I have felt as though I were one of the characters.

When I watch a good movie, I can very easily put myself in the place of a leading character.

When I am reading an interesting story or novel, I imagine how I would feel if the events in the story were happening to me.

I daydream and fantasize, with some regularity, about things that might happen to me.

Concern (Empathic Concern) subscale

I often have tender, concerned feelings for people less fortunate than me.

Sometimes I don’t feel very sorry for other people when they are having problems.

When I see someone being taken advantage of, I feel kind of protective towards them.

Other people’s misfortunes do not usually disturb me a great deal.

When I see someone being treated unfairly, I sometimes don’t feel very much pity for them.

I am often quite touched by things that I see happen.

I would describe myself as a pretty soft-hearted person.

Personal Distress subscale

In emergency situations, I feel apprehensive and ill-at-ease.

I sometimes feel helpless when I am in the middle of a very emotional situation.

When I see someone get hurt, I tend to remain calm.

Being in a tense emotional situation scares me.

I am usually pretty effective in dealing with emergencies.

I tend to lose control during emergencies.

When I see someone who badly needs help in an emergency, I go to pieces.

Novel subscales

Empathy subscale

If I see someone who is excited, I will feel excited myself.

I sometimes find myself feeling the emotions of the people around me, even if I don’t try to feel what they’re feeling.

If I’m watching a movie and a character injures their leg, I will feel pain in my leg.

If I hear a story in which someone is scared, I will imagine how scared I would be in that situation and begin to feel scared myself.

If I hear an awkward story about someone else, I might feel a little embarrassed.

I can’t watch shows in which an animal is being hunted one another because I feel nervous as if I am being hunted.

If I see someone fidgeting, I’ll start feeling anxious too.

Behavioral Contagion subscale

If I see someone else yawn, I am also likely to yawn.

If I see someone vomit, I will gag.

I catch myself crossing my arms or legs just like the person I’m talking to.

If I see a video of a baby smiling, I find myself smiling.

If I see someone suddenly looking away, I’ll automatically look in the direction they are looking.

If I’m watching someone walking on a balance beam, I will lean when they lean.

If I’m having a conversation with someone and they scratch their nose, I will also scratch my nose.

Factor analysis robustness

The Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy for the individual items was 0.919, which is considered “marvelous” and indicates that factor analysis is appropriate (Kaiser). At the subscale-level, the KMO measure of sampling adequacy at the subscale level was 0.703, which is considered “middling” and suggests factor analysis is also appropriate for subscale-composite scores (Kaiser).

Because we had such high within-scale reliability and KMO measures for the subscales, we chose to perform a factor analysis on the standardized composite measures of each subscale. However, we achieved the same qualitative results performing an item-level factor analysis, shown in the table at the end of this section of the SOM-R.

The first two factors were the only two that yielded eigenvalues near to or greater than 1 (factor 1 eigenvalue=2.14, factor 2 eigenvalue=0.81) and these two factors accounted for all of the variance. The table below shows the loadings for the subscale composites.

Table S1. The factor loadings for subscale composite scores after principal factor analysis and promax rotation.

We used the oblique promax rotation because it allows factors to be correlated. However, we get the same qualitative result when we force orthogonal factors by using the varimax rotation. The table below shows the factor loadings that resulted from the orthogonal varimax rotation.

Table S2. The factor loadings for subscale composite scores after principal factor analysis and varimax rotation, which forces orthogonal factors.

We also achieve the same factor solution using the principal components factor, iterative principal factor and maximum likelihood factor methods. Below is the result from the maximum likelihood factor analysis and promax rotation.

Table S3. The factor loadings for subscale composite scores after maximum likelihood factor analysis and promax rotation.

In addition to being robust to a variety of factor and rotation methods, the factor structure described in the main text holds across a number of sub-populations. We created dummy variables to represent median splits on our demographic variables and we conducted the same factor analysis on each half of the dataset for all seven demographic variables. Those analyses are shown below.

Table S4. The factor loadings for subscale composite scores after principal factor analysis and promax rotation, and across genders.

Table S5. The factor loadings for subscale composite scores after principal factor analysis and promax rotation, and across old versus young age groups.

Table S6. The factor loadings for subscale composite scores after principal factor analysis and promax rotation, and across high versus low income categories.

Table S7. The factor loadings for subscale composite scores after principal factor analysis and promax rotation, and across high versus low belief in god.

Table S8. The factor loadings for subscale composite scores after principal factor analysis and promax rotation, and across high versus low education level.

Table S9. The factor loadings for subscale composite scores after principal factor analysis and promax rotation, and across high versus low degree to which they can trust others in their daily lives.

Table S10. The factor loadings for subscale composite scores after principal factor analysis and promax rotation, and across social liberal versus conservativeness.

Across all seven demographic splits, we observe a qualitatively similar factor structure. In fact, the only subscale that shows any movement is the fantasy subscale, which always hovers just above or just below the critical value of 0.4. The Contagion and Other-regarding factors are remarkably stable across a wide age range and a number of other demographic variables that typically do impact social cognitive and emotional abilities like gender—women are typically more social attuned (Baron-Cohen)—and social conservatism—social conservatives tend to be less caring (Hirsh, DeYoung, Xu, & Peterson). These facts are borne out in our data as men (M=0.983, SD=0.846) are less concerned for others than women (M=1.365, SD=0.814), t(478)=5.005, p<0.001. In addition social liberals (M=1.212, SD=0.861) scored marginally higher on concern than social conservatives (M=1.067, SD=0.834), t(478)=1.826, p=0.069. Despite these differences, the factor structure for each of these groups remains the same.

Table S11 below shows that, with just one exception, Empathy and Concern items always load on separate factors.

Table S11. An item level factor analysis reveals the same general structure and almost no crossover between empathy and concern items.

Replication of Previous IRI Factor Solution

Below is a table that shows our replication of Pulos and colleagues’ IRI factor solution (Pulos et al.) using the principal factor method and promax rotation. This is the same factor structure they report, but using the iterated principal factor method, we obtain a different result.

Below we show the results from an iterated principal factor analysis with promax rotation. The results are qualitatively similar, and make sense in light of our results using the six subscale battery, but also demonstrate the instability of previous IRI factor analyses that did not include an empathy measure.

Section 2: Public Goods Game

Instructions

Below are the verbatim stimuli, as presented to participants in Study 2.

You have been randomly assigned to interact with 3 other people. All of you receive this sameset of instructions. You cannot participate in this study more than once.
Each person in your group is given 40 cents for this interaction (in addition to the 50 cents youreceived already for participating).
You each decide how much of your 40 cents to keep for yourself, and how much (if any) tocontribute to the group’s common project.
All money contributed to the common project is doubled, and then split evenly among the 4group members.
Thus for every 2 cents you contribute,you get 1 cent back: so no matter what the other group members contribute, you personally lose money on contributing, but it benefits the group as a whole.

For example:
* If everyone contributes all of their 40 cents, everyone’s money will double: each of you will earn80 cents.

* If everyone else contributes their 40 cents, while you keep your 40 cents, you will earn 100 cents, while the others will earn only 60 cents.

The other peopleare REAL and will really make a decision – there is no deception in this study. Once you and the other people have chosen how much to contribute, the interaction is over. Neither you nor the other people receive any bonus other than what comes out of this interaction.

Comprehension Questions

Participants were also asked to complete the following tow comprehension questions in order to ensure they had fully understood the logic of the game and they payoffs associated with making different contributions.

1. What level of contribution earns the most money for the group as a whole?

2. What level of contribution earns the most money for you personally?

Answers: 1) 40 cents; 2) 0 cents.

Only subjects who correctly answered both of these comprehension questions were included in the analysis reported in the main text.

Further Analyses

In Table S1 below, we show regressions for PGG contributions on each of the six subscales among only those participants who passed the comprehension check. Concern for others is the only robust predictor of cooperation and accounts for 7.4% of the variation in contributions on its own, which the other five subscales combined account for only an additional 1.2% of the variance.

(1) / (2) / (3) / (4) / (5) / (6) / (7) / (8)
VARIABLES / Fraction Contributed to Public Good
Concern / 0.144*** / 0.129** / 0.134*
(0.0402) / (0.0432) / (0.0547)
Empathy / 0.114* / 0.0536 / 0.0687
(0.0531) / (0.0557) / (0.0783)
Perspective Taking / 0.101* / 0.0124
(0.0454) / (0.0564)
Fantasy / 0.0375 / -0.0491
(0.0416) / (0.0487)
Personal Distress / 0.0401 / 0.000928
(0.0411) / (0.0451)
Behavioral Contagion / 0.0932 / 0.0164
(0.0539) / (0.0685)
Constant / 0.346*** / 0.222 / 0.317** / 0.452*** / 0.489*** / 0.239 / 0.226 / 0.193
(0.0588) / (0.141) / (0.0966) / (0.0808) / (0.0457) / (0.165) / (0.138) / (0.188)
Observations / 164 / 164 / 164 / 164 / 164 / 164 / 164 / 164
R-squared / 0.074 / 0.028 / 0.030 / 0.005 / 0.006 / 0.018 / 0.079 / 0.086
Standard errors in parentheses
*** p<0.001, ** p<0.01, * p<0.05

In Table S2 below, we show that same series of regressions as is in Table S1, but including all participants, including those who failed comprehension questions. The pattern of results is very similar, with concern being the most robust predictor. However, empathy appears to be slightly more predictive when we include those who failed comprehension questions than when we exclude such participants.

(1) / (2) / (3) / (4) / (5) / (6) / (7) / (8)
VARIABLES / Public Goods Game Contribution
Concern / 0.113** / 0.0885* / 0.102*
(0.0373) / (0.0396) / (0.0501)
Empathy / 0.130** / 0.0906 / 0.114
(0.0474) / (0.0500) / (0.0709)
Perspective Taking / 0.0740 / 0.00543
(0.0412) / (0.0517)
Fantasy / 0.0218 / -0.0652
(0.0386) / (0.0449)
Personal Distress / 0.0485 / 0.000920
(0.0377) / (0.0423)
Behavioral Contagion / 0.100* / 0.0108
(0.0497) / (0.0647)
Constant / 0.398*** / 0.197 / 0.388*** / 0.495*** / 0.499*** / 0.233 / 0.193 / 0.187
(0.0545) / (0.127) / (0.0866) / (0.0744) / (0.0417) / (0.152) / (0.125) / (0.173)
Observations / 193 / 193 / 193 / 193 / 193 / 193 / 193 / 193
R-squared / 0.046 / 0.038 / 0.017 / 0.002 / 0.009 / 0.021 / 0.062 / 0.073
Standard errors in parentheses
*** p<0.001, ** p<0.01, * p<0.05

Section 3: Identifiable Victim

Stimuli

Below are the verbatim stimuli, as presented to the participants in Study 3.

Identifiable Victim Condition

Chimwemwe, a 13-year-old girl from Malawi, is desperately poor and faces a threat of severe hunger or even starvation. Her life will be changed for the better as a result of your financial gift. With your support, and the support of other caring sponsors, Save the Children will work with Chimwemwe's family and other members of the community to help feed her, provide her with education, as well as basic medical care and hygiene education.
[An image of Chimwewe was shown here]

In addition to your show-up payment, we are now granting you a 30 cent bonus. You can keep as much of this 30 cent bonus as you'd like or give as as much as you'd like to Save the Children, a charity that will help Chimwemwe.
Please indicate below how much of your bonus you would like todonate.(You will receive as a bonus whatever you do not donate.)

Statistics Condition

• Serious food shortages in Malawi are affecting more than 1.6 million people.
• In Zambia, severe rainfall deficits have resulted in a 42% drop in maize production from 2000. As a result, an estimated 3 million Zambians face hunger.
• Four million Angolans - one third of the population - have been forced to flee their homes.
• More than 11 million people in Ethiopia need immediate food assistance.

​
In addition to your show-up payment, we are now granting you a 30 cent bonus. You can keep as much of this 30 cent bonus as you'd like or give as much as you'd like to Save the Children, a charity that will help address the causes above.
Please indicate below how much of your bonus you would like todonate.(You will receive as a bonus whatever you do not donate.)

FurtherAnalyses

Interaction Interpretation

Although we did not replicate the identifiable victim effect, we did analyze our data to see if there were any conditions under which we did find a significant difference between donations in the identifiable victim condition and the statistics condition. To do this, we ran regressions looking at the effects of condition among participants who were high or low on concern and high or low on empathy, where high and low were defined as one standard deviation above and below the mean, respectively.

High concern / Low concern
High empathy / b=-0.069, CI [-0.234, 0.095], p=0.407 / b=0.215, CI [-0.013, 0.444], p=0.065
Low empathy / b=-0.230, CI [-0.448, -0.013], p=0.038 / b=0.055, CI [-0.115, 0.224] p=0.527

The coefficients reported above are the change from thestatistics condition to the identifiable victim condition; that is, negative coefficients mean greater donations in the identifiable victim condition and vice versa. The largest effects we find are in the high concern/low empathy population (who replicate the original effect) and the high empathy/low concern population (who trend in the opposite direction).

Regressions

In the main text, we referred to having found similar results using the extracted factors as when we used the empathy and concern subscales. In Table S3 below, you can see that the results using the factors are very similar, the only notable difference being a lack of condition by Other-regarding interaction.

(1) / (2) / (3) / (4) / (5) / (6)
VARIABLES / Fraction Donated
Condition / -0.0457 / -0.0451 / -0.0342 / -0.0284
(0.0570) / (0.0573) / (0.0563) / (0.0560)
Contagion factor / -0.0158 / -0.0315 / -0.101*
(0.0318) / (0.0419) / (0.0470)
Condition X Contagion / 0.0364 / 0.0593
(0.0645) / (0.0708)
Other-regarding factor / 0.0987** / 0.112* / 0.175**
(0.0344) / (0.0519) / (0.0594)
Condition X Other-regarding / -0.0266 / -0.0727
(0.0697) / (0.0783)
Constant / 0.428*** / 0.359*** / 0.356*** / 0.382*** / 0.372*** / 0.365***
(0.0905) / (0.0286) / (0.0280) / (0.0407) / (0.0402) / (0.0400)
Observations / 192 / 192 / 192 / 192 / 192 / 192
R-squared / 0.003 / 0.001 / 0.042 / 0.006 / 0.044 / 0.070
Standard errors in parentheses
*** p<0.001, ** p<0.01, * p<0.05

In Table S4 below, we show the bivariate and multivariate relationships between the six subscales and donations.

(1) / (2) / (3) / (4) / (5) / (6) / (7)
VARIABLES / Fraction Donated
Concern / 0.146*** / 0.200***
(0.0329) / (0.0414)
Empathy / -0.0286 / -0.106*
(0.0391) / (0.0575)
Perspective Taking / 0.0635* / -0.0168
(0.0365) / (0.0444)
Fantasy / 0.00437 / -0.0342
(0.0336) / (0.0390)
Personal Distress / -0.0410 / -0.0129
(0.0325) / (0.0374)
Behavioral Contagion / -0.00855 / 0.0306
(0.0400) / (0.0511)
Constant / 0.187*** / 0.434*** / 0.232*** / 0.352*** / 0.390*** / 0.384*** / 0.412***
(0.0475) / (0.106) / (0.0785) / (0.0639) / (0.0373) / (0.121) / (0.131)
Observations / 192 / 192 / 192 / 192 / 192 / 192 / 192
R-squared / 0.094 / 0.003 / 0.016 / 0.000 / 0.008 / 0.000 / 0.136
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

Because the crucial analysis centered on the relationship betweenempathy and donations and concern and donations, we checked to see if these two raw correlations were significantly different from one another. A Steiger’s Z test revealed that the correlation between concern and donations (r = 0.306) was significantly larger than the correlations between empathy and donations (r = -0.053), Steiger’s Z = 4.643, p < 0.001.

Mediation Analyses

In the main text, we described a mediation analysis in which the way participants felt about the charitable cause mediated the relationship between concern and donations. In this section of the supplement, we want to add detail to that analysis and show that other models do not account for the data as well.