First Stage 2: Evaluating Messages

The objective of this experiment is to investigate whether naturally occurring messages about children who are confronting serious illnesses can succeed in generating variation in empathic reactions. One message (intervention) is a simple three-sentence description of a social and emotional support program for the children. The second is a fundraising e-mail for the program. The third is a video showing the program in action.

2020-02-21

2020-03-14

Empathic State
After watching the video, all participants will fill out the Emotional Response Scale (e.g., Batson, et al., 1988, 1989, 1991, 1997, 2007). This scale includes a six-item measure of Empathic State.

The Emotional Response Scale has been used in numerous experiments by Batson and colleagues to measure Empathic State after administering a perspective-taking instruction set (“Imagine-other/Remain-objective”) intended to increase/decrease empathy. The scale is made up of 18 emotions (examples: sympathetic, compassionate, alarmed, sad). A participant self-rates (on a scale from 1 to 7) how much they experienced each emotion after reading the three sentences/reading the e-mail/watching the video.

Distress State and Negative State
These are eight- and four-item sub-scales from the Emotional Response Scale.

Principle of Care State
These is an eight-item scale intended to measure a state of thinking about moral principles to help other people. It parallels Bekkers and Ottoni-Wilhelm’s (2016) dispositional measure.

Inclusion of Other in the Self
This is a single-item measurement of how much overlap the participant sees between her/himself and a child in the hospital. The measurement was developed by Aron, Aron, and Smollan (1992).

Interpersonal Reactivity Index
The IRI has been extensively used to have participants self-rate their dispositional tendencies to react to situations (example: “I often have tender, concerned feelings for people less fortunate than me”) on a scale from 1 (does not describe me very well) to 5 (describes me very well). Although theoretically IRI measurements should not respond to temporary changes in state induced by the experimental conditions (because the IRI is a dispositional measure), it may be that self-ratings on dispositional measures nevertheless can be altered by conditions intended to evoke state empathy.


Principle of Care disposition
This is an eight-item scale that measures a participant’s endorsement of a moral principle that one should help people in need (Bekkers and Ottoni-Wilhelm, 2016).

Attention-Check and Manipulation-Check Questions

Distress State and Negative State
Although the conditions are intended to increase Empathic State, the conditions may unintentionally increase/decrease other emotions, in particular the distress one might feel, and/or sadness one might feel, at seeing messages about children who are ill. The Distress State and Negative State scales are intended to detect this, should it occur.

Principle of Care State
Likewise, the conditions are not intended to increase thinking about moral principles to help other people. The PoC State scale is included to check this.

Inclusion of Other in the Self
Some psychologists (e.g., Cialdini et al., 1997) have argued that conditions to increase empathic state instead increase the participant’s experience of oneness with people in the message (in our case, a child in hospital): “they perceive more of themselves in the other” (p. 483). The Inclusion of Other in the Self measurement is intended to measure the extent to which this does, or does not, happen.

Interpersonal Reactivity Index
The IRI has three seven-item sub-scales measuring dispositional tendencies to react with Empathic Concern, Personal Distress, and cognitive Perspective-Taking (Davis, 1983).

As mentioned, the IRI is a dispositional measure, so the responses to its items are not conceptually supposed to be influenced by emotional state. Nevertheless, the conditions may raise the salience of the seven EC-items as the participant reads them and cause measured ECi to move. Therefore, we will check to see if the Empathic Concern items register any variation across conditions.

If the primary outcome Empathic State varies as a function of conditions as planned, it will be the preferred outcome measure. However, if Empathic State does not vary as much as planned, and Empathic Concern improves the measurement of changes evoked by the conditions, we will combine the measures. Details are in the Measurement and Analysis Plan.

Personal Distress provides measurement of whether the conditions change measurement of another emotional response disposition. For example, if the conditions change (dispositional) Empathic Concern, but do not change (dispositional) Personal Distress, that would be an indication that the Empathic Concern measure, although intended to be dispositional, did register a change in empathic state, whereas another emotion-disposition measure did not.

The conditions may also increase Perspective Taking, to the extent that they increase a cognitive (rather than emotional) empathic reaction, and the extent to which the dispositional PT items pick up such a change in state. In this case, because of the close theoretical relationship between affective and cognitive empathic reactions, we will combine Perspective Taking into the outcome measure. Details are in the Measurement and Analysis Plan.

Principle of care - dispositional
Measurement of the (dispositional) endorsement of a moral principle that one should help people in need is at a more advanced stage (Ottoni-Wilhelm and Bekkers, 2010; Bekkers and Ottoni-Wilhelm, 2016) than are state measures of the principle. Therefore we ae including measurement of the dispositional PoC to have a second method to detect any change in conscious thinking about moral principles to help.

Attention-Check
An attention-check will be used to detect participants who are not reading the questions, but simply clicking through the questions to quickly get to the end of the session.

Manipulation-Check questions
All participants (even those randomized into the control group) are asked whether they saw a video, an e-mail, or only a brief description of the children’s hospital.
Participants randomized into the video condition are asked the approximate age of the child featured in the video. Participants randomized into the e-mail condition are asked the approximate age of the child featured in the e-mail.

The experimental design is between-subjects, 3 x 1.

The participants will be Amazon Mechanical Turk workers invited to complete a “Human Intelligence Task” (HIT) called “Evaluating Messages”. The inclusion criteria are: (1) U.S. citizen, (2) 18 years or older, (3) done at least one previous HIT, and (4) have completed 95% or more of their previous HITs.

Approximately one-third of the participants will be randomly assigned to each of the three conditions: control, e-mail, and video.


After watching the conditions, all participants fill out the Emotional Response Scale, the Principle of Care state measurement, the Inclusion of Other in the Self, the Interpersonal Reactivity Index, the Attention-Check question, the Principle of Care dispositional scale, gender, race/ethnicity, and the Manipulation-Check questions.

Here are the three conditions in more detail:

1. Control condition:
RILEY HOSPITAL FOR CHILDREN

Riley Hospital for Children located in Indianapolis, IN is a national leader in providing child- and family-centered care to children.

Child- and family-centered care creates a supportive environment that leads to better health outcomes.

To provide a supportive environment for healing, Riley has a wish list containing books, toys, and craft supplies that a child would love to have.


2. E-mail condition.

Same message as above plus: We would like you to read an email produced by Riley.

The next screen displays the e-mail.


3. Video condition.

Same message as above plus: We would like you to watch a video produced by Riley.

The next screen displays the video.

The randomization is by computer.

Individual (at the level of the individual MTurk worker).

No

n.a.

N = 270 MTurk workers. Because the objective is to test how much empathy can be increased by the conditions, we have a plan to increase the sample size should N = 270 yield insufficient precision to meet our objective. See “Power calculation” below.

N = 90 in each of the three conditions.

Based on previous measurements of the standard deviation of our primary outcome Empathic State (sES = 1.25), we are powered at 80% to detect a difference between conditions of .42sES. We are also 80% powered to detect a .42s difference in our back-up primary outcome–Empathic Concern (the fact that under our projections about standard deviations we are powered to detect exactly .42s in both outcomes is a coincidence). If the standard deviation of Empathic State in the experiment turns out to be much larger than previous measurements, so that with N = 90 in each of the conditions we are unable to say that we have significantly increased the difference in Empathic State across the conditions with an F-stat >= 10, we will increase the number of observations in each condition. Again, the objective of this experiment is to investigate whether the manipulations to increase empathy can succeed, so if the manipulations appear to succeed based on the magnitude of the point estimates, but are imprecisely estimated, the objective will be to find the necessary N to achieve precision. We are monetarily budgeted to increase the number of observations to a total N = 900. In addition, Objective 4 is to investigate whether the manipulations to increase empathy can succeed among women, and also separately among men. If the MTurk workers responding to our HIT are (say) disproportionally men, we will increase the sample size of women to achieve Objective 4. See attachment “fs-Pre-reg-013-v01a-PowerAnalysis-MTurk02.pdf” for further details.