Web Appendix A: Proof of Concept Facebook Study In January 2018, an online petition asking the USDA to allow SNAP funds to be spent on pet food was signed by over 100,000 people (Morris 2018). Networks who shared this news via social media received hundreds of comments from consumers who gave their opinions on this proposed policy. For two such articles shared on social media, we i) coded people’s responses as either being for or against this policy change and ii) labeled respondents as either having a pet or not (codings were made independent of each other). We posited that support for this policy would be predicted by whether or not a commenter showed evidence of having a pet. While a naturalistic study such as this is less controlled, it does serve to capture a spontaneous, real-world illustration of our proposed effect which our experimental studies build on further. Method We analyzed Facebook users’ comments that were made on two articles posted by news sources (NBC affiliate in Madison, WI and Fox affiliate in Cleveland, OH). NBC Madison article. Four hundred and forty-three comments were originally included in the dataset. Eighteen of these comments were removed from analysis because they lacked any meaningful content that could be analyzed (e.g., comment only contained the name of a friend tagged in the post, leaving 425 comments to be analyzed. Comments were coded by MTurk users as having either a favorable or unfavorable opinion of the proposed policy change on a five-point scale (1 = strongly against policy, 5 = strongly supports policy). Coders were randomly assigned to rate 20 of the comments. Each comment received 5 to 10 ratings. The commenters’ Facebook profiles were then explored for evidence of having a pet in their profile picture(s). A priori, profiles that had fewer than ten public pictures were considered uninformative (n = 108). Profiles
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Web Appendix A: Proof of Concept Facebook Study
In January 2018, an online petition asking the USDA to allow SNAP funds to be spent on
pet food was signed by over 100,000 people (Morris 2018). Networks who shared this news via
social media received hundreds of comments from consumers who gave their opinions on this
proposed policy. For two such articles shared on social media, we i) coded people’s responses as
either being for or against this policy change and ii) labeled respondents as either having a pet or
not (codings were made independent of each other). We posited that support for this policy
would be predicted by whether or not a commenter showed evidence of having a pet. While a
naturalistic study such as this is less controlled, it does serve to capture a spontaneous, real-world
illustration of our proposed effect which our experimental studies build on further.
Method
We analyzed Facebook users’ comments that were made on two articles posted by news
sources (NBC affiliate in Madison, WI and Fox affiliate in Cleveland, OH).
NBC Madison article. Four hundred and forty-three comments were originally included
in the dataset. Eighteen of these comments were removed from analysis because they lacked any
meaningful content that could be analyzed (e.g., comment only contained the name of a friend
tagged in the post, leaving 425 comments to be analyzed. Comments were coded by MTurk users
as having either a favorable or unfavorable opinion of the proposed policy change on a five-point
scale (1 = strongly against policy, 5 = strongly supports policy). Coders were randomly assigned
to rate 20 of the comments. Each comment received 5 to 10 ratings. The commenters’ Facebook
profiles were then explored for evidence of having a pet in their profile picture(s). A priori,
profiles that had fewer than ten public pictures were considered uninformative (n = 108). Profiles
with pictures of pets were considered as evidence of having a pet (n = 145), and profiles with
more than ten pictures and no pictures of pets were considered as not having a pet (n = 172).
Fox Cleveland article. Two hundred and twenty-three comments were originally included
in the dataset. One comment was removed from analysis because they lacked any meaningful
content that could be analyzed, leaving 222 comments to be analyzed. MTurk users were
randomly assigned to rate 20 of the comments as having a favorable or unfavorable attitude
toward the policy (1 = strongly against policy, 2 = somewhat against policy, 3 = mixed opinion
of policy, 4 = somewhat supports policy, 5 = strongly supports policy). Each comment received 7
to 10 ratings. To simplify and more objectively define the coding of having a pet, we simply
coded whether the respondent had a picture of a pet specifically in their profile picture (n = 30),
or did not (n = 192). Our coding choices create a conservative test of our hypotheses, insofar as
one can theoretically have a pet or value pets and not post pictures of them on social media. This
also explains why there are seemingly so few people with pets in this dataset compared to the
NBC Wisconsin dataset.
Results and Discussion
Welfare attitudes are often articulated as principled beliefs and a reflection of ideology.
However, we found that the simple fact of whether or not one has a pet was a strong predictor of
policy attitudes. For the NBC article, an ANOVA showed that those with evidence of having a
pet showed significantly more support for the policy (M = 2.70, SD = 1.36) compared to those
showing no evidence of having a pet (M = 2.03, SD = 1.00) or those for whom having a pet was
inconclusive (M = 2.08, SD = 1.21), F (2, 422) = 14.83, p < .001. Said differently, 25.3% of
those identified as having a pet made comments that were on average “somewhat” or “strongly”
supportive of the policy versus only 8.2% of those with no evidence of having a pet (and 10% of
those for who having a pet was inconclusive).
For the Fox article, an independent samples t-test found that those with evidence of
having a pet supported the policy significantly more (M = 3.10, SD = 1.45) than those showing
no evidence of having a pet (M = 2.24, SD = 1.28), t(220) = 3.37, p < .001. Said differently, 40%
of those who had a pet in their profile picture made comments that were on average “somewhat”
or “strongly” supportive of the policy versus only 16.6% of those who did not.
These results provide some preliminary, real-world evidence that people’s attitudes
toward specific welfare policies can vary as a function of egocentric processes. While the nature
of these data are imprecise, they are nevertheless illustrative of the proposed phenomenon.
Web Appendix B: Study 1a Experimental Materials, Pre-test, and Supplementary Analyses
Study 1a Materials
Please RANK the following things according to how you would PRIORITIZE them. That is,
with a limited amount of money, what you would keep in your budget and what would you
remove? Things you would keep in your budget go toward the TOP of the list. Things you would
NOT buy in order to buy other things go toward the BOTTOM of the list:
salty snacks (e.g., pretzels, potato chips)
soda
sugary snack (e.g., candy, candy bars)
frozen processed food
dessert foods (e.g., cake, ice cream)
Now, please rate how much you like each of the following things: (same five items as above
were listed, 1 = not at all, 5 = moderately, 9 = very)
How well do the following statements describe your personality? I see myself as someone who...
…is reserved
…is generally trusting
…tends to be crazy
…is relaxed, handles stress well
…has few artistic interests
…is outgoing, sociable
…tends to find fault with others
…does a thorough job
…gets nervous easily
…has an active imagination
You will next read about a person and answer some questions about your impression of them.
Tim Garrett is 32 years old, and lives in Kansas City Missouri. He has a diploma from a community
college. He has a wife and child that he lives with. He is recently unemployed, and is receiving
welfare from the government. He receives a few hundred dollars a month in the form of cash
assistance as well as Supplemental Nutrition Assistance Program funds (SNAP, i.e., "food
stamps").
Target Profile Image from Minear, Meredith and Denise C. Park (2004), “A Lifespan Database
of Adult Facial Stimuli,” Behavior Research Methods, Instruments, and Computers, 36, 630 –
633).
Image file name: TMWmale22neutral.bmp
When Tim goes grocery shopping with his SNAP (food stamp) funds, he will sometimes buy
some ______________ along with his other groceries [food item varied by condition. Was either
participants most or least preferred item from ranking task].
Manipulation check: How much value do you think Tim will get from buying salty snacks,
considering other ways that he could spend his SNAP funds?
(1 = very little, 5 = a moderate amount, 9 = a lot)
In your opinion, how likely is it that Tim…
(1 = not at all, 5 = moderately, 9 = extremely).
…is irresponsible with his money
…is impulsive
…is easily tempted
…lacks self-control
…violent
…rude
…vulgar
…mean
What was Tim’s occupation? [attention check]
None, was receiving government assistance
None, disability
Restaurant server
Construction
Consumer service
What is your sex? Male, Female
What is your age?
On the scale below please indicate your household’s approximate yearly income before taxes