Data Center Opinion Project - Study 4: Comparing NIMBYism infrastructure projects

We recruited a nationally representative sample of N = 800 from Prolific. Participants were randomly assigned to one of four conditions and evaluated the prospect of a facility being built in their community:

  • Data centers
  • Landfills
  • Distribution warehouses
  • Nuclear Power Plants

Manipulation

Imagine a company is interested in building a [data center / landfill / nuclear power plant / distribution warehouse] within 10 miles of your home to [support AI technology / manage household and commercial waste from the surrounding region / generate electricity for the regional power grid / store and ship products to consumers and businesses]

Would you favor or oppose the construction of this new ${e://Field/projectType} in your local area?

Measures

Price for support

Those who did not report support for the local project build responded to the following:

Imagine a company is interested in building a ${e://Field/projectType} in your community. They are willing to negotiate terms with the local residents. They are in talks with your local government to offer a one-time payment to all residents within a 10-mile radius of the newly constructed ${e://Field/projectType}.

You indicated that you ${q://QID1220560088/ChoiceGroup/SelectedChoices} ${e://Field/about}the construction of a ${e://Field/projectType} in your local area to ${e://Field/projectRole}. Under what conditions - in terms of a one-time payment - would you be willing to say that you support the construction of this ${e://Field/projectType} in your community?

Indicate the dollar amount that the company should offer to local residents to gain your support.

Free response

You indicated that you ${q://QID1220560088/ChoiceGroup/SelectedChoices} ${e://Field/about}the construction of a ${e://Field/projectType} in your local area to ${e://Field/projectRole}.

Write down some thoughts you had as you arrived at your judgment.

Benefit

Consider there is a new ${e://Field/projectType} being built within 10 miles of your home. How much direct benefit do you think it would provide to each of the following? [no benefit/slight benefit/moderate benefit/great benefit]

  • You personally
  • Your neighborhood
  • Your town/city
  • Your state
  • The country

Benefit - free response

Those that reported anything other than “no benefit” to the above question were then given an opportunity to expand on their thinking for that particular beneficiary.

Importance

How important is it that the country has more [project type]s? [not at all important/somewhat important/very important]

Perceived Opportunism

adapted from

  • perceived opportunism: Hsieh, P. L. (2013). Perceived opportunism (PO) in e‐return service encounters. Managing Service Quality: An International Journal, 23(2), 96-110.

  • benevolence: Mayer, R. C., & Davis, J. H. (1999). The effect of the performance appraisal system on trust for management: A field quasi-experiment. Journal of applied psychology, 84(1), 123.

These questions are about the company that is interested in building a ${e://Field/projectType} in your community. Think about the company or the type of company that came to mind when you thought about the hypothetical ${e://Field/projectType}.

How much would you agree or disagree with the following statements about that type of company? [strongly disagree/disagree/neither or not sure/agree/strongly agree]

  • This is the type of company that would alter facts slightly in order to get what it needs
  • This is the type of company that would promise to do things without actually doing them later
  • This type of company would not knowingly do anything to hurt me or the local people
  • This type of company fails to provide the support it should
  • This type of company breaks formal or informal agreements to its own benefit
  • This type of company is honest

Trust in institutions

Below is a list of institutions in American society. Rate how much trust you have in each one. [strongly distrust/somewhat distrust/not sure/somewhat trust/strongly trust]

  • Large technology companies
  • Big business
  • The federal government
  • Your local government

AI use

How often do you use AI? [never/less than once a month/a few times a month/daily]

show code
dcOpinion_study4 = merge(read.csv('study 4/dataCenterOpinions_study4.csv'), read.csv('study 4/prolificDemographics_DCopinion_study4.csv'), by='PROLIFIC_PID') |> 
  filter(Status!='RETURNED' & consent==1) |> 
  mutate(condition= factor(condition, levels=c('dataCenter','landfill','nuclearPowerPlant','warehouse')),
          supporter_bin = factor(ifelse(sentiment > 0, "Supporter", "Non-supporter")))


nice_table(dcOpinion_study4 %>% group_by(condition)  %>% count(), title = 'N per condition')

N per condition

condition

n

dataCenter

206

landfill

203

nuclearPowerPlant

203

warehouse

200

show code
ggerrorplot(data=dcOpinion_study4, color='condition', x = 'condition', y = 'sentiment', title='Support for local facility. Full scale -2 to 2')

show code
benefits_long = dcOpinion_study4 |> select(condition,  benefit_self, benefit_neighborhood, benefit_town, benefit_state, benefit_country) |> 
  pivot_longer(
    cols = c(benefit_self, benefit_neighborhood, benefit_town, benefit_state, benefit_country),
    names_to = "benefit_type",
    names_prefix = "benefit_",
    values_to = "benefit")  

ggerrorplot(data=benefits_long, color='benefit_type', x ='condition', y = 'benefit', position=position_dodge(0), title='Perceived benefits of facility. Full scale 0 - 3')

show code
nice_table(dcOpinion_study4 %>% filter(!is.na(priceOnSupport_1)) |> group_by(condition) %>% summarise(median=median(priceOnSupport_1)/1000, min = min(priceOnSupport_1), max = max(priceOnSupport_1)/1000), title = "Payment required for support, in thousands. (minimum reported in raw)")

Payment required for support, in thousands. (minimum reported in raw)

condition

median

min

max

dataCenter

15.00

0.00

10,000,000,000,000,017,583,607,936,365,707,443,642,779,759,642,417,653,467,751,731,161,425,031,252,187,748,907,155,351,519,777,830,721,715,057,328,371,269,632.00

landfill

10.00

0.00

10,000,000.00

nuclearPowerPlant

100.00

0.00

90,000,000,000,000,006,360,851,714,686,920,526,135,296.00

warehouse

2.85

0.00

5,000.00

show code
ggerrorplot(data=subset(dcOpinion_study4, priceOnSupport_1 < 1000000000), color='condition', x ='condition', y = 'priceOnSupport_1', title='Payment required for support. Cutoff: $1 billion')

show code
ggplot(subset(dcOpinion_study4, priceOnSupport_1 < 1000000) , aes(x=priceOnSupport_1/1000, color=condition, fill=condition))+
  geom_density( alpha=0.4)+
  geom_vline(aes(xintercept=mean(priceOnSupport_1)/1000),linetype="dashed", linewidth=0.5)+
  labs(x="Payment required for support, in thousands.", title="Cutoff: $1 million")+
  theme_classic()

Free responses

Most common words by condition; then a selection of words and the rate at which they appear by condition.

show code
library('stringr'); library('tidytext')

dcOpinion_study4 %>%
  group_by(condition) %>% 
  select(freeResponse) %>% 
  unnest_tokens(word, freeResponse) %>%
  anti_join(stop_words) %>%
  count(word, sort = TRUE) %>% 
  slice(1:7) |> print(n = 28)
Adding missing grouping variables: `condition`
Joining with `by = join_by(word)`
# A tibble: 28 × 3
# Groups:   condition [4]
   condition         word             n
   <fct>             <chr>        <int>
 1 dataCenter        data           152
 2 dataCenter        centers         85
 3 dataCenter        water           81
 4 dataCenter        center          71
 5 dataCenter        ai              43
 6 dataCenter        lot             31
 7 dataCenter        environment     30
 8 landfill          landfill        85
 9 landfill          smell           57
10 landfill          miles           38
11 landfill          close           37
12 landfill          live            31
13 landfill          10              30
14 landfill          bad             25
15 nuclearPowerPlant nuclear        142
16 nuclearPowerPlant power          133
17 nuclearPowerPlant plant           63
18 nuclearPowerPlant energy          54
19 nuclearPowerPlant close           45
20 nuclearPowerPlant live            28
21 nuclearPowerPlant safe            27
22 warehouse         jobs            64
23 warehouse         warehouse       57
24 warehouse         traffic         46
25 warehouse         local           44
26 warehouse         distribution    42
27 warehouse         miles           32
28 warehouse         close           27
show code
#in all other conditions, people are talking about whether 10 miles is close enough to feel the effects of the facility. In data center evaluation, we don't see that as much. Suggests that people are cued into thinking on the abstract when thinking about data centers, and are evaluating the prospect of AI and data centers in general, and not the immediate, concrete facility per se

dcOpinion_study4 %>%
    group_by(condition) %>%
  summarise(
    data = mean(str_detect(str_to_lower(freeResponse), "data")) * 100,
    center = mean(str_detect(str_to_lower(freeResponse), "center")) * 100,
    jobs = mean(str_detect(str_to_lower(freeResponse), "jobs")) * 100,
    environment = mean(str_detect(str_to_lower(freeResponse), "environment")) * 100,
    nature = mean(str_detect(str_to_lower(freeResponse), "natur")) * 100,
    pollution = mean(str_detect(str_to_lower(freeResponse), "pollut")) * 100,
    noise = mean(str_detect(str_to_lower(freeResponse), "nois")) * 100) 
# A tibble: 4 × 8
  condition          data center   jobs environment nature pollution  noise
  <fct>             <dbl>  <dbl>  <dbl>       <dbl>  <dbl>     <dbl>  <dbl>
1 dataCenter        62.6  61.2   14.6          20.4  1.94      13.6  16.5  
2 landfill           0     0.493  0.493        12.3  2.46      10.8   4.43 
3 nuclearPowerPlant  2.46  2.46   6.40         11.3  0.985      2.96  0.985
4 warehouse          2     6.5   31.5           6.5  1.5        5.5  11.5  

Mediation or moderation