Spiritual Identification Affects Political Attitutes

Spirituality and Euthanasia Attitudes

We investigate the relationship between spirituality and attitudes to Euthanasia.

The New Zealand Attitudes and Values Study contains the following Questions, to which participants respond on a 1-7 scale from lowest to highest level of acceptance:

Suppose a person has a painful incurable disease. Do you think that doctors should be allowed by law to end the patient’s life if the patient requests it? (scale 1-7)

I identify as a spiritual person (scale 1-7)

There is only one year in which both of these questions co-occur: 2016.

The New Zealand Attitudes and Values Study also contains a question about religious identification:

Do you identify with a religion and/or spiritual group? (Yes/No)

And for those who indicate that they identify with a religion or spiritual group

How important is your religion to how you see yourself? (scale 1-7)

Our sample consists of \[N = 19805\] participants for which there was complete information.

Show code

# read data
df <- readRDS(here::here("data", "df"))
# data wrangling
dt1<- df %>%
  filter(YearMeasured ==1) %>%
  filter(Wave == 2016 ) %>%
  droplevels() %>%
  arrange(Wave) %>%
  mutate(SpiritualIdentificationS = as.numeric(scale(Spiritual.Identification)),
         RelidS = as.numeric(scale(Relid)),
         AgeS = as.numeric(scale(Age)),
         EduS = as.numeric(scale(as.numeric(Edu))),
         polS = as.numeric(scale(Pol.Orient)))
         
#glimpse(dt1) # inspect data

Notably, religious identification and spiritual identification are correlated:

Show code

library(correlation)
c1 <- correlation::correlation(dt1, select = "RelidS", select2 = "SpiritualIdentificationS") 
  
summary(c1) %>%
  print_html()

Correlation Matrix (pearson-method)
Parameter	SpiritualIdentificationS
RelidS	0.54***
p-value adjustment method: Holm (1979)

However, it is possible that within these populations of identification people differ. Our statistical model investigating the relationship between Spiritual Identification on Attitudes to Euthanasia stratifies accross different levels of both Spiritual Identification and Religious Identification, while also adjusting for potentially confounding co-variates (see Appendix 2).

Show code

#model 
m1 <- glm(Issue.Euthanasia ~  SpiritualIdentificationS * RelidS  + AgeS + EduS + polS , data = dt1)

Focusing on the main effect of Spiritual Identification and support for Euthanasia, we find a negative relationship:

Show code

library(viridisLite)
plot(ggeffects::ggpredict(m1, terms = c( "SpiritualIdentificationS")), add.data = TRUE, dot.alpha = 0.1) + scale_y_continuous(limits = c(1,7)) + scale_colour_viridis_c() + 
  xlab("Level of Spiritual Identification (standarised)") + ylab("Attitudes to Euthanasia (1-7)")

However, stratifying across levels of religious identification, we find this effect is driven entirely by high levels of religious identification

Show code

library(viridisLite)
plot(ggeffects::ggpredict(m1, terms = c( "SpiritualIdentificationS", "RelidS")), add.data = TRUE, dot.alpha = 0.1) + scale_y_continuous(limits = c(1,7)) +
    xlab("Level of Spiritual Identification (standarised)") + ylab("Attitudes to Euthanasia (1-7)") + scale_colour_viridis_c()

Generally, the New Zealand population is supportive of euthanasia for people who seek to end their life with medical help when facing incurable disease. Even highly religious people average above the mid-point of support for such a right. However, on average, highly religious express less support than do secular people.

Our interest is in the relationship of spiritual identification to euthanasia attitudes. Religious identification and spiritual identification are correlated. However, when it comes to ethics at the end of life, people who are highly spiritual but not religious strongly differ from highly religious people. Whereas high levels of religiosity predicts lower levels of support for end of life, high levels of spirituality does not. It is only among the jointly religious and spiritual population that we find lower support. This finding reveals is important because it reveals a separatation between spiritual identification and religious identification on a topic of public importance.

For more information about the NZAVS see: here and here

Other issues

Genetic Engineering

Ensuring that all food and food ingredients sold in New Zealand are free from Genetically Modified Organisms.

Strict regulation limiting the use of genetic engineering in humans.

Show code

#model 
m2 <- glm(Issue.Food.GMO ~  SpiritualIdentificationS * RelidS  + AgeS + EduS + polS , data = dt1)

Show code

library(viridisLite)
plot(ggeffects::ggpredict(m2, terms = c( "SpiritualIdentificationS", "RelidS")), add.data = TRUE, dot.alpha = 0.1) + scale_y_continuous(limits = c(1,7)) + scale_colour_viridis_c() + 
  xlab("Level of Spiritual Identification (standarised)") + ylab("Attitudes to Genetic Engineering (1-7)")

Income Redistribution

Redistributing money and wealth more evenly among a larger percentage of the people in New Zealand through heavy taxes on the rich.

No differences and greater acceptance of taxation among both highly religiously identified and highly spiritually identified. Here we find that having either a high religious or a high spiritual identity (or both) drives support for income redistribution.

Show code

#model 
m3 <- glm(Issue.IncomeRedistribution ~  SpiritualIdentificationS * RelidS  + AgeS + EduS + polS , data = dt1)

Show code

library(viridisLite)
plot(ggeffects::ggpredict(m3, terms = c( "SpiritualIdentificationS", "RelidS")), add.data = TRUE, dot.alpha = 0.1) + scale_y_continuous(limits = c(1,7)) + scale_colour_viridis_c() + 
  xlab("Level of Spiritual Identification (standarised)") + ylab("Attitudes to Income Redistribution (1-7)")

Healthy Lifestyle

Government initiatives to inform and promote healthy lifestyle choices.

Here we see a convergence: high religious identification and high sprituality support Governmet investment in health. At the low end of spiritual identification, among those who are high in religious identification there is markedly lower support, whereas there is only only moderately reduced support among those who are low in both spiritual identification and religious identification. This reveals the importance of spirituality as a moderator of attitudes for governement investment in health.

Show code

#model 
m4 <- glm(Issue.GovtHealthPolicy ~  SpiritualIdentificationS * RelidS  + AgeS + EduS + polS , data = dt1)

Show code

library(viridisLite)
plot(ggeffects::ggpredict(m4, terms = c( "SpiritualIdentificationS", "RelidS")), add.data = TRUE, dot.alpha = 0.1) + scale_y_continuous(limits = c(1,7)) + scale_colour_viridis_c() + 
  xlab("Level of Spiritual Identification (standarised)") + ylab("Attitudes to Government Support for Healthy Lifestyles (1-7)")

Appendix: Results table

Show code

sjPlot::tab_model(m1)

	Issue.Euthanasia
Predictors	Estimates	CI	p
(Intercept)	5.83	5.80 – 5.85	<0.001
SpiritualIdentificationS	-0.18	-0.20 – -0.15	<0.001
RelidS	-0.56	-0.60 – -0.53	<0.001
AgeS	-0.08	-0.10 – -0.06	<0.001
EduS	-0.11	-0.13 – -0.08	<0.001
polS	-0.22	-0.24 – -0.19	<0.001
SpiritualIdentificationS * RelidS	-0.25	-0.28 – -0.22	<0.001
Observations	19805
R2	0.259

Show code

sjPlot::tab_model(m2)

	Issue.Food.GMO
Predictors	Estimates	CI	p
(Intercept)	5.05	5.02 – 5.08	<0.001
SpiritualIdentificationS	0.45	0.42 – 0.48	<0.001
RelidS	-0.04	-0.08 – -0.01	0.021
AgeS	0.04	0.01 – 0.06	0.003
EduS	-0.19	-0.22 – -0.17	<0.001
polS	-0.07	-0.09 – -0.04	<0.001
SpiritualIdentificationS * RelidS	-0.09	-0.13 – -0.06	<0.001
Observations	19813
R2	0.070

Show code

sjPlot::tab_model(m3)

	Issue.IncomeRedistribution
Predictors	Estimates	CI	p
(Intercept)	4.47	4.44 – 4.50	<0.001
SpiritualIdentificationS	0.17	0.14 – 0.20	<0.001
RelidS	-0.00	-0.04 – 0.03	0.810
AgeS	-0.02	-0.05 – 0.00	0.101
EduS	0.03	0.00 – 0.05	0.028
polS	-0.63	-0.65 – -0.60	<0.001
SpiritualIdentificationS * RelidS	0.04	0.01 – 0.07	0.019
Observations	19798
R2	0.119

Show code

sjPlot::tab_model(m4)

	Issue.GovtHealthPolicy
Predictors	Estimates	CI	p
(Intercept)	5.87	5.85 – 5.89	<0.001
SpiritualIdentificationS	0.11	0.09 – 0.13	<0.001
RelidS	-0.08	-0.11 – -0.06	<0.001
AgeS	-0.05	-0.07 – -0.03	<0.001
EduS	0.14	0.12 – 0.15	<0.001
polS	-0.08	-0.10 – -0.07	<0.001
SpiritualIdentificationS * RelidS	0.04	0.02 – 0.06	<0.001
Observations	19758
R2	0.036

Appendix: Causal Assumptions

Show code

library(ggdag)

library(ggdag)
dg <- dagify(
  euth ~ age + spirit + relid + pol + eth + edu + U,
  edu ~ eth + age + U,
  relid ~ age + pol + eth + male,
  spirit ~ relid + age + pol + eth + edu + male,
  exposure =  "spirit",
  outcome =   "euth",
  latent = "U",
  labels = c(
    "euth" = "Euthanasia Attitudes",
    "male" = "male",
    "spirit" = "Spirituality",
    "eth" = "ethnicity",
    "edu" = "education",
    "relid" = "religious identification",
    "age" = "age",
    "pol" = "political orientation",
    "U" = "Unmeasured"
  )
) %>%
  tidy_dagitty() 
#%>%
#  control_for("cs")

# dplot1 <-
#   ggdag::ggdag_collider(dg, use_labels = "label", text = FALSE) + theme_dag_blank() +
#   labs(title = "")
# dplot1


ggdag::ggdag_adjustment_set(dg, use_labels = "label", text = FALSE) + theme_dag_blank() +
  labs(title = "Adjustment set for variables needed to obtain an unbaiased causal estimate of spirituality on attitudes to Euthanasia ")

Our model is:

Show code

equatiomatic::extract_eq(m1,
                         wrap = TRUE, 
                         terms_per_line = 1)

\[ \begin{aligned} E( \operatorname{Issue.Euthanasia} ) &= \alpha\ + \\ &\quad \beta_{1}(\operatorname{SpiritualIdentificationS})\ + \\ &\quad \beta_{2}(\operatorname{RelidS})\ + \\ &\quad \beta_{3}(\operatorname{AgeS})\ + \\ &\quad \beta_{4}(\operatorname{EduS})\ + \\ &\quad \beta_{5}(\operatorname{polS})\ + \\ &\quad \beta_{6}(\operatorname{SpiritualIdentificationS} \times \operatorname{RelidS}) \end{aligned} \]