Skip to main content

Table 2 Regression results for recursive bivariate probit models

From: Euthanasia, religiosity and the valuation of health states: results from an Irish EQ5D5L valuation study and their implications for anchor values

Dependant Variable

Worse-than-Dead

Euthanasia

Coefficient

Standard Error

Coefficient

Standard Error

Third Level Education

−0.155

(0.167)

−0.077

(0.255)

Age Group (Base: 18–35)

 36–45

0.304

(0.207)

0.328

(0.32)

 46–60

0.471*

(0.228)

0.318

(0.42)

 61+

1.168***

(0.274)

0.585

(0.489)

Male

−0.298

(0.17)

−0.113

(0.251)

Married/Living as Married

−0.158

(0.204)

−0.07

(0.303)

Dependants Under 18 (Y/N)

−0.098

(0.208)

0.412

(0.359)

Visual Analogue Scale

0.004

(0.006)

−0.012

(0.008)

Experienced a serious illness

−0.287

(0.264)

−0.765*

(0.3)

In favour of Legalisation of Euthanasia (see scenario in Fig. 1)

1.003*

(0.394)

Religiosity (Base: A few times a year or less)

 Monthly

−0.454

(0.318)

 Weekly

−1.369***

(0.328)

Mobility (Base: No Problems)

 Slight problems

0.194

(0.114)

 Moderate problems

0.326**

(0.109)

 Severe problems

0.476***

(0.12)

 Unable

0.419***

(0.112)

Self-care (Base: No Problems)

 Slight problems

0.263*

(0.128)

 Moderate problems

0.34*

(0.131)

 Severe problems

0.528***

(0.143)

 Unable

0.444***

(0.102)

Usual Activities (Base: No Problems)

 Slight problems

0.304*

(0.134)

 Moderate problems

0.477**

(0.156)

 Severe problems

0.443***

(0.118)

 Unable

0.335***

(0.091)

Paid/Discomfort (Base: No Problems)

 Slight problems

0.253*

(0.116)

 Moderate problems

0.182

(0.116)

 Severe problems

0.614***

(0.111)

 Extreme problems

0.852***

(0.138)

Anxiety/Depression (Base: No Problems)

 Slight problems

0.221

(0.128)

 Moderate problems

0.313*

(0.14)

 Severe problems

0.867***

(0.159)

 Extreme problems

0.858***

(0.133)

Constant

−3.331***

(0.626)

2.05**

(0.748)

Rhoa

  

−0.801*

(0.402)

Number of observationsa

   

1600

Number of clustersa

   

160

  1. The table shows the relationship between the dependent and independent variables in a recursive bivariate probit regression. Coefficients can be interpreted in the usual manner – the significant positive coefficient associated with being aged 61+ for example shows that the probability of assigning a WTD value is higher among this age group relative to those aged 18–35. Similarly, the negative and significant coefficient on those who attend religious services weekly compared to those who attend a few times a year or less, shows that the frequently attending group are less likely to support access to euthanasia than the infrequently attending group. Other coefficients can be interpreted in a similar fashion. The negative and significant value for Rho shows the correlation in errors between the two regressions is negative and large (0.8 in absolute terms)
  2. aPertains to both models
  3. *p < 0.05, **p < 0.01, ***p < 0.001