Attaching package: 'dplyr'

The following objects are masked from 'package:stats':

    filter, lag

The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union


=== TABLE 1: Available N per variable (verified cases) ===
Verified cases:             1678 
Intermission time:          1319 
Death unexpectedness:       1307 
Naturalness of death:       1526 
Age at death:               1317 
Verification precision:     1678 
Memory contamination:       895 
Family association:         1520 
Social status difference:   751 
Reincarnation belief:       1316 

=== Descriptive statistics ===
Unnatural deaths: 57% (n = 868)
Natural deaths:   43% (n = 658)
Completely unexpected (rating 4): 68%

Naturalness x Unexpectedness (chi-square):

	Pearson's Chi-squared test

data:  table(dat$NATURAL_DEATH, dat$death_unexp)
X-squared = 509.07, df = 4, p-value < 2.2e-16

Naturalness x Age at death (Wilcoxon):

	Wilcoxon rank sum test with continuity correction

data:  agedhp by NATURAL_DEATH
W = 131371, p-value < 2.2e-16
alternative hypothesis: true location shift is not equal to 0

Median age: unnatural = 25 years, natural = 40 years
Unexpectedness x Age at death (Pearson):
r(1079) = -0.239, 95% CI [-0.294, -0.182], t = -8.08, p = 1.77e-15, N = 1081

IT skewness = 6.82

	D'Agostino skewness test

data:  x
skew = 6.8164, z = 34.5044, p-value < 2.2e-16
alternative hypothesis: data have a skewness

Median IT: 16 months (IQR: 8-48)
Maximum IT: 150 years

=== TABLE 2: Correlations with log(IT) ===
Unexpectedness x log(IT):
r(1073) = -0.126, 95% CI [-0.184, -0.066], t = -4.15, p = 3.64e-05, N = 1075
Naturalness x log(IT):
r(1233) = 0.132, 95% CI [0.077, 0.186], t = 4.67, p = 3.38e-06, N = 1235
Age at death x log(IT):
r(1155) = 0.042, 95% CI [-0.016, 0.099], t = 1.42, p = 0.157, N = 1157

Spearman sensitivity check:
Unexpectedness: Spearman r = -0.133, p = 1.16e-05, N = 1075
Naturalness:    Spearman r = 0.134, p = 2.21e-06, N = 1235
Age at death:   Spearman r = 0.030, p = 0.307, N = 1157

Multiple regression (unexpectedness + naturalness):

Call:
lm(formula = intermp_log ~ death_unexp + NATURAL_DEATH, data = dat)

Residuals:
    Min      1Q  Median      3Q     Max 
-0.8946 -0.3106 -0.1065  0.3329  1.9151 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept)    1.43292    0.05519  25.965   <2e-16 ***
death_unexp   -0.03059    0.01380  -2.216   0.0269 *  
NATURAL_DEATH  0.06375    0.04039   1.578   0.1148    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.4952 on 1050 degrees of freedom
  (625 observations deleted due to missingness)
Multiple R-squared:  0.01772,	Adjusted R-squared:  0.01585 
F-statistic: 9.472 on 2 and 1050 DF,  p-value: 8.378e-05


Median IT by unexpectedness rating:
# A tibble: 5 x 3
  death_unexp median_IT     n
        <int>     <dbl> <int>
1           0        21   114
2           1        20    69
3           2        17   108
4           3        17    76
5           4        12   708
Some anticipation (0-3): 18 months; completely unexpected (4): 12 months (33% shorter)
Natural deaths: 19 months; unnatural deaths: 12 months (37% shorter)

=== Moderator 1: Verification precision ===
Stringently verified cases:
r(741) = -0.150, 95% CI [-0.220, -0.079], t = -4.14, p = 3.86e-05, N = 743
Non-stringently verified cases:
r(330) = -0.067, 95% CI [-0.174, 0.041], t = -1.23, p = 0.221, N = 332
Interaction model:

Call:
lm(formula = intermp_log ~ death_unexp * stringent_verif, data = dat)

Residuals:
     Min       1Q   Median       3Q      Max 
-1.41120 -0.31965 -0.07088  0.32610  1.82054 

Coefficients:
                                Estimate Std. Error t value Pr(>|t|)    
(Intercept)                      1.50918    0.06422  23.499   <2e-16 ***
death_unexp                     -0.02449    0.01908  -1.284    0.200    
stringent_verifTRUE             -0.02272    0.07830  -0.290    0.772    
death_unexp:stringent_verifTRUE -0.02837    0.02312  -1.227    0.220    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.494 on 1071 degrees of freedom
  (603 observations deleted due to missingness)
Multiple R-squared:  0.02748,	Adjusted R-squared:  0.02475 
F-statistic: 10.09 on 3 and 1071 DF,  p-value: 1.482e-06


=== Moderator 2: Memory contamination ===
No or inaccurate memory of mode of death (no possible contamination):
r(200) = -0.161, 95% CI [-0.293, -0.024], t = -2.31, p = 0.022, N = 202
Accurate memory of mode of death (possible contamination):
r(475) = -0.117, 95% CI [-0.204, -0.027], t = -2.56, p = 0.0107, N = 477
Interaction model:

Call:
lm(formula = intermp_log ~ death_unexp * no_or_wrong_memory, 
    data = dat)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.93366 -0.35293 -0.07389  0.36666  1.62994 

Coefficients:
                                   Estimate Std. Error t value Pr(>|t|)    
(Intercept)                         1.49031    0.07663  19.448   <2e-16 ***
death_unexp                        -0.05307    0.02111  -2.514   0.0122 *  
no_or_wrong_memoryTRUE              0.04541    0.10607   0.428   0.6687    
death_unexp:no_or_wrong_memoryTRUE -0.00208    0.03115  -0.067   0.9468    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.5016 on 675 degrees of freedom
  (999 observations deleted due to missingness)
Multiple R-squared:  0.02208,	Adjusted R-squared:  0.01774 
F-statistic: 5.081 on 3 and 675 DF,  p-value: 0.001746


=== Moderator 3: Family association ===
Distribution of family association levels:

 0  1  2  3  4 
28 10 14  8 39 
Main-effect model (controlling for family association):

Call:
lm(formula = intermp_log ~ death_unexp + family_assoc, data = dat)

Residuals:
    Min      1Q  Median      3Q     Max 
-1.3097 -0.3185 -0.1056  0.3317  1.8916 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept)   1.494024   0.046055  32.440  < 2e-16 ***
death_unexp  -0.043135   0.011227  -3.842  0.00013 ***
family_assoc -0.002946   0.009578  -0.308  0.75844    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.4985 on 1014 degrees of freedom
  (661 observations deleted due to missingness)
Multiple R-squared:  0.01445,	Adjusted R-squared:  0.01251 
F-statistic: 7.434 on 2 and 1014 DF,  p-value: 0.0006238

Interaction model:

Call:
lm(formula = intermp_log ~ death_unexp * family_assoc, data = dat)

Residuals:
     Min       1Q   Median       3Q      Max 
-1.32650 -0.31481 -0.09899  0.33287  1.88411 

Coefficients:
                          Estimate Std. Error t value Pr(>|t|)    
(Intercept)               1.579178   0.073822  21.392  < 2e-16 ***
death_unexp              -0.069017   0.020824  -3.314 0.000951 ***
family_assoc             -0.035298   0.023926  -1.475 0.140439    
death_unexp:family_assoc  0.010286   0.006972   1.475 0.140418    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.4982 on 1013 degrees of freedom
  (661 observations deleted due to missingness)
Multiple R-squared:  0.01656,	Adjusted R-squared:  0.01365 
F-statistic: 5.687 on 3 and 1013 DF,  p-value: 0.0007314


=== Moderator 4: Social status ===
PP higher status: 203; same: 405; PP lower status: 143
Binomial test (higher vs. lower, H0: p = .5):

	Exact binomial test

data:  pp_higher and pp_higher + pp_lower
number of successes = 203, number of trials = 346, p-value = 0.001479
alternative hypothesis: true probability of success is not equal to 0.5
95 percent confidence interval:
 0.5328158 0.6390976
sample estimates:
probability of success 
             0.5867052 

Main-effect model (controlling for social status difference):

Call:
lm(formula = intermp_log ~ death_unexp + social_status_diff_recoded, 
    data = dat)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.82257 -0.28009 -0.09521  0.24174  1.92339 

Coefficients:
                            Estimate Std. Error t value Pr(>|t|)    
(Intercept)                 1.424627   0.048056  29.645   <2e-16 ***
death_unexp                -0.030593   0.014276  -2.143   0.0326 *  
social_status_diff_recoded -0.002922   0.009588  -0.305   0.7607    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.4573 on 543 degrees of freedom
  (1132 observations deleted due to missingness)
Multiple R-squared:  0.008507,	Adjusted R-squared:  0.004855 
F-statistic:  2.33 on 2 and 543 DF,  p-value: 0.09831

Interaction model:

Call:
lm(formula = intermp_log ~ death_unexp * social_status_diff_recoded, 
    data = dat)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.84013 -0.27416 -0.08839  0.24196  1.92436 

Coefficients:
                                        Estimate Std. Error t value Pr(>|t|)
(Intercept)                             1.416527   0.048474  29.222   <2e-16
death_unexp                            -0.028215   0.014397  -1.960   0.0505
social_status_diff_recoded              0.025663   0.024952   1.029   0.3042
death_unexp:social_status_diff_recoded -0.008925   0.007193  -1.241   0.2152
                                          
(Intercept)                            ***
death_unexp                            .  
social_status_diff_recoded                
death_unexp:social_status_diff_recoded    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.457 on 542 degrees of freedom
  (1132 observations deleted due to missingness)
Multiple R-squared:  0.01132,	Adjusted R-squared:  0.005843 
F-statistic: 2.068 on 3 and 542 DF,  p-value: 0.1035


=== Moderator 5: Belief in reincarnation ===
Distribution of reincarnation-favorable religious affiliation:

 0  1 
28 72 
Believing families:
r(624) = -0.070, 95% CI [-0.148, 0.008], t = -1.75, p = 0.0798, N = 626
Non-believing families:
r(253) = -0.211, 95% CI [-0.325, -0.090], t = -3.43, p = 0.000707, N = 255
Interaction model:

Call:
lm(formula = intermp_log ~ death_unexp * believeReinc_S, data = dat)

Residuals:
     Min       1Q   Median       3Q      Max 
-1.17455 -0.32945 -0.07241  0.30219  1.85946 

Coefficients:
                           Estimate Std. Error t value Pr(>|t|)    
(Intercept)                 1.53700    0.08811  17.444  < 2e-16 ***
death_unexp                -0.09061    0.02412  -3.757 0.000183 ***
believeReinc_S             -0.07022    0.09949  -0.706 0.480512    
death_unexp:believeReinc_S  0.06728    0.02782   2.418 0.015809 *  
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.4827 on 877 degrees of freedom
  (797 observations deleted due to missingness)
Multiple R-squared:  0.04303,	Adjusted R-squared:  0.03976 
F-statistic: 13.15 on 3 and 877 DF,  p-value: 2.114e-08