Start with SOS. Importing and cleaning data.

First looking at main effects of condition.

No effect of condition on overall health in the last week.

s$condition[s$Area_type == 2] <- 0
s$condition[s$Area_type == 1] <- 1
table(s$condition)
## 
##   0   1 
## 373 382
summary(lm(overall.health ~ condition, s))
## 
## Call:
## lm(formula = overall.health ~ condition, data = s)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.5420 -1.5157  0.4580  0.4843  2.4843 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.54201    0.06866  37.022   <2e-16 ***
## condition   -0.02630    0.09627  -0.273    0.785    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.319 on 749 degrees of freedom
##   (15 observations deleted due to missingness)
## Multiple R-squared:  9.962e-05,  Adjusted R-squared:  -0.001235 
## F-statistic: 0.07462 on 1 and 749 DF,  p-value: 0.7848

No effect on symptoms experiences in the last week.

summary(lm(symptoms ~ condition, s))
## 
## Call:
## lm(formula = symptoms ~ condition, data = s)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -2.7355 -1.7290  0.2645  1.2710  4.2710 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.73554    0.09315   29.37   <2e-16 ***
## condition   -0.00654    0.13120   -0.05     0.96    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.775 on 730 degrees of freedom
##   (34 observations deleted due to missingness)
## Multiple R-squared:  3.404e-06,  Adjusted R-squared:  -0.001366 
## F-statistic: 0.002485 on 1 and 730 DF,  p-value: 0.9603

Economic DVs. First analyzed as a composite of the 3 items. No effect of condition.

library(psych)
## Warning: package 'psych' was built under R version 3.2.5
alpha(s[,c(33:35)], na.rm=TRUE)
## 
## Reliability analysis   
## Call: alpha(x = s[, c(33:35)], na.rm = TRUE)
## 
##   raw_alpha std.alpha G6(smc) average_r S/N  ase mean  sd
##        0.6      0.58    0.54      0.32 1.4 0.02  2.3 1.8
## 
##  lower alpha upper     95% confidence boundaries
## 0.56 0.6 0.64 
## 
##  Reliability if an item is dropped:
##     raw_alpha std.alpha G6(smc) average_r  S/N alpha se
## qE2      0.17      0.21    0.12      0.12 0.27    0.047
## qE3      0.36      0.45    0.29      0.29 0.80    0.035
## qE4      0.70      0.70    0.54      0.54 2.39    0.021
## 
##  Item statistics 
##       n raw.r std.r r.cor r.drop mean  sd
## qE2 292  0.88  0.83  0.73   0.59  3.1 3.1
## qE3 707  0.87  0.75  0.59   0.49  2.2 3.0
## qE4 758  0.52  0.64  0.30   0.23  2.2 1.4
## 
## Non missing response frequency for each item
##        0    1    2    3    4    5    6    7 miss
## qE2 0.43 0.03 0.07 0.04 0.03 0.04 0.02 0.34 0.62
## qE3 0.58 0.03 0.04 0.04 0.04 0.04 0.00 0.24 0.08
## qE4 0.00 0.53 0.04 0.25 0.06 0.12 0.00 0.00 0.01
s$econ<-(s$qE2 + s$qE3 + s$qE4)/3
summary(lm(econ~condition, s))
## 
## Call:
## lm(formula = econ ~ condition, data = s)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -2.2174 -1.9475 -0.5507  1.4493  3.9891 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   2.3443     0.1761  13.309   <2e-16 ***
## condition     0.2065     0.2335   0.884    0.377    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.946 on 281 degrees of freedom
##   (483 observations deleted due to missingness)
## Multiple R-squared:  0.002774,   Adjusted R-squared:  -0.0007749 
## F-statistic: 0.7816 on 1 and 281 DF,  p-value: 0.3774

Analyze the 3 items separately: ability to work, do chores, and stress about money. Significant effect on the last item only. Since condition is 0 for control and 1 for intervention, this effect shows that people were MORE stressed in the intervention condition.

s$unable.to.work<-s$qE2
s$unable.do.chores<-s$qE3
s$money.stress<-s$qE4

summary(lm(unable.to.work~condition, s))
## 
## Call:
## lm(formula = unable.to.work ~ condition, data = s)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.2147 -2.9048 -0.9048  3.7853  4.0952 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   2.9048     0.2799  10.379   <2e-16 ***
## condition     0.3100     0.3727   0.832    0.406    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 3.141 on 287 degrees of freedom
##   (477 observations deleted due to missingness)
## Multiple R-squared:  0.002405,   Adjusted R-squared:  -0.001071 
## F-statistic: 0.6919 on 1 and 287 DF,  p-value: 0.4062
summary(lm(unable.do.chores~condition, s))
## 
## Call:
## lm(formula = unable.do.chores ~ condition, data = s)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -2.339 -2.339 -2.059  2.941  4.941 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   2.0593     0.1610  12.790   <2e-16 ***
## condition     0.2795     0.2232   1.252    0.211    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.956 on 701 degrees of freedom
##   (63 observations deleted due to missingness)
## Multiple R-squared:  0.002232,   Adjusted R-squared:  0.0008087 
## F-statistic: 1.568 on 1 and 701 DF,  p-value: 0.2109
summary(lm(money.stress~condition, s))
## 
## Call:
## lm(formula = money.stress ~ condition, data = s)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.3605 -1.0699 -1.0699  0.9301  2.9301 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.06989    0.07453  27.772  < 2e-16 ***
## condition    0.29063    0.10485   2.772  0.00571 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.437 on 750 degrees of freedom
##   (14 observations deleted due to missingness)
## Multiple R-squared:  0.01014,    Adjusted R-squared:  0.008822 
## F-statistic: 7.684 on 1 and 750 DF,  p-value: 0.005709

Very low alpha for the two “mental outcomes” (hopeful and depressed), so analyze them separately. Only effect is that people in the intervention condition felt LESS hopeful.

s$hopeful<-s$qM1
s$depressed<-s$qM2
alpha(s[,c(54:55)], na.rm=TRUE, check.keys=TRUE)
## Warning in matrix(unlist(drop.item), ncol = 8, byrow = TRUE): data length
## [12] is not a sub-multiple or multiple of the number of columns [8]
## 
## Reliability analysis   
## Call: alpha(x = s[, c(54:55)], na.rm = TRUE, check.keys = TRUE)
## 
##   raw_alpha std.alpha G6(smc) average_r  S/N   ase mean  sd
##       0.21      0.21    0.12      0.12 0.26 0.057  2.7 1.1
## 
##  lower alpha upper     95% confidence boundaries
## 0.09 0.21 0.32 
## 
##  Reliability if an item is dropped:
##           raw_alpha std.alpha G6(smc) average_r  S/N alpha se
## hopeful       0.116      0.12   0.013      0.12   NA       NA
## depressed     0.013      0.12      NA        NA 0.12    0.012
## 
##  Item statistics 
##             n raw.r std.r r.cor r.drop mean  sd
## hopeful   760  0.80  0.75  0.25   0.12  3.0 1.6
## depressed 758  0.69  0.75  0.25   0.12  2.4 1.3
## 
## Non missing response frequency for each item
##             1    2    3    4   5 miss
## hopeful   0.3 0.06 0.28 0.07 0.3 0.01
## depressed 0.4 0.07 0.38 0.05 0.1 0.01
summary(lm(hopeful~condition, s)) 
## 
## Call:
## lm(formula = hopeful ~ condition, data = s)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -2.1774 -1.8534  0.1466  1.8226  2.1466 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.17742    0.08207   38.72  < 2e-16 ***
## condition   -0.32402    0.11530   -2.81  0.00508 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.583 on 752 degrees of freedom
##   (12 observations deleted due to missingness)
## Multiple R-squared:  0.01039,    Adjusted R-squared:  0.009077 
## F-statistic: 7.897 on 1 and 752 DF,  p-value: 0.005079
summary(lm(depressed~condition, s)) 
## 
## Call:
## lm(formula = depressed ~ condition, data = s)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.4368 -1.3226  0.5632  0.6774  2.6774 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.32258    0.06875  33.782   <2e-16 ***
## condition    0.11426    0.09672   1.181    0.238    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.326 on 750 degrees of freedom
##   (14 observations deleted due to missingness)
## Multiple R-squared:  0.001858,   Adjusted R-squared:  0.0005266 
## F-statistic: 1.396 on 1 and 750 DF,  p-value: 0.2378

Very low alpha for the two “social outcomes” (how often illness interfered with social activities, and how often patient felt rejected), so analyze them separately. No significant effects here.

alpha(s[,c(37,38)], na.rm=TRUE)
## Warning in matrix(unlist(drop.item), ncol = 8, byrow = TRUE): data length
## [12] is not a sub-multiple or multiple of the number of columns [8]
## 
## Reliability analysis   
## Call: alpha(x = s[, c(37, 38)], na.rm = TRUE)
## 
##   raw_alpha std.alpha G6(smc) average_r S/N   ase mean sd
##       0.52      0.52    0.35      0.35 1.1 0.034  2.1  1
## 
##  lower alpha upper     95% confidence boundaries
## 0.45 0.52 0.59 
## 
##  Reliability if an item is dropped:
##     raw_alpha std.alpha G6(smc) average_r  S/N alpha se
## qM2      0.35      0.35    0.13      0.35   NA       NA
## qS1      0.13      0.35      NA        NA 0.35    0.022
## 
##  Item statistics 
##       n raw.r std.r r.cor r.drop mean  sd
## qM2 758  0.84  0.82  0.49   0.35  2.4 1.3
## qS1 759  0.80  0.82  0.49   0.35  1.7 1.2
## 
## Non missing response frequency for each item
##        1    2    3    4    5 miss
## qM2 0.40 0.07 0.38 0.05 0.10 0.01
## qS1 0.69 0.04 0.18 0.03 0.06 0.01
s$social<-s$qS1
s$rejected<-s$qS2

summary(lm(social ~ condition, s))
## 
## Call:
## lm(formula = social ~ condition, data = s)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -0.7500 -0.7500 -0.6595  1.2500  3.3405 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.65952    0.06125  27.093   <2e-16 ***
## condition    0.09048    0.08623   1.049    0.294    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.183 on 751 degrees of freedom
##   (13 observations deleted due to missingness)
## Multiple R-squared:  0.001464,   Adjusted R-squared:  0.0001345 
## F-statistic: 1.101 on 1 and 751 DF,  p-value: 0.2943
summary(lm(rejected ~ condition, s))
## 
## Call:
## lm(formula = rejected ~ condition, data = s)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -0.4182 -0.4182 -0.4079 -0.4079  3.5921 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.41823    0.04653  30.479   <2e-16 ***
## condition   -0.01034    0.06550  -0.158    0.875    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8987 on 751 degrees of freedom
##   (13 observations deleted due to missingness)
## Multiple R-squared:  3.315e-05,  Adjusted R-squared:  -0.001298 
## F-statistic: 0.0249 on 1 and 751 DF,  p-value: 0.8747

Moving on to the Midline survey. No effect of condition.

table(m$Area_Type)
## 
##   1   2 
## 232 216
m$condition[m$Area_Type==2]<-0
m$condition[m$Area_Type==1]<-1

summary(lm(health1_rating~condition, m))
## 
## Call:
## lm(formula = health1_rating ~ condition, data = m)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.5129 -1.4259 -0.4259  0.5741  2.5741 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.42593    0.08828  27.480   <2e-16 ***
## condition    0.08701    0.12267   0.709    0.479    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.297 on 446 degrees of freedom
## Multiple R-squared:  0.001127,   Adjusted R-squared:  -0.001113 
## F-statistic: 0.503 on 1 and 446 DF,  p-value: 0.4786

No effect of condition on symptoms.

m$symptoms<-(m$health2_coughing + m$health2_fever + m$health2_appetite+m$health2_fatigue + m$health2_sputum + m$health2_sweats + m$health2_breathing)

table(m$symptoms)
## 
##   0   1   2   3   4   5   6   7 889 890 894 
##  41  77  68  87  67  49  47   8   2   1   1
m$symptoms[m$symptoms>7]<-NA

summary(lm(symptoms~condition, m))
## 
## Call:
## lm(formula = symptoms ~ condition, data = m)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -3.048 -1.916 -0.048  1.084  4.084 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   2.9163     0.1280  22.783   <2e-16 ***
## condition     0.1318     0.1782   0.739     0.46    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.877 on 442 degrees of freedom
##   (4 observations deleted due to missingness)
## Multiple R-squared:  0.001235,   Adjusted R-squared:  -0.001025 
## F-statistic: 0.5465 on 1 and 442 DF,  p-value: 0.4602

Very low reliability for economic outcomes, so analyze them separately. No effects of condition on any of the economic outcomes.

alpha(m[,c(53:55)], na.rm=TRUE, check.keys=TRUE)
## 
## Reliability analysis   
## Call: alpha(x = m[, c(53:55)], na.rm = TRUE, check.keys = TRUE)
## 
##   raw_alpha std.alpha G6(smc) average_r  S/N   ase mean sd
##     0.0071      0.25    0.21       0.1 0.33 0.011  3.4 24
## 
##  lower alpha upper     95% confidence boundaries
## -0.01 0.01 0.03 
## 
##  Reliability if an item is dropped:
##                         raw_alpha std.alpha G6(smc) average_r     S/N
## econ2_missed_work_days   -0.00051   -0.0084 -0.0042   -0.0042 -0.0083
## econ3_missed_chore_days   0.35101    0.4292  0.2732    0.2732  0.7518
## econ4_worry_money         0.00789    0.0607  0.0313    0.0313  0.0647
##                         alpha se
## econ2_missed_work_days    0.0057
## econ3_missed_chore_days   0.0470
## econ4_worry_money         0.0118
## 
##  Item statistics 
##                           n raw.r std.r r.cor r.drop mean   sd
## econ2_missed_work_days  173 0.897  0.69 0.438  0.040  2.3  3.0
## econ3_missed_chore_days 448 0.999  0.54 0.033  0.024  4.2 47.2
## econ4_worry_money       448 0.026  0.67 0.396  0.013  2.5  1.4
m$econ<-(m$econ2_missed_work_days + m$econ3_missed_chore_days + m$econ4_worry_money)

m$econ3_missed_chore_days[m$econ3_missed_chore_days>7]<-NA

summary(lm(econ2_missed_work_days ~ condition, m))
## 
## Call:
## lm(formula = econ2_missed_work_days ~ condition, data = m)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -2.670 -2.670 -1.951  3.049  5.049 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   1.9512     0.3292   5.927 1.66e-08 ***
## condition     0.7191     0.4539   1.584    0.115    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.981 on 171 degrees of freedom
##   (275 observations deleted due to missingness)
## Multiple R-squared:  0.01446,    Adjusted R-squared:  0.008699 
## F-statistic: 2.509 on 1 and 171 DF,  p-value: 0.115
summary(lm(econ3_missed_chore_days ~ condition, m))
## 
## Call:
## lm(formula = econ3_missed_chore_days ~ condition, data = m)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -2.100 -2.100 -1.903  2.097  5.097 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   1.9028     0.1958   9.720   <2e-16 ***
## condition     0.1968     0.2723   0.723     0.47    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.877 on 445 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.001172,   Adjusted R-squared:  -0.001072 
## F-statistic: 0.5222 on 1 and 445 DF,  p-value: 0.4703
summary(lm(econ4_worry_money ~ condition, m))
## 
## Call:
## lm(formula = econ4_worry_money ~ condition, data = m)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.5862 -1.4861  0.4138  1.4138  2.5139 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.48611    0.09763  25.465   <2e-16 ***
## condition    0.10010    0.13567   0.738    0.461    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.435 on 446 degrees of freedom
## Multiple R-squared:  0.001219,   Adjusted R-squared:  -0.00102 
## F-statistic: 0.5443 on 1 and 446 DF,  p-value: 0.461

Decent alpha for mental outcomes when excluding the “control over life” item, so analyze the items as a composite but excluding that item. No effect of condition on this composite.

alpha(m[,c(56:63)], na.rm=TRUE, check.keys=TRUE)
## Warning in alpha(m[, c(56:63)], na.rm = TRUE, check.keys = TRUE): Some items were negatively correlated with total scale and were automatically reversed.
##  This is indicated by a negative sign for the variable name.
## 
## Reliability analysis   
## Call: alpha(x = m[, c(56:63)], na.rm = TRUE, check.keys = TRUE)
## 
##   raw_alpha std.alpha G6(smc) average_r S/N   ase mean sd
##       0.02      0.57    0.59      0.14 1.3 0.061  376 10
## 
##  lower alpha upper     95% confidence boundaries
## -0.1 0.02 0.14 
## 
##  Reliability if an item is dropped:
##                       raw_alpha std.alpha G6(smc) average_r  S/N alpha se
## mental1_hopeful-          0.010      0.60    0.61      0.18 1.49    0.055
## mental2_depressed         0.016      0.45    0.46      0.10 0.82    0.063
## mental3_calm-             0.017      0.46    0.48      0.11 0.86    0.063
## mental4_nervous           0.017      0.48    0.49      0.12 0.92    0.063
## mental5_concentrating     0.017      0.50    0.53      0.13 1.01    0.063
## mental6_sharp-            0.017      0.52    0.54      0.13 1.07    0.063
## mental7_control           0.021      0.60    0.61      0.18 1.51    0.054
## mental8_confidence        0.023      0.61    0.62      0.18 1.54    0.054
## 
##  Item statistics 
##                         n raw.r std.r r.cor r.drop  mean   sd
## mental1_hopeful-      448 0.581  0.31 0.082 0.0121 993.9 47.0
## mental2_depressed     448 0.120  0.69 0.698 0.1041   2.5  1.3
## mental3_calm-         448 0.098  0.66 0.644 0.0834 996.6  1.2
## mental4_nervous       448 0.094  0.63 0.599 0.0773   2.5  1.4
## mental5_concentrating 448 0.095  0.57 0.476 0.0790   2.3  1.3
## mental6_sharp-        448 0.106  0.54 0.429 0.0907 996.5  1.3
## mental7_control       448 0.576  0.30 0.068 0.0055   5.7 47.1
## mental8_confidence    448 0.576  0.29 0.052 0.0044   4.4 47.1
m$mental1_hopeful[m$mental1_hopeful>7]<-NA
m$mental7_control[m$mental1_control>7]<-NA
m$mental8_confidence[m$mental8_confidence>7]<-NA

m$mental<-(m$mental1_hopeful + (6-m$mental2_depressed) + m$mental3_calm +(6-m$mental4_nervous) + m$mental5_concentrating + m$mental6_sharp +m$mental8_confidence)/7

summary(lm(mental~condition, data=m))
## 
## Call:
## lm(formula = mental ~ condition, data = m)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.57702 -0.37235 -0.00559  0.34193  1.70870 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.22950    0.03568  90.515   <2e-16 ***
## condition   -0.08105    0.04968  -1.631    0.104    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.5244 on 444 degrees of freedom
##   (2 observations deleted due to missingness)
## Multiple R-squared:  0.005958,   Adjusted R-squared:  0.003719 
## F-statistic: 2.661 on 1 and 444 DF,  p-value: 0.1035

Looking at the mental outcomes individually. Positive coefficients for condition indicate that the intervention condition INCREASED that mental outcome.

#LESS hopeful in the intervention condition. 
summary(lm(m$mental1_hopeful~condition, data=m))
## 
## Call:
## lm(formula = m$mental1_hopeful ~ condition, data = m)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -2.98148 -0.77056  0.01852  1.01852  1.22944 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.98148    0.07431   53.58   <2e-16 ***
## condition   -0.21092    0.10337   -2.04   0.0419 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.092 on 445 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.009269,   Adjusted R-squared:  0.007043 
## F-statistic: 4.163 on 1 and 445 DF,  p-value: 0.0419
summary(lm(m$mental2_depressed~condition, data=m))
## 
## Call:
## lm(formula = m$mental2_depressed ~ condition, data = m)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.5474 -1.4167  0.4526  0.5833  2.5833 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.41667    0.08952  26.997   <2e-16 ***
## condition    0.13075    0.12440   1.051    0.294    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.316 on 446 degrees of freedom
## Multiple R-squared:  0.002471,   Adjusted R-squared:  0.0002342 
## F-statistic: 1.105 on 1 and 446 DF,  p-value: 0.2938
summary(lm(m$mental3_calm~condition, data=m))
## 
## Call:
## lm(formula = m$mental3_calm ~ condition, data = m)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -2.4306 -0.4306 -0.4181  0.8288  1.5819 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.43056    0.08348  41.093   <2e-16 ***
## condition   -0.01245    0.11601  -0.107    0.915    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.227 on 446 degrees of freedom
## Multiple R-squared:  2.583e-05,  Adjusted R-squared:  -0.002216 
## F-statistic: 0.01152 on 1 and 446 DF,  p-value: 0.9146
summary(lm(m$mental4_nervous~condition, data=m))
## 
## Call:
## lm(formula = m$mental4_nervous ~ condition, data = m)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.4698 -1.4676  0.5302  0.5324  2.5324 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 2.467593   0.092235  26.753   <2e-16 ***
## condition   0.002235   0.128172   0.017    0.986    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.356 on 446 degrees of freedom
## Multiple R-squared:  6.818e-07,  Adjusted R-squared:  -0.002241 
## F-statistic: 0.0003041 on 1 and 446 DF,  p-value: 0.9861
summary(lm(m$mental5_concentrating~condition, data=m))
## 
## Call:
## lm(formula = m$mental5_concentrating ~ condition, data = m)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.3621 -1.3056 -0.3056  0.6944  2.6944 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.30556    0.09121  25.278   <2e-16 ***
## condition    0.05651    0.12674   0.446    0.656    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.34 on 446 degrees of freedom
## Multiple R-squared:  0.0004456,  Adjusted R-squared:  -0.001796 
## F-statistic: 0.1988 on 1 and 446 DF,  p-value: 0.6559
#LESS sharp in the intervention condition. 
summary(lm(m$mental6_sharp~condition, data=m))
## 
## Call:
## lm(formula = m$mental6_sharp ~ condition, data = m)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -2.6852 -0.6852  0.3148  1.3148  1.7586 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.68519    0.08688  42.419  < 2e-16 ***
## condition   -0.44381    0.12072  -3.676 0.000266 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.277 on 446 degrees of freedom
## Multiple R-squared:  0.02941,    Adjusted R-squared:  0.02723 
## F-statistic: 13.51 on 1 and 446 DF,  p-value: 0.0002655
summary(lm(m$mental7_control~condition, data=m))
## 
## Call:
## lm(formula = m$mental7_control ~ condition, data = m)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
##  -6.61  -4.61  -2.61   0.44 991.39 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)
## (Intercept)    3.556      3.202   1.110    0.267
## condition      4.057      4.450   0.912    0.363
## 
## Residual standard error: 47.07 on 446 degrees of freedom
## Multiple R-squared:  0.00186,    Adjusted R-squared:  -0.0003785 
## F-statistic: 0.8309 on 1 and 446 DF,  p-value: 0.3625
summary(lm(m$mental8_confidence~condition, data=m))
## 
## Call:
## lm(formula = m$mental8_confidence ~ condition, data = m)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.2511 -1.0880 -0.2511  0.9120  2.9120 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.08796    0.08861  23.563   <2e-16 ***
## condition    0.16312    0.12327   1.323    0.186    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.302 on 445 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.00392,    Adjusted R-squared:  0.001681 
## F-statistic: 1.751 on 1 and 445 DF,  p-value: 0.1864

Social outcomes. Very low reliability so analyze them separately.

m$social1_normal_act[m$social1_normal_act>5]<-NA
m$social3_interact_closely[m$social3_interact_closely>5]<-NA
m$social5_support[m$social5_support>5]<-NA
m$social6_comfortable[m$social6_comfortable>5]<-NA

alpha(m[,c(64:69)], na.rm=TRUE, check.keys=TRUE)
## Warning in alpha(m[, c(64:69)], na.rm = TRUE, check.keys = TRUE): Some items were negatively correlated with total scale and were automatically reversed.
##  This is indicated by a negative sign for the variable name.
## 
## Reliability analysis   
## Call: alpha(x = m[, c(64:69)], na.rm = TRUE, check.keys = TRUE)
## 
##   raw_alpha std.alpha G6(smc) average_r S/N   ase mean   sd
##       0.64      0.66    0.65      0.25   2 0.026  4.2 0.73
## 
##  lower alpha upper     95% confidence boundaries
## 0.59 0.64 0.69 
## 
##  Reliability if an item is dropped:
##                          raw_alpha std.alpha G6(smc) average_r S/N
## social1_normal_act-           0.63      0.65    0.63      0.27 1.9
## social2_rejected-             0.60      0.61    0.58      0.24 1.6
## social3_interact_closely      0.60      0.63    0.61      0.25 1.7
## social4_mean_words-           0.60      0.61    0.58      0.24 1.5
## social5_support               0.54      0.58    0.54      0.22 1.4
## social6_comfortable           0.62      0.65    0.61      0.27 1.8
##                          alpha se
## social1_normal_act-         0.027
## social2_rejected-           0.029
## social3_interact_closely    0.030
## social4_mean_words-         0.030
## social5_support             0.034
## social6_comfortable         0.028
## 
##  Item statistics 
##                            n raw.r std.r r.cor r.drop mean   sd
## social1_normal_act-      447  0.55  0.54  0.38   0.29  4.1 1.27
## social2_rejected-        448  0.55  0.63  0.53   0.39  4.6 0.88
## social3_interact_closely 447  0.65  0.60  0.46   0.38  3.6 1.44
## social4_mean_words-      448  0.57  0.64  0.54   0.40  4.6 0.89
## social5_support          447  0.71  0.69  0.63   0.52  4.0 1.22
## social6_comfortable      447  0.61  0.55  0.41   0.32  4.0 1.47
## 
## Non missing response frequency for each item
##                             1    2    3    4    5 miss
## social1_normal_act       0.61 0.11 0.12 0.09 0.06    0
## social2_rejected         0.78 0.10 0.08 0.03 0.02    0
## social3_interact_closely 0.13 0.11 0.16 0.18 0.41    0
## social4_mean_words       0.80 0.08 0.07 0.04 0.01    0
## social5_support          0.08 0.05 0.12 0.26 0.50    0
## social6_comfortable      0.15 0.05 0.08 0.14 0.58    0
#Disease interfered with social activity arginally more often in intervention condition.
summary(lm(m$social1_normal_act~m$condition))
## 
## Call:
## lm(formula = m$social1_normal_act ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -0.9610 -0.9610 -0.7639  1.0390  3.2361 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.76389    0.08592  20.529   <2e-16 ***
## m$condition  0.19715    0.11952   1.649   0.0998 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.263 on 445 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.006077,   Adjusted R-squared:  0.003843 
## F-statistic: 2.721 on 1 and 445 DF,  p-value: 0.09976
#People felt rejected MORE often in the intervention condition. 
summary(lm(m$social2_rejected~m$condition))
## 
## Call:
## lm(formula = m$social2_rejected ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -0.5388 -0.5388 -0.2685 -0.2685  3.7315 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.26852    0.05924  21.413  < 2e-16 ***
## m$condition  0.27027    0.08232   3.283  0.00111 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8706 on 446 degrees of freedom
## Multiple R-squared:  0.0236, Adjusted R-squared:  0.02141 
## F-statistic: 10.78 on 1 and 446 DF,  p-value: 0.001107
summary(lm(m$social3_interact_closely~m$condition))
## 
## Call:
## lm(formula = m$social3_interact_closely ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -2.7407 -0.7407  0.2593  1.2593  1.4632 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.74074    0.09809  38.134   <2e-16 ***
## m$condition -0.20394    0.13646  -1.495    0.136    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.442 on 445 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.004995,   Adjusted R-squared:  0.002759 
## F-statistic: 2.234 on 1 and 445 DF,  p-value: 0.1357
#People were spoken to in a mean way more often in the intervention condition. 
summary(lm(m$social4_mean_words~m$condition))
## 
## Call:
## lm(formula = m$social4_mean_words ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -0.4741 -0.4741 -0.3102 -0.3102  3.6898 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.31019    0.06011  21.797   <2e-16 ***
## m$condition  0.16395    0.08353   1.963   0.0503 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8834 on 446 degrees of freedom
## Multiple R-squared:  0.008565,   Adjusted R-squared:  0.006342 
## F-statistic: 3.853 on 1 and 446 DF,  p-value: 0.05028
summary(lm(m$social5_support~m$condition))
## 
## Call:
## lm(formula = m$social5_support ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.1944 -0.1944  0.0866  0.8056  1.0866 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  4.19444    0.08287  50.615   <2e-16 ***
## m$condition -0.28102    0.11528  -2.438   0.0152 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.218 on 445 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.01318,    Adjusted R-squared:  0.01096 
## F-statistic: 5.943 on 1 and 445 DF,  p-value: 0.01517
summary(lm(m$social6_comfortable~m$condition))
## 
## Call:
## lm(formula = m$social6_comfortable ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.1435 -0.8009  0.8565  1.1991  1.1991 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   4.1435     0.0997  41.561   <2e-16 ***
## m$condition  -0.3427     0.1387  -2.471   0.0139 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.465 on 445 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.01353,    Adjusted R-squared:  0.01132 
## F-statistic: 6.104 on 1 and 445 DF,  p-value: 0.01386
#People spoke to fewer other people in the intervention condition. 
summary(lm(m$social7_spoken_ppl_num~m$condition))
## 
## Call:
## lm(formula = m$social7_spoken_ppl_num ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -12.616  -7.616  -4.480  -0.616 237.384 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   12.616      1.357   9.298   <2e-16 ***
## m$condition   -4.271      1.885  -2.265    0.024 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 19.94 on 446 degrees of freedom
## Multiple R-squared:  0.01137,    Adjusted R-squared:  0.009158 
## F-statistic: 5.131 on 1 and 446 DF,  p-value: 0.02398
summary(lm(m$social8_gatherings~m$condition))
## 
## Call:
## lm(formula = m$social8_gatherings ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -0.3426 -0.3426 -0.2931 -0.2931  5.6574 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  0.34259    0.04978   6.882 2.01e-11 ***
## m$condition -0.04949    0.06917  -0.715    0.475    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.7316 on 446 degrees of freedom
## Multiple R-squared:  0.001146,   Adjusted R-squared:  -0.001093 
## F-statistic: 0.5118 on 1 and 446 DF,  p-value: 0.4747

For comprehension questions, re-coding “don’t know” as incorrect, incorrect answers as 0 and correct as 1. No effect of condition on comprehension.

m$comprehension1[m$comprehension1==2 |m$comprehension1==888]<-0
m$comprehension1[m$comprehension1==999]<-NA

m$comprehension2[m$comprehension2==1 |m$comprehension2==888]<-0
m$comprehension2[m$comprehension2==2 ]<-1

m$comprehension3[m$comprehension3==2 |m$comprehension3==888]<-0

m$comprehension4[m$comprehension4==1 |m$comprehension4==888]<-0
m$comprehension4[m$comprehension4==2 ]<-1

m$comprehension5[m$comprehension5==1 |m$comprehension5==888]<-0
m$comprehension5[m$comprehension5==2 ]<-1
m$comprehension5[m$comprehension5==999 ]<-NA

m$comprehension6[m$comprehension6==2 | m$comprehension6==888]<-0
m$comprehension6[m$comprehension6==999 ]<-NA

m$comprehension7[m$comprehension7==999 ]<-NA
m$comprehension7[m$comprehension7==2 | m$comprehension7==888]<-0

m$comprehension8[m$comprehension8==1 |m$comprehension8==888]<-0
m$comprehension8[m$comprehension8==2 ]<-1
m$comprehension8[m$comprehension8==999 ]<-NA

m$comprehension9[m$comprehension9==1 |m$comprehension9==888]<-0
m$comprehension9[m$comprehension9==2 ]<-1
m$comprehension9[m$comprehension9==999 ]<-NA


m$comp<-(m$comprehension1 + m$comprehension2+m$comprehension4+m$comprehension3+m$comprehension5+m$comprehension6+m$comprehension7+m$comprehension8+m$comprehension9)

summary(lm(comp~condition, data=m))
## 
## Call:
## lm(formula = comp ~ condition, data = m)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -14.34  -8.34  -7.34   0.08 986.66 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)  
## (Intercept)    7.920      4.598   1.722   0.0857 .
## condition      8.422      6.387   1.319   0.1880  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 66.95 on 438 degrees of freedom
##   (8 observations deleted due to missingness)
## Multiple R-squared:  0.003954,   Adjusted R-squared:  0.00168 
## F-statistic: 1.739 on 1 and 438 DF,  p-value: 0.188

High alpha for engagement. No effect of condition.

alpha(m[,c(189:195)], na.rm=TRUE, check.keys=TRUE)
## 
## Reliability analysis   
## Call: alpha(x = m[, c(189:195)], na.rm = TRUE, check.keys = TRUE)
## 
##   raw_alpha std.alpha G6(smc) average_r S/N    ase mean sd
##       0.91      0.92    0.98      0.62  12 0.0069  8.5 57
## 
##  lower alpha upper     95% confidence boundaries
## 0.89 0.91 0.92 
## 
##  Reliability if an item is dropped:
##                         raw_alpha std.alpha G6(smc) average_r  S/N
## engagement1_interest         0.89      0.90    0.96      0.61  9.2
## engagement2_participate      0.89      0.90    0.96      0.61  9.2
## engagement3_attention        0.89      0.91    0.97      0.62  9.6
## engagment4_engrossed         0.88      0.90    0.97      0.61  9.4
## engagment5_entertaining      0.88      0.90    0.97      0.61  9.4
## engagement6_focus            0.89      0.91    0.96      0.64 10.5
## engagement7_questions        0.92      0.93    0.98      0.68 12.8
##                         alpha se
## engagement1_interest      0.0079
## engagement2_participate   0.0079
## engagement3_attention     0.0083
## engagment4_engrossed      0.0089
## engagment5_entertaining   0.0089
## engagement6_focus         0.0081
## engagement7_questions     0.0059
## 
##  Item statistics 
##                           n raw.r std.r r.cor r.drop mean sd
## engagement1_interest    448  0.82  0.87  0.88   0.78  6.0 47
## engagement2_participate 448  0.82  0.87  0.88   0.78  5.8 47
## engagement3_attention   448  0.83  0.84  0.83   0.77  8.3 66
## engagment4_engrossed    448  0.88  0.86  0.87   0.82 10.4 81
## engagment5_entertaining 448  0.88  0.86  0.87   0.82 10.3 81
## engagement6_focus       448  0.81  0.79  0.77   0.73 10.2 81
## engagement7_questions   448  0.68  0.67  0.62   0.54  8.6 81
m$engaged<-(m$engagement1_interest+m$engagement2_participate+m$engagement3_attention+m$engagment4_engrossed+m$engagment5_entertaining+m$engagement6_focus+m$engagement7_questions)/7

summary(lm(engaged~condition, m))
## 
## Call:
## lm(formula = engaged ~ condition, data = m)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -11.07  -8.36  -3.42  -1.63 986.93 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)   
## (Intercept)   12.073      3.864   3.124   0.0019 **
## condition     -6.871      5.370  -1.279   0.2014   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 56.79 on 446 degrees of freedom
## Multiple R-squared:  0.003657,   Adjusted R-squared:  0.001423 
## F-statistic: 1.637 on 1 and 446 DF,  p-value: 0.2014

Desensitization. Low alpha, so look at the items individually.

alpha(m[,c(196:201)], na.rm=TRUE, check.keys=TRUE)
## Warning in alpha(m[, c(196:201)], na.rm = TRUE, check.keys = TRUE): Some items were negatively correlated with total scale and were automatically reversed.
##  This is indicated by a negative sign for the variable name.
## 
## Reliability analysis   
## Call: alpha(x = m[, c(196:201)], na.rm = TRUE, check.keys = TRUE)
## 
##   raw_alpha std.alpha G6(smc) average_r  S/N   ase mean sd
##     0.0058      0.16    0.14      0.03 0.19 0.068  335 19
## 
##  lower alpha upper     95% confidence boundaries
## -0.13 0.01 0.14 
## 
##  Reliability if an item is dropped:
##                          raw_alpha std.alpha G6(smc) average_r   S/N
## desensitize1_worry         0.00917     0.165   0.141     0.038 0.198
## desensitize2_cure          0.00649     0.155   0.133     0.035 0.184
## desensitize3_afraid        0.00393     0.081   0.068     0.017 0.088
## desensitize4_identity      0.00467     0.146   0.125     0.033 0.170
## desensitize5_healthy-      0.00428     0.079   0.067     0.017 0.086
## desensitize6_threatened-   0.00032     0.177   0.152     0.041 0.215
##                          alpha se
## desensitize1_worry          0.066
## desensitize2_cure           0.066
## desensitize3_afraid         0.071
## desensitize4_identity       0.067
## desensitize5_healthy-       0.071
## desensitize6_threatened-    0.066
## 
##  Item statistics 
##                            n raw.r std.r r.cor  r.drop  mean    sd
## desensitize1_worry       448 0.406  0.41 0.079 -0.0017   4.8 47.09
## desensitize2_cure        448 0.576  0.42 0.108  0.0015   8.0 66.45
## desensitize3_afraid      448 0.087  0.49 0.308  0.0757   2.5  1.31
## desensitize4_identity    448 0.410  0.43 0.134  0.0024   4.2 47.12
## desensitize5_healthy-    448 0.095  0.49 0.312  0.0872 995.8  0.95
## desensitize6_threatened- 448 0.580  0.40 0.044  0.0050 993.0 66.51
m$desensitize1_worry[m$desensitize1_worry>5]<-NA
m$desensitize2_cure[m$desensitize2_cure>5]<-NA
m$desensitize4_identity[m$desensitize4_identity>5]<-NA
m$desensitize6_threatened[m$desensitize6_threatened>5]<-NA

#Marginally higher worry in the intervention condition. 
summary(lm(m$desensitize1_worry ~ m$condition))
## 
## Call:
## lm(formula = m$desensitize1_worry ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.6897 -1.4930  0.3103  1.3103  2.5070 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.49302    0.08513  29.284   <2e-16 ***
## m$condition  0.19663    0.11817   1.664   0.0968 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.248 on 445 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.006184,   Adjusted R-squared:  0.00395 
## F-statistic: 2.769 on 1 and 445 DF,  p-value: 0.09682
summary(lm(m$desensitize2_cure ~ m$condition))
## 
## Call:
## lm(formula = m$desensitize2_cure ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -2.5694 -0.5694  0.4306  0.5000  1.5000 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.56944    0.07939  44.961   <2e-16 ***
## m$condition -0.06944    0.11055  -0.628     0.53    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.167 on 444 degrees of freedom
##   (2 observations deleted due to missingness)
## Multiple R-squared:  0.0008879,  Adjusted R-squared:  -0.001362 
## F-statistic: 0.3946 on 1 and 444 DF,  p-value: 0.5302
#Marginally more afraid in the intervention condition. 
summary(lm(m$desensitize3_afraid ~ m$condition))
## 
## Call:
## lm(formula = m$desensitize3_afraid ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.5862 -1.3657 -0.3657  0.8291  2.6343 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.36574    0.08896  26.593   <2e-16 ***
## m$condition  0.22047    0.12362   1.783   0.0752 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.307 on 446 degrees of freedom
## Multiple R-squared:  0.007081,   Adjusted R-squared:  0.004855 
## F-statistic: 3.181 on 1 and 446 DF,  p-value: 0.0752
summary(lm(m$desensitize4_identity ~ m$condition))
## 
## Call:
## lm(formula = m$desensitize4_identity ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.0000 -1.0000 -0.8657  1.0000  3.1343 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.86574    0.07987  23.361   <2e-16 ***
## m$condition  0.13426    0.11110   1.208    0.228    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.174 on 445 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.003271,   Adjusted R-squared:  0.001031 
## F-statistic:  1.46 on 1 and 445 DF,  p-value: 0.2275
summary(lm(m$desensitize5_healthy ~ m$condition))
## 
## Call:
## lm(formula = m$desensitize5_healthy ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.2315 -0.2315 -0.1552  0.8448  0.8448 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  4.23148    0.06499  65.114   <2e-16 ***
## m$condition -0.07631    0.09031  -0.845    0.399    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.9551 on 446 degrees of freedom
## Multiple R-squared:  0.001598,   Adjusted R-squared:  -0.0006401 
## F-statistic: 0.714 on 1 and 446 DF,  p-value: 0.3986
summary(lm(m$desensitize6_threatened ~ m$condition))
## 
## Call:
## lm(formula = m$desensitize6_threatened ~ m$condition)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.6652 -1.5093 -0.5093  1.3348  2.4907 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.50926    0.08563  29.304   <2e-16 ***
## m$condition  0.15596    0.11924   1.308    0.192    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.258 on 444 degrees of freedom
##   (2 observations deleted due to missingness)
## Multiple R-squared:  0.003838,   Adjusted R-squared:  0.001594 
## F-statistic: 1.711 on 1 and 444 DF,  p-value: 0.1916

Story-telling (potential mechanism) - moderately high reliability, so analyze as a composite. No effect of condition.

alpha(m[,c(202:207)], na.rm=TRUE, check.keys=TRUE)
## 
## Reliability analysis   
## Call: alpha(x = m[, c(202:207)], na.rm = TRUE, check.keys = TRUE)
## 
##   raw_alpha std.alpha G6(smc) average_r S/N   ase mean sd
##       0.83      0.87    0.89      0.54 6.9 0.012  8.5 53
## 
##  lower alpha upper     95% confidence boundaries
## 0.81 0.83 0.86 
## 
##  Reliability if an item is dropped:
##               raw_alpha std.alpha G6(smc) average_r S/N alpha se
## storytelling1      0.83      0.88    0.90      0.59 7.1    0.013
## storytelling2      0.78      0.82    0.81      0.47 4.4    0.016
## storytelling3      0.83      0.88    0.89      0.59 7.0    0.013
## storytelling4      0.80      0.86    0.88      0.56 6.3    0.015
## storytelling5      0.81      0.86    0.88      0.55 6.1    0.014
## storytelling6      0.78      0.82    0.81      0.47 4.4    0.016
## 
##  Item statistics 
##                 n raw.r std.r r.cor r.drop mean sd
## storytelling1 448  0.68  0.67  0.56   0.50  9.6 81
## storytelling2 448  0.88  0.93  0.98   0.85  5.6 47
## storytelling3 448  0.73  0.68  0.58   0.54 12.3 94
## storytelling4 448  0.76  0.74  0.65   0.62 10.5 81
## storytelling5 448  0.73  0.76  0.69   0.60  8.2 66
## storytelling6 448  0.88  0.93  0.98   0.84  4.8 47
m$storytelling1[m$storytelling1>5]<-NA
m$storytelling2[m$storytelling2>5]<-NA
m$storytelling3[m$storytelling3>5]<-NA
m$storytelling4[m$storytelling4>5]<-NA
m$storytelling5[m$storytelling5>5]<-NA
m$storytelling6[m$storytelling6>5]<-NA

m$story<-(m$storytelling1+m$storytelling2+m$storytelling3+m$storytelling4+m$storytelling5+m$storytelling6)/6

summary(lm(story~condition, m))
## 
## Call:
## lm(formula = story ~ condition, data = m)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -2.1580 -0.4716  0.1019  0.5087  1.6950 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.32464    0.04767  69.742   <2e-16 ***
## condition   -0.01970    0.06608  -0.298    0.766    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.6925 on 438 degrees of freedom
##   (8 observations deleted due to missingness)
## Multiple R-squared:  0.0002028,  Adjusted R-squared:  -0.00208 
## F-statistic: 0.08884 on 1 and 438 DF,  p-value: 0.7658

IGNORE THIS FOR NOW! HAVEN’T UPDATED THE INTERACTION ANALYSES YET.

Now looking for interactions between condition and demographic variables.

#merge

summary(lm(overall.health ~ condition * Age + condition*Gender + condition*education, all))

summary(lm(symptoms ~ condition * Age + condition*Gender + condition*education, all))

summary(lm(unable.to.work ~ condition * Age + condition*Gender + condition*education, all))

summary(lm(unable.do.chores ~ condition * Age + condition*Gender + condition*education, all))

summary(lm(money.stress ~ condition * Age + condition*Gender + condition*education, all))

summary(lm(depressed ~ condition * Age + condition*Gender + condition*education, all))

summary(lm(hopeful ~ condition * Age + condition*Gender + condition*education, all))

summary(lm(social ~ condition * Age + condition*Gender + condition*education, all))

summary(lm(rejected ~ condition * Age + condition*Gender + condition*education, all))

Now looking for interactions between condition and other covariates. Frequent interaction with smoking, but only 16 people smoke so probably not worth interpreting yet.

m$tech2_computer[m$tech2_computer>1]<-NA
m$tech6_tablet_freq[m$tech6_tablet_freq>4]<-NA
m$drink_alcohol[m$drink_alcohol>1]<-NA

summary(lm(overall.health ~ condition * tech1_watch_tv + condition*tech2_computer + condition*tech3_smartphone + condition*tech5_tablet + condition * smoker  + condition*drink_alcohol , all))

#some interactions are significant here 
summary(lm(symptoms ~ condition * tech1_watch_tv + condition*tech2_computer + condition*tech3_smartphone + condition*tech5_tablet + condition * smoker  + condition*drink_alcohol , all))

#here too
summary(lm(unable.to.work ~ condition * tech1_watch_tv + condition*tech2_computer + condition*tech3_smartphone + condition*tech5_tablet + condition * smoker  + condition*drink_alcohol , all))

#here too
summary(lm(unable.do.chores ~ condition * tech1_watch_tv + condition*tech2_computer + condition*tech3_smartphone + condition*tech5_tablet + condition * smoker  + condition*drink_alcohol , all))

#here too
summary(lm(money.stress~ condition * tech1_watch_tv + condition*tech2_computer + condition*tech3_smartphone + condition*tech5_tablet + condition * smoker  + condition*drink_alcohol , all))

#here too
summary(lm(depressed ~ condition * tech1_watch_tv + condition*tech2_computer + condition*tech3_smartphone + condition*tech5_tablet + condition * smoker  + condition*drink_alcohol , all))

summary(lm(hopeful ~ condition * tech1_watch_tv + condition*tech2_computer + condition*tech3_smartphone + condition*tech5_tablet + condition * smoker  + condition*drink_alcohol , all))

summary(lm(social ~ condition * tech1_watch_tv + condition*tech2_computer + condition*tech3_smartphone + condition*tech5_tablet + condition * smoker  + condition*drink_alcohol , all))

summary(lm(rejected ~ condition * tech1_watch_tv + condition*tech2_computer + condition*tech3_smartphone + condition*tech5_tablet + condition * smoker  + condition*drink_alcohol , all))

Following up on interactions

table(m$tech1_watch_tv)
TVeveryday<-subset(all, tech1_watch_tv==1)
TVnoteveryday<-subset(all, tech1_watch_tv!=1)

summary(lm(money.stress ~ condition, data=TVeveryday))
summary(lm(money.stress ~ condition, data=TVnoteveryday))

#condition had the desired effect among people who don't watch TV every day (although the sample is small, about 75)
summary(lm(unable.do.chores ~ condition, data=TVeveryday))
summary(lm(unable.do.chores ~ condition, data=TVnoteveryday))