Follow-up on the previous study. Last study, we found that self-reported creativity predicted compromise (more creative = less compromise) and number of reasons listed for the disjunctive question. We found no relationship between a behavioral-based measure of the RAT at the beginning of the survey with decisions.

This study was a follow-up where we 1) included a new measure of creativity: pasta names, 2) included measures of intelligence and reasoning, and 3) changed the order of when we measured about how much one enjoyed hawaiii vacations (this could have messed things up).

Creativity Measures

The alpha for the 4 creativity items was 0.78.

Mean self-reported creativity was 4.56 out of 7.

SD self-reported creativity was 1.04.

Mean # of pasta names generated was 6.92.

Mean # of of pasta names ending in “i” was 2.14.

Thus, mean percent of pasta names ending in “i” was 30%.

Percent people who wrote at least one name ending in “i” was 61.8%.

Distributions

qplot(create_total, data=d, xlab = "Self-reported Creativity")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## Warning: Removed 1 rows containing non-finite values (stat_bin).

qplot(total_pasta, data=d, xlab = "Total # of Pasta Names")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

qplot(perc_i, data=d, xlab = "% names ending in i")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## Warning: Removed 3 rows containing non-finite values (stat_bin).

Intelligence/Reasoning Measures

Mean CRT correct was 0.96 out of 3.

Mean logic type 1: correctly generalizing from group was 1.53 correct out of 2.

Mean logic type 2: NOT incorrectly generalizing from group was 0.98 correct out of 2.

Mean total logic correct was 2.5 correct out of 4.

Examining the relationship between self-reported creativity, the pasta task, and the intelligence/reasnoning measures

While the number of ideas listed in the pasta task does positively relate to self-reported creativity (p<.001), the percent of the names that end in i does not (although the realtionship is negative, p = .25). This is interesting. We need to think about what this measure is capturing.

Neither self-reported creativity nor number of pasta names relate to CRT or any of the logic questions.

Percent of names that end in i psotiviely relates to logic questions, but has no relationship to CRT.

CRT positively relates to both types of logic questions, but more so with logic type 1 questions (correctly generalizing to the group).

See below for correlations

#self-reported creativity
cor.test(d$create_total, d$total_pasta)
## 
##  Pearson's product-moment correlation
## 
## data:  d$create_total and d$total_pasta
## t = 7.1727, df = 760, p-value = 1.747e-12
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.1840676 0.3171490
## sample estimates:
##       cor 
## 0.2517983
cor.test(d$create_total, d$perc_i)
## 
##  Pearson's product-moment correlation
## 
## data:  d$create_total and d$perc_i
## t = -1.15, df = 757, p-value = 0.2505
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.11258955  0.02948891
## sample estimates:
##         cor 
## -0.04176144
cor.test(d$create_total, d$crt_total) 
## 
##  Pearson's product-moment correlation
## 
## data:  d$create_total and d$crt_total
## t = 1.2526, df = 760, p-value = 0.2107
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.02571452  0.11603911
## sample estimates:
##        cor 
## 0.04539078
cor.test(d$create_total, d$logic_type1) 
## 
##  Pearson's product-moment correlation
## 
## data:  d$create_total and d$logic_type1
## t = -0.85167, df = 760, p-value = 0.3947
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.10167800  0.04023205
## sample estimates:
##         cor 
## -0.03087858
cor.test(d$create_total, d$logic_type2) 
## 
##  Pearson's product-moment correlation
## 
## data:  d$create_total and d$logic_type2
## t = -0.83594, df = 760, p-value = 0.4035
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.10111353  0.04080145
## sample estimates:
##         cor 
## -0.03030879
cor.test(d$create_total, d$logic_all) 
## 
##  Pearson's product-moment correlation
## 
## data:  d$create_total and d$logic_all
## t = -1.0926, df = 760, p-value = 0.2749
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.11031378  0.03150938
## sample estimates:
##         cor 
## -0.03960165
#pasta task & reasoning/intelligence
cor.test(d$total_pasta, d$logic_type1)
## 
##  Pearson's product-moment correlation
## 
## data:  d$total_pasta and d$logic_type1
## t = 0.82301, df = 761, p-value = 0.4108
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.04124216  0.10058388
## sample estimates:
##        cor 
## 0.02982095
cor.test(d$total_pasta, d$logic_type2)
## 
##  Pearson's product-moment correlation
## 
## data:  d$total_pasta and d$logic_type2
## t = -1.7464, df = 761, p-value = 0.08115
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.133557186  0.007830668
## sample estimates:
##         cor 
## -0.06318027
cor.test(d$total_pasta, d$crt_total)
## 
##  Pearson's product-moment correlation
## 
## data:  d$total_pasta and d$crt_total
## t = -1.0456, df = 761, p-value = 0.2961
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.10856104  0.03318798
## sample estimates:
##         cor 
## -0.03787706
cor.test(d$perc_i, d$logic_type1) #significant
## 
##  Pearson's product-moment correlation
## 
## data:  d$perc_i and d$logic_type1
## t = 3.8565, df = 758, p-value = 0.0001248
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.06827938 0.20778759
## sample estimates:
##       cor 
## 0.1387216
cor.test(d$perc_i, d$logic_type2)
## 
##  Pearson's product-moment correlation
## 
## data:  d$perc_i and d$logic_type2
## t = 2.2428, df = 758, p-value = 0.02519
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.01013765 0.15143638
## sample estimates:
##        cor 
## 0.08119497
cor.test(d$perc_i, d$crt_total)
## 
##  Pearson's product-moment correlation
## 
## data:  d$perc_i and d$crt_total
## t = 1.534, df = 758, p-value = 0.1254
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.01554585  0.12624788
## sample estimates:
##       cor 
## 0.0556315
#reasoning and intelligence
cor.test(d$logic_type2, d$logic_type1)
## 
##  Pearson's product-moment correlation
## 
## data:  d$logic_type2 and d$logic_type1
## t = 4.8894, df = 761, p-value = 1.234e-06
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.1048435 0.2424927
## sample estimates:
##       cor 
## 0.1745206
cor.test(d$crt_total, d$logic_type1)
## 
##  Pearson's product-moment correlation
## 
## data:  d$crt_total and d$logic_type1
## t = 4.75, df = 761, p-value = 2.432e-06
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.09991604 0.23780030
## sample estimates:
##       cor 
## 0.1696885
cor.test(d$crt_total, d$logic_type2)
## 
##  Pearson's product-moment correlation
## 
## data:  d$crt_total and d$logic_type2
## t = 2.884, df = 761, p-value = 0.004037
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.03324864 0.17367320
## sample estimates:
##       cor 
## 0.1039791

Reason-based choice

Disjunctive

48% said that they would delay the choice of vacation.

We also recorded the number of reasons listed for choosing to go on the vacation. People on average listed 3.48 reasons, with a SD of 1.61

Compromise

Out of the 3 choices, the average number of compromise options chosen was 1.09

Percent compromise for hotel: 54% Dogs: 30% Ticket: 25%

Reason-based choice and Creativity

Disjunctive

We do get a marginal promising effect of self-reported creativity on number of reasons generated:

summary(lm(reasons_total ~ create_total, data=d))
## 
## Call:
## lm(formula = reasons_total ~ create_total, data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -2.7213 -1.3286 -0.4022  0.6469  8.6960 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   3.03403    0.26230   11.57   <2e-16 ***
## create_total  0.09818    0.05611    1.75   0.0806 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.607 on 752 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.004055,   Adjusted R-squared:  0.002731 
## F-statistic: 3.062 on 1 and 752 DF,  p-value: 0.08056
ggplot(d, aes(x= create_total, y = reasons_total)) +
      geom_point() +
      geom_smooth(method="lm", formula = y ~ x)
## Warning: Removed 1 rows containing non-finite values (stat_smooth).
## Warning: Removed 1 rows containing missing values (geom_point).

summary(lm(reasons_total ~ create_total*disjunction2, data=d))
## 
## Call:
## lm(formula = reasons_total ~ create_total * disjunction2, data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -2.9415 -0.9291 -0.0957  1.0292  8.2631 
## 
## Coefficients:
##                                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                     3.57618    0.34790  10.279   <2e-16 ***
## create_total                    0.05845    0.07481   0.781   0.4348    
## disjunction2delay              -1.23010    0.51262  -2.400   0.0167 *  
## create_total:disjunction2delay  0.10813    0.10960   0.987   0.3242    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.565 on 750 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.05772,    Adjusted R-squared:  0.05396 
## F-statistic: 15.32 on 3 and 750 DF,  p-value: 1.113e-09

As creativity increases, the number of reasons listed increases. This is not moderated by the choice they made (to make a choice or delay).

Now let’s look at the pasta names and reasons listed:

summary(lm(reasons_total ~ i_num + total_pasta, data=d))
## 
## Call:
## lm(formula = reasons_total ~ i_num + total_pasta, data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.3025 -1.0761 -0.2561  0.8298  8.5536 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.83048    0.11653  24.290  < 2e-16 ***
## i_num       -0.06345    0.02332  -2.721  0.00666 ** 
## total_pasta  0.11323    0.01685   6.720 3.58e-11 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.564 on 752 degrees of freedom
## Multiple R-squared:  0.05731,    Adjusted R-squared:  0.05481 
## F-statistic: 22.86 on 2 and 752 DF,  p-value: 2.303e-10

Interesting! the number of names ending in “i” negatively predict number of reasons listed (when controlling for total number of names listed).

Now let’s look at the reasoning/intelligence measures:

CRT positively predicts (p = .06) number of ideas listed (but controlling for crt doesn’t attenuate effect of self-reported creativity). Controlling for CRT also doesn’t attenuate effect of percent names ending in i negatively predicting reasons generated. Logic type 1 is a significant prerdcitor, logic type 2 is not.

See below for regressions

summary(glm(reasons_total ~ crt_total, data=d))
## 
## Call:
## glm(formula = reasons_total ~ crt_total, data = d)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -2.7148  -1.3706  -0.3706   0.6294   8.6294  
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.37064    0.08415  40.056   <2e-16 ***
## crt_total    0.11471    0.06305   1.819   0.0692 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for gaussian family taken to be 2.57893)
## 
##     Null deviance: 1950.5  on 754  degrees of freedom
## Residual deviance: 1941.9  on 753  degrees of freedom
## AIC: 2861.9
## 
## Number of Fisher Scoring iterations: 2
summary(glm(reasons_total ~ crt_total + create_total, data=d))
## 
## Call:
## glm(formula = reasons_total ~ crt_total + create_total, data = d)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -2.7145  -1.2770  -0.3706   0.7230   8.7932  
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   2.94928    0.26643  11.069   <2e-16 ***
## crt_total     0.10984    0.06308   1.741   0.0820 .  
## create_total  0.09363    0.05609   1.669   0.0955 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for gaussian family taken to be 2.575927)
## 
##     Null deviance: 1950.2  on 753  degrees of freedom
## Residual deviance: 1934.5  on 751  degrees of freedom
##   (1 observation deleted due to missingness)
## AIC: 2858.2
## 
## Number of Fisher Scoring iterations: 2
summary(glm(reasons_total ~ logic_type1, data=d))
## 
## Call:
## glm(formula = reasons_total ~ logic_type1, data = d)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -2.5624  -1.2093  -0.3858   0.6142   8.6142  
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.20926    0.15687  20.458   <2e-16 ***
## logic_type1  0.17658    0.09467   1.865   0.0625 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for gaussian family taken to be 2.578356)
## 
##     Null deviance: 1950.5  on 754  degrees of freedom
## Residual deviance: 1941.5  on 753  degrees of freedom
## AIC: 2861.7
## 
## Number of Fisher Scoring iterations: 2
summary(glm(reasons_total ~ logic_type2, data=d))
## 
## Call:
## glm(formula = reasons_total ~ logic_type2, data = d)
## 
## Deviance Residuals: 
##    Min      1Q  Median      3Q     Max  
## -2.527  -1.432  -0.432   0.568   8.568  
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.52674    0.08973  39.302   <2e-16 ***
## logic_type2 -0.04739    0.07013  -0.676    0.499    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for gaussian family taken to be 2.588698)
## 
##     Null deviance: 1950.5  on 754  degrees of freedom
## Residual deviance: 1949.3  on 753  degrees of freedom
## AIC: 2864.7
## 
## Number of Fisher Scoring iterations: 2
summary(glm(reasons_total ~ crt_total + perc_i + total_pasta, data=d))
## 
## Call:
## glm(formula = reasons_total ~ crt_total + perc_i + total_pasta, 
##     data = d)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -3.1488  -1.0769  -0.2215   0.8659   8.6989  
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.79361    0.14065  19.863  < 2e-16 ***
## crt_total    0.13826    0.06194   2.232   0.0259 *  
## perc_i      -0.32527    0.17677  -1.840   0.0662 .  
## total_pasta  0.09361    0.01473   6.354 3.65e-10 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for gaussian family taken to be 2.454542)
## 
##     Null deviance: 1949.7  on 751  degrees of freedom
## Residual deviance: 1836.0  on 748  degrees of freedom
##   (3 observations deleted due to missingness)
## AIC: 2815.3
## 
## Number of Fisher Scoring iterations: 2

Summary: no effect of creativity on choice in the disjunction paradigm. This is a replication of study 1. We do get a marginal effect of self-reported creativity on number of reasons people generated, which is in line with our thinking. We get a significant effect of percent of names ending in i negatively predicting the number of reasons generated.

Compromise

We replicate the effect on compromise! When self-reported creativity increases, people compromise less. However, there is no relationship between total number of ideas generated nor number of ideas generated that end in i with compromise. We need to think about what exactly we are measuring with self-reported creativity.

See below for regressions

summary(lm(comp_total ~ create_total ,data=d))
## 
## Call:
## lm(formula = comp_total ~ create_total, data = d)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.43529 -0.90525 -0.07263  0.81578  2.06685 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   1.60267    0.14236  11.258  < 2e-16 ***
## create_total -0.11159    0.03045  -3.664 0.000265 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8722 on 752 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.01754,    Adjusted R-squared:  0.01624 
## F-statistic: 13.43 on 1 and 752 DF,  p-value: 0.0002653
summary(lm(comp_total ~ total_pasta ,data=d))
## 
## Call:
## lm(formula = comp_total ~ total_pasta, data = d)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.13232 -1.06345 -0.09376  0.88972  1.95032 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.13232    0.06534  17.329   <2e-16 ***
## total_pasta -0.00551    0.00820  -0.672    0.502    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8791 on 753 degrees of freedom
## Multiple R-squared:  0.0005992,  Adjusted R-squared:  -0.000728 
## F-statistic: 0.4515 on 1 and 753 DF,  p-value: 0.5018
summary(lm(comp_total ~ perc_i ,data=d))
## 
## Call:
## lm(formula = comp_total ~ perc_i, data = d)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.11163 -1.07298 -0.09955  0.88837  1.93668 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.11163    0.04377  25.394   <2e-16 ***
## perc_i      -0.04831    0.09892  -0.488    0.625    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8792 on 750 degrees of freedom
##   (3 observations deleted due to missingness)
## Multiple R-squared:  0.0003179,  Adjusted R-squared:  -0.001015 
## F-statistic: 0.2385 on 1 and 750 DF,  p-value: 0.6254
summary(lm(comp_total ~ i_num + total_pasta + create_total ,data=d))
## 
## Call:
## lm(formula = comp_total ~ i_num + total_pasta + create_total, 
##     data = d)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.46613 -0.92377 -0.07622  0.82604  2.07642 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   1.596865   0.144249  11.070  < 2e-16 ***
## i_num        -0.006435   0.013038  -0.494 0.621742    
## total_pasta   0.004497   0.009696   0.464 0.642908    
## create_total -0.114139   0.031528  -3.620 0.000314 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8732 on 750 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.01795,    Adjusted R-squared:  0.01402 
## F-statistic: 4.568 on 3 and 750 DF,  p-value: 0.003519

Compromise with logic/CRT

Answering the logic questions correctly actualy positively predicts compromise. CRT does not predict compromise. Controlling for logic or CRT doesn’t impact the effect of self-reported creativity on compromise.

See below for regressions

summary(lm(comp_total ~ logic_type1 ,data=d))
## 
## Call:
## lm(formula = comp_total ~ logic_type1, data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.1564 -1.0215 -0.1564  0.8436  2.1135 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  0.88653    0.08553  10.366  < 2e-16 ***
## logic_type1  0.13494    0.05161   2.614  0.00912 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8754 on 753 degrees of freedom
## Multiple R-squared:  0.008996,   Adjusted R-squared:  0.00768 
## F-statistic: 6.835 on 1 and 753 DF,  p-value: 0.009116
summary(lm(comp_total ~ logic_type2 ,data=d))
## 
## Call:
## lm(formula = comp_total ~ logic_type2, data = d)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.15114 -1.04032 -0.09573  0.84886  1.95968 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.04032    0.04898  21.241   <2e-16 ***
## logic_type2  0.05541    0.03828   1.448    0.148    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8782 on 753 degrees of freedom
## Multiple R-squared:  0.002775,   Adjusted R-squared:  0.001451 
## F-statistic: 2.095 on 1 and 753 DF,  p-value: 0.1482
summary(lm(comp_total ~ logic_all ,data=d))
## 
## Call:
## lm(formula = comp_total ~ logic_all, data = d)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.19923 -0.98782 -0.05829  0.80077  2.08265 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  0.91735    0.07755  11.829   <2e-16 ***
## logic_all    0.07047    0.02820   2.499   0.0127 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8758 on 753 degrees of freedom
## Multiple R-squared:  0.008227,   Adjusted R-squared:  0.00691 
## F-statistic: 6.246 on 1 and 753 DF,  p-value: 0.01266
summary(lm(comp_total ~ crt_total ,data=d))
## 
## Call:
## lm(formula = comp_total ~ crt_total, data = d)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.14692 -1.06914 -0.09507  0.87900  1.93086 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.06914    0.04606  23.211   <2e-16 ***
## crt_total    0.02593    0.03451   0.751    0.453    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8791 on 753 degrees of freedom
## Multiple R-squared:  0.0007489,  Adjusted R-squared:  -0.0005781 
## F-statistic: 0.5644 on 1 and 753 DF,  p-value: 0.4527
summary(lm(comp_total ~ crt_total + create_total, data=d))
## 
## Call:
## lm(formula = comp_total ~ crt_total + create_total, data = d)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.44059 -0.90429 -0.07003  0.81616  2.07110 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   1.57810    0.14481  10.898  < 2e-16 ***
## crt_total     0.03184    0.03428   0.929 0.353292    
## create_total -0.11291    0.03049  -3.703 0.000228 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8723 on 751 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.01867,    Adjusted R-squared:  0.01606 
## F-statistic: 7.144 on 2 and 751 DF,  p-value: 0.0008442
summary(lm(comp_total ~ logic_all + create_total, data=d))
## 
## Call:
## lm(formula = comp_total ~ logic_all + create_total, data = d)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.49789 -0.90067 -0.06658  0.79766  2.16265 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   1.42265    0.16106   8.833  < 2e-16 ***
## logic_all     0.06632    0.02805   2.364 0.018323 *  
## create_total -0.10860    0.03039  -3.574 0.000374 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.8696 on 751 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.0248, Adjusted R-squared:  0.0222 
## F-statistic:  9.55 on 2 and 751 DF,  p-value: 8.025e-05

Here are the results for each individual item. Directional for each item, only significant (p<.0001) for dogs.

#hotel
summary(glm(comp_1 ~ create_total, family="binomial",data=d))
## 
## Call:
## glm(formula = comp_1 ~ create_total, family = "binomial", data = d)
## 
## Deviance Residuals: 
##    Min      1Q  Median      3Q     Max  
## -1.388  -1.237   1.043   1.108   1.219  
## 
## Coefficients:
##              Estimate Std. Error z value Pr(>|z|)  
## (Intercept)   0.64145    0.32964   1.946   0.0517 .
## create_total -0.10566    0.07039  -1.501   0.1333  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 1040.5  on 753  degrees of freedom
## Residual deviance: 1038.2  on 752  degrees of freedom
##   (1 observation deleted due to missingness)
## AIC: 1042.2
## 
## Number of Fisher Scoring iterations: 3
#dogs
summary(glm(comp_2 ~ create_total, family="binomial",data=d))
## 
## Call:
## glm(formula = comp_2 ~ create_total, family = "binomial", data = d)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -1.1873  -0.8437  -0.7931   1.3896   1.8790  
## 
## Coefficients:
##              Estimate Std. Error z value Pr(>|z|)    
## (Intercept)   0.45984    0.35556   1.293 0.195917    
## create_total -0.29107    0.07809  -3.727 0.000193 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 920.84  on 753  degrees of freedom
## Residual deviance: 906.59  on 752  degrees of freedom
##   (1 observation deleted due to missingness)
## AIC: 910.59
## 
## Number of Fisher Scoring iterations: 4
#tickets
summary(glm(comp_3 ~ create_total, family="binomial",data=d))
## 
## Call:
## glm(formula = comp_3 ~ create_total, family = "binomial", data = d)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -0.9084  -0.7787  -0.7457   1.5177   1.7723  
## 
## Coefficients:
##              Estimate Std. Error z value Pr(>|z|)  
## (Intercept)  -0.47225    0.37021  -1.276   0.2021  
## create_total -0.13308    0.08038  -1.656   0.0978 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 855.61  on 753  degrees of freedom
## Residual deviance: 852.85  on 752  degrees of freedom
##   (1 observation deleted due to missingness)
## AIC: 856.85
## 
## Number of Fisher Scoring iterations: 4

Conclusion: This is promising! It looks like, for at least the self-reported creativity, compromising is reduced for more creative individuals.

Things to discuss:

  1. What is our self-reported creativity measure capturing? In the previous study, it did not relate to the RAT. In this study, while it does relate to number of pasta names, it doesn’t relate to percent ending in “i.” Also, the number of pasta names did not relate to compromise. It might be that we are capturing a self-perception and this isn’t a cognitive story. I think that’s fine and still interesting, but we should discuss what that mechanism is (one possibility is that people who feel creative have less need to justify or feel less constrained by situations. Could be related to conformity, as well. I do think it would be a big contribution to the creativity lit to separate the self-perception of creativity vs. actual creative thinking). Alternatively, our measures of creativity might be crappy. Coding the actual pasta names for creativitiy might be a good step. Either way, let’s talk mechansim.

  2. The number of reasons listed in the disjunctive study was not significant this time (with a much larger sample). However, we changed the order of when we measured how much they like Hawaii vacations. That might have primed that reason, making other reaosns less salient. This is supported by the fact that the number of ideas listed was lower in this study than our previous study.

  3. It was interesting that the percent of names ending in i did negatively predict number of reasons listed. We should discuss if this is something we should follow up on.

Possible next steps:

  1. I think it might be fruitful to code the pasta names– this could help with deciding mechanism. This is especially true because a lot of people didn’t list any names ending in i. this is suprising to me and not what I normally find. Coding the ideas might provide a more nuanced perspective.

  2. Re-measure reasons listed and have it after the compromise question, potentially.

  3. The next study can begin to test mechanism. I’d prefer to use self-reported creativity as our IV one more time. Once we have more idea about mechanism it will be easier to determine how we should manipulate our creativity IV (is it about perception or the cognitive process of divergent thinking?)