Cultural Tightness and Social Values

Introduction

The project attempts to investigate the impact of cultural tightness on social values of people in different societies. Cultural tightness, a new cultural dimension developed by Gelfand in 2006, refers to the strength of norms and degree of sanctioning within societies. Research has identified its influence on creativity, leadership, and personality, acknowledging its vital role on shaping people’s mind.

I will use the tightness score data from Gelfand’s research, which calculates the tightness score of 33 countries. The social values data will be collected from the Association of Religion Data Archives. In order to fulfill the objective of my project, I will use the data from the Association of Religion Data Archives to create new variables such as gender belief and environment belief. Then I will use regression to determine whether cultural tightness will influence these variables. Data filtering, regression, and visualization will be employed in my project.

A better grasp of cultural variation across societies will inform conversations and improve communication. My project might help us understand why values and beliefs of people from different cultural backgrounds differ and help us develop common ground.

Packages Required

Following are the packages used in this project:

library(tidyverse)    #creating tidy data
library(readxl)       #reading data in excel
library(DT)           #creating interactive tables

Data Preparation

Data Import

  • Cultural Tightness Score: The data is collect from Gelfand’s research, which calculates the tightness score of 33 countries.
TS <- read_excel("/Users/shienhao/Desktop/R/working directory/Tightness score.XLSX")
WVS <- read_excel("/Users/shienhao/Desktop/R/working directory/World Values Survey Aggregate Data.XLSX", sheet = "Data")

Data Transformation

In order to examine the relationship between cultural tightness score and social values, we first need to merge the two datasets by country names.

mergeddata <- merge(WVS, TS, by= "CNAME")

The World Values Survey Aggregate Data consist research result of wave 1-4. Compared with wave 1,2, and 4, wave 3 contains the most comprehensive result. Therefore we will the wave 3 data to conduct analysis. First we need to filter the data to get the result of wave 3.

WVS3 <-select(mergeddata, ends_with("3"))

Then we create varibles that are of interest to this project. They include:environment belief, gender belief, passiveness, and religious devotion.

envir_belief <- select(WVS3, starts_with("ENV"))%>%
  rowMeans(na.rm=TRUE)
gender_belief <- select(WVS3, "TOLERA_3", "JOBS1_3",
                      "MENPOL_3", "BOYSCH_3")%>%
  rowMeans(na.rm=TRUE)
passiveness <-select(WVS3, "POL1_3", "POL2_3",
                    "POL3_3", "POL4_3")%>%
  rowMeans(na.rm=TRUE)
religious_devotion <- select(WVS3, "IMPREL_3", "EVIL_3",
                           "ATTEND_3","GODIMP_3")%>%
  rowMeans(na.rm=TRUE)

Adding these variables to the merged dataset, we have the final dataset now.

Final <- cbind(WVS3, envir_belief=envir_belief, gender_belief=gender_belief,passiveness=passiveness)

The final dataset is displayed below.

datatable(Final, caption = "Table 1: Cleaned Dataset")

Summary of Variables

Although there are 139 variables in the “Final” dataset, only a few are of concern to the project. Following is the summary of these variables.

Tightness_score: A higher scores means more emphasis for environment.

gender_belief: A higher score means lower gender equality.

passiveness: A higher score means greater passive views toward politics.

religious_devotion: A higher score means stronger devotion to religion

Exploratory Data Analysis

We first visualize the tightness score for 33 countries. As is depicted in Figure 1, the tightness score varies dramatically across 33 countries. Pakistan has the highest tightness score, which is 12.3, while the United Kingdom has the lowest tightness score, which is 1.6.

ggplot(data = TS, aes(x = CNAME, y = Tightness_score))+
  geom_col()+
  theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust = 1))+
  ggtitle("Figure 1:Tightness Score of 33 Countries")

Then we calculate the correlation between tightness score and the 4 variables and create scatterplots to visualize their relationship. As is shown below, gender belief and religious devetion have relatively strong correlations with tightness score while environment belief and passiveness have weak correlations with tightness score.

cor(envir_belief, Tightness_score, use = "complete.obs")
## [1] -0.02124435
ggplot(Final, aes(x = Tightness_score, y = envir_belief))+
  geom_point()+
  geom_smooth(method = "lm")+
  ggtitle("Figure2: tightness score vs environment belief")

cor(gender_belief, Tightness_score,use = "complete.obs")
##[1] 0.3189363
ggplot(Final, aes(x = Tightness_score, y = gender_belief))+
  geom_point()+
  geom_smooth(method = "lm")+
  ggtitle("Figure3: tightness score vs gender belief")

cor(passiveness, Tightness_score, use = "complete.obs")
##[1] 0.01750742
ggplot(Final, aes(x=Tightness_score, y=passiveness))+
  geom_point()+
  geom_smooth(method = "lm")+
  ggtitle("Figure4: tightness score vs passiveness")

cor(religious_devotion, Tightness_score, use = "complete.obs")
##[1] 0.2243921
ggplot(Final, aes(x=Tightness_score, y=religious_devotion))+
  geom_point()+
  geom_smooth(method = "lm")+
  ggtitle("Figure5: tightness score vs religious devotion")

Based on the correlation result, we decide to develop a linear regression model for tightness score and gender belief and for tightness score and religious devotion. The regression model has an acceptable R square of 0.1017, but the p value for tightness score is 0.247, which is not statistically significant. Therefore, tightness score cannot be considered a useful predictor for gender belief.

model1 <- lm(gender_belief~Tightness_score, data=Final)
summary(model1)
## Call:
## lm(formula = gender_belief ~ Tightness_score, data = Final)

## Residuals:
##      Min       1Q   Median       3Q      Max 
## -27.3269  -8.6942   0.2755   9.3843  20.6379 

## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)   
## (Intercept)       26.930      8.609   3.128    0.008 **
## Tightness_score    1.487      1.226   1.213    0.247   
---
## Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

## Residual standard error: 14.21 on 13 degrees of freedom
##   (10 observations deleted due to missingness)
## Multiple R-squared:  0.1017, Adjusted R-squared:  0.03262 
## F-statistic: 1.472 on 1 and 13 DF,  p-value: 0.2466

The R square of the regression model for religious devotion and tightness score is 0.05035, and the p value for tightness score is 0.4214, which is not statistically significant. Therefore, tightness score cannot be considered a useful predictor for religious devotion.

model2 <- lm(religious_devotion~Tightness_score, data=Final)
summary(model2)
## Call:
## lm(formula = religious_devotion ~ Tightness_score, data = Final)

## Residuals:
##     Min      1Q  Median      3Q     Max 
## -33.897 -15.501   0.418  15.409  37.181 

## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)  
## (Intercept)       28.197     13.254   2.128   0.0531 .
## Tightness_score    1.567      1.887   0.830   0.4214  
---
## Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

## Residual standard error: 21.87 on 13 degrees of freedom
##   (10 observations deleted due to missingness)
## Multiple R-squared:  0.05035,    Adjusted R-squared:  -0.0227 
## F-statistic: 0.6893 on 1 and 13 DF,  p-value: 0.4214

Summary

Our analysis identifies the relationship between tightness score and some social values. According to our analysis, tightness score has weak correlations with environment belief and passiveness, but has relatively strong positive correlations with gender belief and religious devotion. The results suggest that environment belief and passiveness do not differ much between loose and tight societies but gender belief and religious devotion tend to have wide variation between loose and tight societies.

The linear regression model we developed sheds light on whether we can use tightness score to predict certain social values. Although the regression model for tightness score and gender belief has an acceptable R square, the p value for tightness score is concerning. Therefore tightness score might not be considered useful predictor of certain social values. This might result from the limited size of the tightness score dataset.