Big Five Data

Author

Hugo Loyola

Introduction

This file is my version of the work presented by Allen B. Downey in his book Probably Overthingking It. The reason for replicate the text is apply R tools for the data analysis and learn to use Quarto also in R. The details on how the data is prepared can be founded in Allen Downey Github repository . I followed the steps presented there.

Data

The data can be downloaded from the Open-Source Psychometrics Project. The downloadable is described as “Answers to the IPIP Big Five Factor Markers”. The download is a .zip file containing the dataset (.csv file) and a .txt codebook.

The following is the procedure to download and unzip the file.

Code

url <- (
  "https://openpsychometrics.org/_rawdata/IPIP-FFM-data-8Nov2018.zip")
file_name <- "b5d.zip"
file_path <- ("/media/drive2/Quarto/")
download.file(url = url, destfile = paste0(file_path, file_name, sep = ""))
unzip("b5d.zip")

Dataset

The dataset will be set available to work with using the tidyverse tools.

Code

library(tidyverse)
library(gridExtra)
library(grid)
library(corrplot)

Code

dat1 <- read_tsv(
  "/media/drive2/Quarto/IPIP-FFM-data-8Nov2018/data-final.csv",col_select = c(1:50))

Rows: 1015341 Columns: 50
── Column specification ────────────────────────────────────────────────────────
Delimiter: "\t"
chr (50): EXT1, EXT2, EXT3, EXT4, EXT5, EXT6, EXT7, EXT8, EXT9, EXT10, EST1,...

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

To make it easier the work with this dataset we will limit ourselves to the answers of the test only and then not including additional information available in the downloaded file.

All variables in this subset are of character type. The first step will be to transform it to numeric variable type.

Code

dat1 <- dat1 %>%
  mutate(across(where(is.character), as.numeric))

Let start by looking at some of the data.

Code

apply(dat1[,1:10],2,table)

    EXT1   EXT2   EXT3   EXT4   EXT5   EXT6   EXT7   EXT8   EXT9  EXT10
0   3720   6099   6075   6893   9781   6183   8179   6372   6845   6261
1 249711 210104  76730 102032  98458 266568 236795  78165 173361  76504
2 199165 235528 191723 208184 178508 338198 227788 175759 221323 158282
3 288622 244683 269135 280161 234463 191408 194827 229962 214960 191349
4 191638 190216 284124 257148 299214 139539 197374 274953 242310 268237
5  80702 126928 185771 159140 193134  71662 148595 248347 154759 312925

The first thing to notice is the existence of 0 in the first 10 data columns and this value it is not expected as a valid response. The expected values ranges from 1 to 5.

We will take a look at missing values across variables if any.

Code

apply(is.na(dat1[,1:10]),2,sum)

 EXT1  EXT2  EXT3  EXT4  EXT5  EXT6  EXT7  EXT8  EXT9 EXT10 
 1783  1783  1783  1783  1783  1783  1783  1783  1783  1783

And yes, we have missing (NA) values in at least the first 10 variables of the dataset.

Before continue on any posterior analysis we need to remove the cases with a 0 and or NA.

First we will transform all 0 values to NA.

Code

#|label:  Change 0 to NA
dat1[dat1==0] <- NA

And second, remove all cases with NA

Code

dat1 <- dat1 %>% drop_na()

And let look the dataset again

Code

#label: check for changes
apply(dat1[,1:10],2,table)

    EXT1   EXT2   EXT3   EXT4   EXT5   EXT6   EXT7   EXT8   EXT9  EXT10
1 216949 179468  66519  86970  85967 229996 207648  66244 150045  64801
2 176207 205925 168457 181098 158114 297333 200884 152483 194494 136922
3 247835 212344 234223 242115 204662 165199 168070 198204 185618 164581
4 166599 167015 247551 226729 261179 121334 171441 241798 211865 235322
5  66844 109682 157684 137522 164512  60572 126391 215705 132412 272808

Code

dat1

# A tibble: 874,434 × 50
    EXT1  EXT2  EXT3  EXT4  EXT5  EXT6  EXT7  EXT8  EXT9 EXT10  EST1  EST2  EST3
   <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
 1     4     1     5     2     5     1     5     2     4     1     1     4     4
 2     3     5     3     4     3     3     2     5     1     5     2     3     4
 3     2     3     4     4     3     2     1     3     2     5     4     4     4
 4     2     2     2     3     4     2     2     4     1     4     3     3     3
 5     3     3     3     3     5     3     3     5     3     4     1     5     5
 6     3     3     4     2     4     2     2     3     3     4     3     4     3
 7     4     3     4     3     3     3     5     3     4     3     2     4     4
 8     3     1     5     2     5     2     5     2     3     2     2     4     2
 9     1     5     3     5     2     3     2     4     5     4     3     3     3
10     3     3     2     3     3     2     4     3     3     5     4     3     4
# ℹ 874,424 more rows
# ℹ 37 more variables: EST4 <dbl>, EST5 <dbl>, EST6 <dbl>, EST7 <dbl>,
#   EST8 <dbl>, EST9 <dbl>, EST10 <dbl>, AGR1 <dbl>, AGR2 <dbl>, AGR3 <dbl>,
#   AGR4 <dbl>, AGR5 <dbl>, AGR6 <dbl>, AGR7 <dbl>, AGR8 <dbl>, AGR9 <dbl>,
#   AGR10 <dbl>, CSN1 <dbl>, CSN2 <dbl>, CSN3 <dbl>, CSN4 <dbl>, CSN5 <dbl>,
#   CSN6 <dbl>, CSN7 <dbl>, CSN8 <dbl>, CSN9 <dbl>, CSN10 <dbl>, OPN1 <dbl>,
#   OPN2 <dbl>, OPN3 <dbl>, OPN4 <dbl>, OPN5 <dbl>, OPN6 <dbl>, OPN7 <dbl>, …

There was a reduction in the number of rows from 1015341 to 874,434, i.e., a reduction of 140907 rows (13.8%).

Personality traits.

In the documentation associated is said:

“The big five personality traits are the best accepted and most commonly used model of personality in academic psychology” and “This test uses the Big-Five Factor Markers from the International Personality Item Pool, developed by Goldberg (1992).”

In the text aforementioned is said that the test consist of fifty items that are rated on how true they are about you on a five point scale where 1=Disagree, 3=Neutral and 5=Agree.

The questions about people personality are: extraversion (EXT), neuroticism (EST), agreeableness (AGR), conscientiousness (CSN) and openness (OPN) to experience.

The personality traits score can be obtained by aggregating answers. As result of using a technique called factor analysis, the answers to the questions could be positive or negative keyed. Then each trait is developed by adding the keyed responses.

Code

#|label: Traits
EXT <- dat1[,1]-dat1[,2]+dat1[,3]-dat1[,4]+dat1[,5]-dat1[,6]+dat1[,7]-dat1[,8]+dat1[,9]-dat1[,10]
EST <- dat1[,11]-dat1[,12]+dat1[,13]-dat1[,14]+dat1[,15]-dat1[,16]+dat1[,17]-dat1[,18]+dat1[,19]-dat1[,20]
AGR <- dat1[,21]-dat1[,22]+dat1[,23]-dat1[,24]+dat1[,25]-dat1[,26]+dat1[,27]-dat1[,28]+dat1[,29]-dat1[,30]
CSN <- dat1[,31]-dat1[,32]+dat1[,33]-dat1[,34]+dat1[,35]-dat1[,36]+dat1[,37]-dat1[,38]+dat1[,39]-dat1[,40]
OPN <- dat1[,41]-dat1[,42]+dat1[,43]-dat1[,44]+dat1[,45]-dat1[,46]+dat1[,47]-dat1[,48]+dat1[,49]-dat1[,50]

Collect the traits score in a dataset

Code

#|label: Explore the score distributions for each personality trait.
bfd <- data.frame(EXT,EST,AGR,CSN,OPN)
bfd <- bfd %>%
  rename("EXT" = "EXT1", "EST"="EST1", "AGR" = "AGR1",
         "CSN"="CSN1","OPN"="OPN1")

Plot the traits

Code

#|label: Traits histograms
#|warning: false

ext <- ggplot(bfd,aes(x=EXT)) +
  geom_histogram(binwidth = 1)
est <- ggplot(bfd,aes(x=EST)) +
  geom_histogram(binwidth = 1)
agr <- ggplot(bfd,aes(x=AGR)) +
  geom_histogram(binwidth = 1)
csn <- ggplot(bfd,aes(x=CSN)) +
  geom_histogram(binwidth = 1)
opn <- ggplot(bfd,aes(x=OPN)) +
  geom_histogram(binwidth = 1)

grid.arrange(ext,est,agr,csn,opn, ncol=2)

Distribution frequency of scores show a clustering around some central value and dispersion of deviations with different level of symmetry.

The level of correlation between the traits can be seen as follows:

Code

cormat <- round(cor(bfd),2)
corrplot(cormat, type = "lower",order = "hclust",tl.col = "black",
         tl.srt = 45)