1.Introduction

In psychology, 5 main personality traits have been defined since the 1980s, commonly coined as the “Big Five”: openness, conscientiousness, extraversion, agreeableness and neuroticism - each existing on a unique spectrum within every individual. These personality traits are linked to different patterns of behavior and personality.

Big 5 Personality Traits
OCEAN
Openness
Conscientiousness
Extraversion
Agreeableness
Neuroticism

Openness to experience is characterized by a curiosity and willingness to explore novel ideas and sensations. Conscientiousness reflects a tendency towards organization, diligence, and goal-directed behavior. Extraversion relates to sociability, assertiveness, and positive emotionality. Agreeableness denotes empathy, compassion, and cooperation, while neuroticism pertains to emotional instability and susceptibility to negative affect.

Interestingly, research has revealed correlations between certain personality traits and substance use. Specifically, high levels of neuroticism, combined with high openness to experience and low agreeableness, have been linked to the use of various drugs [https://pubmed.ncbi.nlm.nih.gov/34766786/].

As a result, the aim of this project is to look specifically at the associations between openness to experience (O) and illegal drug use, employing weed as a proxy for drug consumption patterns. The research question guiding this investigation is:

What is the association between openness to experience (O) and the use of both illegal and legal drugs, specifically focusing on cannabis consumption as a proxy for drug consumption patterns?

Data Set Information:

Database contains records for 1885 respondents. For each respondent 12 attributes are known:

Attribute	Description
1-5	Personality measurements (NEO-FFI-R): Neuroticism, Extraversion, Openness to Experience, Agreeableness, Conscientiousness
6	BIS-11: Impulsivity
7	ImpSS: Sensation seeking
8	Level of education
9	Age
10	Gender
11	Country of residence
12	Ethnicity

All input attributes are originally categorical and are quantified. After quantification values of all input features can be considered as real-valued. In addition, participants were questioned concerning their use of 18 legal and illegal drugs (alcohol, amphetamines, amyl nitrite, benzodiazepine, cannabis, chocolate, cocaine, caffeine, crack, ecstasy, heroin, ketamine, legal highs, LSD, methadone, mushrooms, nicotine and volatile substance abuse and one fictitious drug (Semeron) which was introduced to identify over-claimers. For each drug they have to select one of the answers: never used the drug, used it over a decade ago, or in the last decade, year, month, week, or day.

Detailed description of database and process of data quantification are presented in E. Fehrman, A. K. Muhammad, E. M. Mirkes, V. Egan and A. N. Gorban, “The Five Factor Model of personality and evaluation of drug consumption risk.,” arXiv [https://arxiv.org/pdf/1506.06297.pdf], 2015 The paper above solve binary classification problem for all drugs. For most of drugs sensitivity and specificity are greater than 75%

Dimensions: n = 1884, 32 variables

Data Dictionary
Variable	Description
ID	Numerical identifier
Age	Participant’s age
Gender	Participant’s gender
Education	Level of completed education
Country	Country of origin
Ethnicity	Participant’s ethnicity
Nscore	Neuroticism score (NEO-FFI-R)
Escore	Extraversion score (NEO-FFI-R)
Ascore	Agreeableness score (NEO-FFI-R)
Oscore	Openness to Experience score (NEO-FFI-R)
Cscore	Conscientiousness score (NEO-FFI-R)
Impulsive	Impulsivity score (BIS-11)
SS	Sensation seeking score (ImpSS)
Alcohol	Alcohol use
Amphet	Amphetamines use
Amyl	Amyl nitrite use
Benzos	Benzodiazepine use
Caff	Caffeine use
Cannabis	Cannabis use
Choc	Chocolate use
Coke	Cocaine use
Crack	Crack cocaine use
Ecstasy	Ecstasy use
Heroin	Heroin use
Ketamine	Ketamine use
Legalh	Legal highs use (like synthetic cannabinoids & other research chemicals)
Meth	Methamphetamine use
Mushrooms	‘Magic’ mushrooms use
Nicotine	Nicotine use
Semer	Fictitious drug ‘Semeron’ to assess overreporting
VSA	Volatile substance abuse (i.e. nitrous oxide)

Personality scores are given in z-scores as compared to the rest of the group.

All drugs used are assessed on a scale of 0-6 with the following classes: “Never Used”, “Used over a Decade Ago”, “Used in Last Decade”, “Used in Last Year”, “Used in Last Month”, “Used in Last Week”, and “Used in Last Day”.

# 1. load the appropriate libraries
library(tidyverse)
library(broom)
library(GGally)
library(car)
library(easystats)
library(datawizard)
library(marginaleffects)
library(modelsummary)
library(table1)

# 2. use the read_rds file to read the dataset
drug_consumption <- read_csv("data/Drug_Consumption.csv")

drug_consumption <- drug_consumption %>%
  rename(Ascore = AScore)

2. Exploratory data analysys (EDA)

ggplot(drug_consumption) +
  geom_bar(aes(x = Education, fill = Education)) +
  labs(y = "Drug Consumption", title = "Drug Consumption by Education Level") +
  theme(axis.text.x = element_blank()) +
  theme(legend.title = element_blank(),
        axis.ticks.x = element_blank(),
        legend.text = element_text(size = 8))

Most of the population seems to have had at least some sort of university education and above.

ggplot(drug_consumption) +
  geom_bar(aes(x=Age, fill = Cannabis))+
    labs(y = "Cannabis Consumption", title = "Cannabis Consumption by Age")

It seems that from this bar graph, the number of people using cannabis decreases with age overall. People aged 18-24 have the highest use of cannabis, with most of them using cannabis within the last day (CL6).

ggplot(drug_consumption) +
  geom_bar(aes(x=Gender, fill = Gender))+
   labs(y = "Drug Consumption", title = "Drug Consumption by Gender")+
   theme(axis.text.x = element_blank(),  
        axis.ticks.x = element_blank())

ggplot(drug_consumption) +
  geom_bar(aes(x=Country, fill = Country)) +
  labs(y = "Drug Consumption", title = "Drug Consumption by Country") +
  theme(axis.text.x = element_blank(),  
        axis.ticks.x = element_blank())

ggplot(drug_consumption) +
  geom_bar(aes(x=Ethnicity, fill = Ethnicity)) +
   labs(y = "Drug Consumption", title = "Drug Consumption by Ethnicity") +
  theme(axis.text.x = element_blank(),  
        axis.ticks.x = element_blank())

The bar graph drug consumption by gender showcases little to none differences for drug use. As most data comes from US & UK, there is little value in the graph drug consumption by country. The majority of the population are white, as most of them are from the UK and USA.

Table 1

table1(~Age + Education + Country + Ethnicity + Nscore + Escore + Oscore + Ascore + Cscore + Impulsive + SS | Gender, data=drug_consumption)

	F (N=941)	M (N=943)	Overall (N=1884)
Age
18-24	237 (25.2%)	406 (43.1%)	643 (34.1%)
25-34	294 (31.2%)	187 (19.8%)	481 (25.5%)
35-44	201 (21.4%)	154 (16.3%)	355 (18.8%)
45-54	158 (16.8%)	136 (14.4%)	294 (15.6%)
55-64	43 (4.6%)	50 (5.3%)	93 (4.9%)
65+	8 (0.9%)	10 (1.1%)	18 (1.0%)
Education
Doctorate degree	57 (6.1%)	32 (3.4%)	89 (4.7%)
Left school at 16 years	43 (4.6%)	56 (5.9%)	99 (5.3%)
Left school at 17 years	14 (1.5%)	16 (1.7%)	30 (1.6%)
Left school at 18 years	37 (3.9%)	63 (6.7%)	100 (5.3%)
Left school before 16 years	12 (1.3%)	16 (1.7%)	28 (1.5%)
Masters degree	179 (19.0%)	104 (11.0%)	283 (15.0%)
Professional certificate/ diploma	135 (14.3%)	134 (14.2%)	269 (14.3%)
Some college or university, no certificate or degree	171 (18.2%)	335 (35.5%)	506 (26.9%)
University degree	293 (31.1%)	187 (19.8%)	480 (25.5%)
Country
Australia	21 (2.2%)	33 (3.5%)	54 (2.9%)
Canada	41 (4.4%)	46 (4.9%)	87 (4.6%)
New Zealand	1 (0.1%)	4 (0.4%)	5 (0.3%)
Other	36 (3.8%)	82 (8.7%)	118 (6.3%)
Republic of Ireland	9 (1.0%)	11 (1.2%)	20 (1.1%)
UK	622 (66.1%)	421 (44.6%)	1043 (55.4%)
USA	211 (22.4%)	346 (36.7%)	557 (29.6%)
Ethnicity
Asian	15 (1.6%)	11 (1.2%)	26 (1.4%)
Black	17 (1.8%)	16 (1.7%)	33 (1.8%)
Mixed-Black/Asian	2 (0.2%)	1 (0.1%)	3 (0.2%)
Mixed-White/Asian	10 (1.1%)	9 (1.0%)	19 (1.0%)
Mixed-White/Black	11 (1.2%)	9 (1.0%)	20 (1.1%)
Other	29 (3.1%)	34 (3.6%)	63 (3.3%)
White	857 (91.1%)	863 (91.5%)	1720 (91.3%)
Nscore
Mean (SD)	0.0743 (0.983)	-0.0744 (1.01)	-0.000119 (0.998)
Median [Min, Max]	0.0426 [-2.76, 3.27]	-0.0519 [-3.46, 2.61]	0.0426 [-3.46, 3.27]
Escore
Mean (SD)	0.0582 (0.953)	-0.0578 (1.04)	0.000143 (0.998)
Median [Min, Max]	0.00332 [-2.73, 3.01]	-0.155 [-3.27, 3.27]	0.00332 [-3.27, 3.27]
Oscore
Mean (SD)	-0.131 (1.02)	0.130 (0.953)	-0.000225 (0.996)
Median [Min, Max]	-0.178 [-3.27, 2.90]	0.141 [-2.86, 2.90]	-0.0193 [-3.27, 2.90]
Ascore
Mean (SD)	0.220 (0.976)	-0.219 (0.970)	0.000242 (0.997)
Median [Min, Max]	0.288 [-2.79, 3.16]	-0.302 [-3.46, 3.46]	-0.0173 [-3.46, 3.46]
Cscore
Mean (SD)	0.183 (0.986)	-0.184 (0.976)	-0.000383 (0.998)
Median [Min, Max]	0.260 [-3.16, 3.46]	-0.276 [-3.46, 3.01]	-0.00665 [-3.46, 3.46]
Impulsive
Mean (SD)	-0.153 (0.974)	0.167 (0.908)	0.00734 (0.955)
Median [Min, Max]	-0.217 [-2.56, 2.90]	0.193 [-2.56, 2.90]	-0.217 [-2.56, 2.90]
SS
Mean (SD)	-0.238 (0.985)	0.232 (0.882)	-0.00267 (0.964)
Median [Min, Max]	-0.216 [-2.08, 1.92]	0.401 [-2.08, 1.92]	0.0799 [-2.08, 1.92]

Table 1 shows the general characteristics of the population used in this analysis. Interestingly, there is a difference in the openness score between males and females. In the dataset, females have a lower openness score than males.

To carry on further with the data analysis, we reclassified the columns of drug use into dummy variables, where we define CL0 and 1 (CL0: Never used, CL1: used over a decade ago) as someone not using the drug. For example, a person with a score of CL1 for cannabis would be classified as not using the drug, whereas any score higher than that would classify them for using it.

This table shows the descriptive statistics of people who use illegal drugs (using both weed and ecstasy).

#table 1 & ggpairs split by drug use

#create variables of use

drug_consumption <- drug_consumption %>%
  mutate(drinks =
        case_when(
        Alcohol == "CL0" ~ F,
        Alcohol == "CL1" ~ F,
        T ~ T
        ) )

drug_consumption <- drug_consumption %>%
  mutate(weed =
        case_when(
        Cannabis == "CL0" ~ F,
        Cannabis == "CL1" ~ F,
        T ~ T
        ) )


drug_consumption <- drug_consumption %>%
  mutate(xtc =
        case_when(
        Ecstasy == "CL0" ~ F,
        Ecstasy == "CL1" ~ F,
        T ~ T
        ) )


drug_consumption <- drug_consumption %>%
  mutate(smokes =
        case_when(
        Nicotine == "CL0" ~ F,
        Nicotine == "CL1" ~ F,
        T ~ T
        ) )

drug_consumption <- drug_consumption %>%
  mutate(legalDrugs =
        case_when(
        smokes == F & drinks == F ~ F,
        T ~ T
        ) )

drug_consumption <- drug_consumption %>%
  mutate(illegalDrugs =
        case_when(
        weed == F & xtc == F ~ "Doesnt use",
        T ~ "Uses"
        ) )

#table 

table1(~Age + Education + drinks + smokes + xtc + weed | illegalDrugs, data=drug_consumption)

	Doesnt use (N=598)	Uses (N=1286)	Overall (N=1884)
Age
18-24	57 (9.5%)	586 (45.6%)	643 (34.1%)
25-34	130 (21.7%)	351 (27.3%)	481 (25.5%)
35-44	158 (26.4%)	197 (15.3%)	355 (18.8%)
45-54	180 (30.1%)	114 (8.9%)	294 (15.6%)
55-64	56 (9.4%)	37 (2.9%)	93 (4.9%)
65+	17 (2.8%)	1 (0.1%)	18 (1.0%)
Education
Doctorate degree	39 (6.5%)	50 (3.9%)	89 (4.7%)
Left school at 16 years	44 (7.4%)	55 (4.3%)	99 (5.3%)
Left school at 17 years	7 (1.2%)	23 (1.8%)	30 (1.6%)
Left school at 18 years	22 (3.7%)	78 (6.1%)	100 (5.3%)
Left school before 16 years	8 (1.3%)	20 (1.6%)	28 (1.5%)
Masters degree	139 (23.2%)	144 (11.2%)	283 (15.0%)
Professional certificate/ diploma	112 (18.7%)	157 (12.2%)	269 (14.3%)
Some college or university, no certificate or degree	47 (7.9%)	459 (35.7%)	506 (26.9%)
University degree	180 (30.1%)	300 (23.3%)	480 (25.5%)
drinks
Yes	556 (93.0%)	1260 (98.0%)	1816 (96.4%)
No	42 (7.0%)	26 (2.0%)	68 (3.6%)
smokes
Yes	183 (30.6%)	1080 (84.0%)	1263 (67.0%)
No	415 (69.4%)	206 (16.0%)	621 (33.0%)
xtc
Yes	0 (0%)	751 (58.4%)	751 (39.9%)
No	598 (100%)	535 (41.6%)	1133 (60.1%)
weed
Yes	0 (0%)	1265 (98.4%)	1265 (67.1%)
No	598 (100%)	21 (1.6%)	619 (32.9%)

From this table, illegal drug use seems to be higher in people with a higher education, but this could be due to the larger sample size for people in higher education. In terms of drinking and smoking, the majority of illegal drug users do both, whereas if they don’t use illegal drugs, there is a larger chance of them not smoking or drinking.

#ggpairs

pm <- ggpairs(drug_consumption,
            aes(color = sober),
            columns = c("Nscore", "Escore", "Oscore", "Ascore", "Cscore","Impulsive", "SS"),
            columnLabels = c("Neuroticism ", "Extraversion", "Openness", "Agreeableness", "Conscientiousness", "Impulsiveness", "Sensation Seeking") ,  
  lower = list(
    continuous = wrap("cor", size = 2.0),
    combo = "facethist",
    mapping = aes(color = illegalDrugs)
  ),
  upper = list(
    continuous = "density",
    combo = "facethist",
    discrete = "colbar",
    mapping = aes(color = illegalDrugs)
  ),
  diag = list(
    continuous = wrap("densityDiag", alpha = 0.5, color = FALSE),
    discrete = "densityDiag",
    mapping = aes(color = illegalDrugs)), labels = list(textSize = 4))

pm + theme(strip.text.x = element_text(size = 6.5),
           strip.text.y = element_text(size = 4))

In this plot, blue shows the illegal drug users and red is the legal drug users. It can be seen that out of the big five personality traits, the trend between openness and illegal drug use is the most clear. People who have higher openness scores tend to have more use of illegal drugs. The other characteristics that are also seen to be linked with illegal drug use that are not in the big five personality traits are impulsiveness and sensation seeking.

custom_labels <- c("Never Used", "Used over a Decade Ago", "Used in Last Decade", 
                   "Used in Last Year", "Used in Last Month", "Used in Last Week", 
                   "Used in Last Day")

ggplot(drug_consumption, aes(x = Cannabis, fill = Cannabis))+
  geom_bar() +
   labs(y = "Frequency", x= 'Cannabis use', title = "Frequency of Cannabis Consumption")+
  scale_fill_discrete(labels = custom_labels)

drug_consumption <- drug_consumption %>%
  mutate(weedYear =
        case_when(
        Cannabis == "CL0" ~ F,
        Cannabis == "CL1" ~ F,
        Cannabis == "CL2" ~ F,
        T ~ T
        ) )

custom_labels <- c("Never Used", "Used over a Decade Ago", "Used in Last Decade", 
                   "Used in Last Year", "Used in Last Month", "Used in Last Week", 
                   "Used in Last Day")

ggplot(drug_consumption, aes(x=Alcohol, y=Oscore, fill=Alcohol)) +
  geom_boxplot()+
  scale_fill_discrete(labels = custom_labels)

3. Regression analysys

weedOpen <- glm(weed~Oscore, family="binomial", data=drug_consumption)
summary(weedOpen)

## 
## Call:
## glm(formula = weed ~ Oscore, family = "binomial", data = drug_consumption)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)    
## (Intercept)  0.82450    0.05428   15.19   <2e-16 ***
## Oscore       0.85422    0.05992   14.26   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 2385.7  on 1883  degrees of freedom
## Residual deviance: 2139.6  on 1882  degrees of freedom
## AIC: 2143.6
## 
## Number of Fisher Scoring iterations: 4

The model output indicates that both the intercept and the coefficient for Oscore are statistically significant, with p-values < 0.001. This suggests that there is a significant relationship between openness to experience and cannabis use.

Specifically, the estimated coefficient for Oscore is 0.85422, indicating that for every one-unit increase in openness to experience score, the log-odds of cannabis use increase by approximately 0.85422, holding other variables constant.

The null deviance and residual deviance are measures of model fit. The residual deviance of 2139.6 on 1882 degrees of freedom indicates that the model explains a significant amount of variability in the data.

weedBig5 <- glm(weed~Oscore+Escore+Ascore+Cscore+Nscore, family="binomial", data=drug_consumption)
summary(weedBig5)

## 
## Call:
## glm(formula = weed ~ Oscore + Escore + Ascore + Cscore + Nscore, 
##     family = "binomial", data = drug_consumption)
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)    
## (Intercept)  0.91986    0.05881  15.640  < 2e-16 ***
## Oscore       0.94140    0.06585  14.295  < 2e-16 ***
## Escore      -0.09746    0.06717  -1.451    0.147    
## Ascore      -0.32706    0.05993  -5.458 4.83e-08 ***
## Cscore      -0.56202    0.06579  -8.543  < 2e-16 ***
## Nscore       0.04354    0.06712   0.649    0.517    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 2385.7  on 1883  degrees of freedom
## Residual deviance: 1958.3  on 1878  degrees of freedom
## AIC: 1970.3
## 
## Number of Fisher Scoring iterations: 4

The model output indicates that the intercept and the coefficients for Oscore, Ascore, and Cscore are statistically significant (p < 0.001), suggesting that these variables have a significant relationship with cannabis use. However, the coefficients for Escore (extraversion) and Nscore (neuroticism) are not statistically significant at the 0.05 level, suggesting that these variables may not have a significant association with cannabis use.

drugOpen <- lm(Oscore~weed+drinks+smokes+xtc, data=drug_consumption)
summary(drugOpen)

## 
## Call:
## lm(formula = Oscore ~ weed + drinks + smokes + xtc, data = drug_consumption)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.2349 -0.6186  0.0487  0.6568  2.6817 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) -0.37218    0.11411  -3.262  0.00113 ** 
## weedTRUE     0.56691    0.05893   9.620  < 2e-16 ***
## drinksTRUE  -0.15628    0.11537  -1.355  0.17569    
## smokesTRUE   0.02805    0.05412   0.518  0.60427    
## xtcTRUE      0.30892    0.05146   6.003 2.32e-09 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.9249 on 1879 degrees of freedom
## Multiple R-squared:  0.1402, Adjusted R-squared:  0.1384 
## F-statistic:  76.6 on 4 and 1879 DF,  p-value: < 2.2e-16

This regression analysis explores how openness to experience (Oscore) is influenced by various drug use behaviors, including weed, alcohol, cigarettes, and ecstasy. The results indicate that individuals who use weed or ecstasy tend to have higher Oscores compared to non-users, while the use of alcohol and cigarettes does not show a significant association with Oscore. Overall, the model suggests that drug use behaviors, particularly involving weed and ecstasy, may be linked to higher levels of openness to experience.

	Doesnt use (N=549)	Uses (N=1123)	Weed (N=212)	Overall (N=1884)
Age
18-24	49 (8.9%)	536 (47.7%)	58 (27.4%)	643 (34.1%)
25-34	118 (21.5%)	304 (27.1%)	59 (27.8%)	481 (25.5%)
35-44	144 (26.2%)	164 (14.6%)	47 (22.2%)	355 (18.8%)
45-54	169 (30.8%)	91 (8.1%)	34 (16.0%)	294 (15.6%)
55-64	53 (9.7%)	27 (2.4%)	13 (6.1%)	93 (4.9%)
65+	16 (2.9%)	1 (0.1%)	1 (0.5%)	18 (1.0%)
Education
Doctorate degree	35 (6.4%)	41 (3.7%)	13 (6.1%)	89 (4.7%)
Left school at 16 years	38 (6.9%)	49 (4.4%)	12 (5.7%)	99 (5.3%)
Left school at 17 years	7 (1.3%)	21 (1.9%)	2 (0.9%)	30 (1.6%)
Left school at 18 years	20 (3.6%)	74 (6.6%)	6 (2.8%)	100 (5.3%)
Left school before 16 years	8 (1.5%)	18 (1.6%)	2 (0.9%)	28 (1.5%)
Masters degree	133 (24.2%)	110 (9.8%)	40 (18.9%)	283 (15.0%)
Professional certificate/ diploma	105 (19.1%)	131 (11.7%)	33 (15.6%)	269 (14.3%)
Some college or university, no certificate or degree	38 (6.9%)	422 (37.6%)	46 (21.7%)	506 (26.9%)
University degree	165 (30.1%)	257 (22.9%)	58 (27.4%)	480 (25.5%)
drinks
Yes	510 (92.9%)	1098 (97.8%)	208 (98.1%)	1816 (96.4%)
No	39 (7.1%)	25 (2.2%)	4 (1.9%)	68 (3.6%)
smokes
Yes	163 (29.7%)	948 (84.4%)	152 (71.7%)	1263 (67.0%)
No	386 (70.3%)	175 (15.6%)	60 (28.3%)	621 (33.0%)

drugsOpen <- lm(Oscore~weed+drinks+smokes+xtc+xanax+paddos+specialK+acid+crystal+speed+rocks+blow+smack+cocoa+coffee+poppers, data=drug_consumption)
summary(drugsOpen)

## 
## Call:
## lm(formula = Oscore ~ weed + drinks + smokes + xtc + xanax + 
##     paddos + specialK + acid + crystal + speed + rocks + blow + 
##     smack + cocoa + coffee + poppers, data = drug_consumption)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -3.13742 -0.59349 -0.01192  0.61699  2.68087 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  -0.735924   0.214311  -3.434 0.000608 ***
## weedTRUE      0.475591   0.060451   7.867 6.08e-15 ***
## drinksTRUE   -0.150545   0.114399  -1.316 0.188347    
## smokesTRUE    0.014935   0.053418   0.280 0.779826    
## xtcTRUE       0.135376   0.067987   1.991 0.046604 *  
## xanaxTRUE     0.123710   0.052650   2.350 0.018894 *  
## paddosTRUE    0.189392   0.064534   2.935 0.003379 ** 
## specialKTRUE  0.004958   0.066257   0.075 0.940354    
## acidTRUE      0.353373   0.068031   5.194 2.28e-07 ***
## crystalTRUE   0.097918   0.063704   1.537 0.124443    
## speedTRUE    -0.041989   0.061870  -0.679 0.497438    
## rocksTRUE    -0.149137   0.083272  -1.791 0.073462 .  
## blowTRUE     -0.121066   0.063558  -1.905 0.056958 .  
## smackTRUE     0.051409   0.085750   0.600 0.548900    
## cocoaTRUE     0.136916   0.156074   0.877 0.380464    
## coffeeTRUE    0.221927   0.154271   1.439 0.150444    
## poppersTRUE  -0.168597   0.060068  -2.807 0.005056 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.9023 on 1867 degrees of freedom
## Multiple R-squared:  0.187,  Adjusted R-squared:  0.1801 
## F-statistic: 26.84 on 16 and 1867 DF,  p-value: < 2.2e-16

This extended regression analysis investigates the relationship between openness to experience (Oscore) and a broader range of drug use, including various illegal and legal substances. The results reveal that while the use of weed, acid, and poppers is significantly associated with higher Oscores, other substances such as drinks, smokes, and specialK show no significant correlation.

drug_consumption <- drug_consumption %>%
  mutate(psychedelics =
        case_when(
        paddos ~ T ,
        acid ~ T ,
        xtc ~ T ,
        weed ~ T ,
        poppers ~ T,
        T  ~ F
        ) )

drugsBest <- lm(Oscore~weed+xtc+paddos+acid+poppers, data=drug_consumption)
summary(drugsBest)

## 
## Call:
## lm(formula = Oscore ~ weed + xtc + paddos + acid + poppers, data = drug_consumption)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -2.93184 -0.61044 -0.00021  0.61918  2.66182 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) -0.50858    0.03643 -13.961  < 2e-16 ***
## weedTRUE     0.48951    0.05432   9.011  < 2e-16 ***
## xtcTRUE      0.09141    0.06104   1.498 0.134380    
## paddosTRUE   0.19630    0.06400   3.067 0.002192 ** 
## acidTRUE     0.36919    0.06604   5.590  2.6e-08 ***
## poppersTRUE -0.19464    0.05752  -3.384 0.000729 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.9047 on 1878 degrees of freedom
## Multiple R-squared:  0.1777, Adjusted R-squared:  0.1755 
## F-statistic: 81.18 on 5 and 1878 DF,  p-value: < 2.2e-16

The regression analysis suggests that while weed usage positively correlates with openness to experience, other psychedelics such as xtc, paddos, and acid do not significantly influence openness. However, the use of poppers shows a negative association with openness. Overall, the model indicates that these factors explain about 17.55% of the variance in openness to experience.

4. Conclusion

In conclusion, our analysis suggests a discernible association between openness to experience and the consumption of illicit drugs, particularly cannabis, while such a relationship is not observed with legal drugs. However, the findings also highlight the nuanced interplay between personality traits and drug consumption behaviors, indicating that other factors such as agreeableness and conscientiousness also play significant roles.

Limitations:

The cross-sectional nature of the data limits the ability to establish causality. It’s unclear whether personality traits influence drug use or vice versa. The sample primarily consists of individuals from the UK and USA, potentially limiting the generalizability of the findings to other populations. The reliance on self-reported drug use and personality assessments may introduce bias and underreporting.

Lab 6 - Drug Consumption

Andreas, Anouar, Olga, Leo

2024-03-21