rm(list=ls())

#BSS8     (values: 0, 1, 2) (rare, frequent, continuous)
#   I accept the idea of killing myself.
#SLCS1   0
#Owing to my capabilities, I have much potential.
#SDT19   !
#I often do not feel very capable.   
#SLCS10  
#I am a capable person.  
#SDT15   B
#In my life I do not get much of a chance to show how capable I am.  

# Question
#2.	Is level of perceived burdensomeness related to level of suicidal ideation?
# Answer
# I will use correlation to establish any relation

#----------------------------------------------------------------------
# all numerical data, but this data links to textual answers
file = "VanOrden_for dist.csv"
dall = read.csv(file)
d = dall[c("bss8","slcs1","sdt19","slcs10","sdt15")]
d

# clean up unused variables
rm(dall)

# bss8 has many 999 undefines. Must get rid of them
d = d[d$bss8 != 999,]
d

v1 = d$slcs1
v2 = d$sdt19
v3 = d$slcs10
v4 = d$sdt15

# plot one variable versus the other
par(mfcol=c(2,2))
attach(d)
z = max(bss8)-min(bss8)
y = jitter(bss8,amount=z/100)
p = plot(v1,y)
p = plot(v2,y)
p = plot(v3,y)
p = plot(v4,y)
detach(d)

# There are 22 values, but only 10 points. Must change the type of 
# plot. Either add jitter, or a 3D plot, with count as the 3rd axis. 
# Or perform a correlation

# Conclusion from plot
# most results are that

# bss8 is mostly 0 (6 out of 9) across bel2 (only 7 values)
# 

print("*** Correlation of variable 1 with ideation ****")
attach(d)
cp = cor(bss8,v1, method="pearson")
ck = cor(bss8,v1, method="kendall")
cs = cor(bss8,v1, method="spearman")
detach(d)

# I need better way of printing
print("cp= "); print(cp)
print("ck= "); print(ck)
print("cs= "); print(cs)
# correlation of 0.4 : not much relation


print("*** Correlation of variable 2 with ideation ****")
attach(d)
cp = cor(bss8,v2, method="pearson")
ck = cor(bss8,v2, method="kendall")
cs = cor(bss8,v2, method="spearman")
detach(d)

# I need better way of printing
print("cp= "); print(cp)
print("ck= "); print(ck)
print("cs= "); print(cs)
# correlation of 0.4 : not much relation