Let’s move on to your challenge: You want to test the relationship between satisfaction with democracy and the variables trstprl
, trstlgl
, stfeco
, and agea
. Try using ggpairs()
!
Try it out before looking at the other tabs!
oneSolution <- ggpairs(
pss,
columns = c(
"stfdem",
"trstprl",
"trstlgl",
"stfeco"
),
lower = list(
continuous = wrap(
"points",
position = position_jitter(width = 0.5)
),
combo = wrap(
"facethist",
binwidth = 1
)
),
upper = list(
continuous = "cor",
combo = "box_no_facet"
),
diag = list(
continuous = wrap("densityDiag", bw = 1),
discrete = "barDiag"
)
)
oneSolution
library("psych")
pairs.panels(
pss[c(
"stfdem",
"trstprl",
"trstlgl",
"stfeco"
)
],
method = "pearson",
jiggle = TRUE, # für pseudometrische Daten
stars = TRUE # Konvention für Signifikanzen
)
cor <- corr.test(
pss[c(
"stfdem",
"trstprl",
"trstlgl",
"stfeco"
)
],
method = "pearson",
use = "complete.obs"
)
library("corrplot")
corrplot(
cor$r,
p.mat = cor$p, # Matrix mit p-Werten
insig = "blank", # nicht signifikante = leer
type = "upper", # auch lower möglich
method = "circle" # verschiedene Optionen möglich
)
On the next page, you can explore a ggplot
alternative to corrplot
, or you can skip ahead to comparing means!