x= sample(x = -1000:1000, size = 500)
y= sample(x = -1000:1000, size = 500)
xpos = sample(x = 0:1000, size = 250)
ypos = sample(x = 0:1000, size = 250)
s = c()
t = c()
for (q in 1:1000) {
x= sample(x = -1000:1000, size = 500)
y= sample(x = -1000:1000, size = 500)
xpos = sample(x = 0:1000, size = 250)
ypos = sample(x = 0:1000, size = 250)
s[q] = c(summary(lm(x~y))$coefficients[1,4])
t[q] = c(summary(lm(xpos~ypos))$coefficients[1,4])
}
summary(s)
summary(t)
e= sample(x = 0:1000, size = 500)
f= sample(x = 0:1000, size = 500)
epos = sample(x = 0:500, size = 250)
fpos = sample(x = 0:500, size = 250)
k = c()
h = c()
for (q in 1:1000) {
e= sample(x = 0:1000, size = 250)
f= sample(x = 0:1000, size = 250)
epos = sample(x = 0:500, size = 125)
fpos = sample(x = 0:500, size = 125)
k[q] = c(summary(lm(e~f))$coefficients[1,4])
h[q] = c(summary(lm(epos~fpos))$coefficients[1,4])
}
summary(k)
summary(h)
Summary of findings: splitting the data into a half-sized category with a positive/negative split creates a nearly immeasurable difference where the normal data reveals no relationship, while the split data "reveals" a statistical significance strong enough to be considered a cosmic or particle physics discovery. With just the positive data, both come out as "significant" but the significance gets cut in half with the split data.
No comments:
Post a Comment