Trend analysis (5 of 5)
Next section: Formal model
The
degrees of freedom for the t is equal to
the degrees of freedom
error in the analysis of
variance
which is: N - a = 40 - 4 = 36. A
t table can
be used to find that the
probability value (p)
is less than .0001. Therefore, the decrease in running time associated
with the increase in magnitude of reward is significant. Rats run faster
if the reward is larger.
The quadratic component of trend is used to test
whether the decreasing slope that shows up as a flattening out of the
function is significant. The table on a
previous
page
shows that the coefficients for the quadratic component of trend with
four groups are: 1, -1, -1, 1.
L = ΣM
ia
i
= (10)(1) + (6)(-1) + (4)(-1) +(3)(1) = 3
s
L=
1.414
t = 3/1.414 = 2.12,
df = 36, and p = .041. The quadratic component of trend is
significant.
Next section: Formal model