Transcript Mike`s Seminar

ANOVA
Seminar #30
(Economic Value of Tropical Forest to Coffee Production)
Mike DeDad
12/3/07
Source: http://www.pnas.org/cgi/reprint/0405147101v1.pdf
Economic Value of Tropical Forest to
Coffee Production
Societal benefits are immense
 Not a big enough motivator
Economic value
 pollination services to agriculture
Purpose
 To “conduct pollination experiments to
examine the effects of these patterns on
coffee production and quality, and to estimate
the resulting economic value of tropical forest
fragments to coffee farms.”
 Previously determined that “bee species richness
and visitation rate decline with distance from
forest.”
Methods
Studied a 1,065 hectare coffee farm in
the Valle General, Costa Rica
Took branches from three types of sites
 Near, Intermediate, and Far
Took 4 branches from each of 5 healthy
plants from each site divided randomly
between hand-pollination and ambient
pollination
Methods
Explanatory Variables
 Distance (Near, Intermediate, or Far)
 Pollination (By hand or ambient)
Response Variables
 seed mass
 Fruit set
 peaberry frequency
Methods
What are some possible confounding
variables?
 Length of the branch
 number of leaves
 relative shade
Dismissed these because r<.19
 Not significantly related
What are the hypotheses?
Null Hypothesis: population means of
each response variable are equal for
each explanatory variable
Alternative Hypothesis: population
means differ
Two-way ANOVA
DFG= product of the two degrees of
freedom for each factor
 Distance dfg= (I-1)= 3-1= 2
 Pollination dfg= (I-1)= 2-1= 1
 Two-way ANOVA dfg= 2 x 1= 2
F2,24 = 8.13, P < 0.0001
Reject H0
F2,24 = 2.96, P = 0.0710
Borderline
F2,24 = 7.28, P = 0.0034
Reject H0
Conclusions
Hand pollinating increased seed mass
and fruit set for far sites while
decreasing peaberry frequency (higher
quality)
Ambient pollination decreased seed
mass and fruit set for far sites while
increasing peaberry frequency (lower
quality)
Conclusions
The presence of native tropical forest
increased both the quantity and quality
of harvested coffee (near and int. sites)
The lack of nearby forest decreases
quantity and quality of harvested coffee
(far sites)
What are the consequences of
committing error?
Type 1: Rejecting the null hypothesis
even though it is true
 Convincing farmers to not expand even
though the presence of forests has no
effect on coffee production
Type 2: Failing to reject the null
hypothesis even though it is false
 Allowing deforestation even though forests
have an effect on coffee quality and yield
Are there any questions?