Indoor Yield-O-Rama

Analysis of Medium Types

As with our previous analyses of general medium, we convert the original field variable into a series of quantitative dummy variables, coded 1 vs. zero for presence vs. absence of the particular medium type or soil mix. When we then correlate each of these single dummy variables with crop yield, we see that only three are statistically significant: Lava Rock; Multiple Media; and Soil, Fertility & Drainage Mix. The remaining dummy variables will not be used in subsequent modeling analyses, because their simple correlations with crop yield are not statistically significant.

Cypher 2 - Medium Type or Soil Mix
Correlation Significance
A=Soilless mix
-.115 .158
B=Soilless mix (FORTified)
-.089 .276
C=Crushed Aggregate
-.045 .577
D=Coconut fiber
.109 .178
.010 .904
.108 .185
I=Expanded Clay Pellets
.098 .228
-.035 .671
L=Lava Rock
.181* .025
M=Multiple Media
.227** .005
.019 .817
R=Rockwool, 4 inch cube(s) only
.028 .734
S=Rockwool, slab
.046 .571
T=Rockwool, flock absorbent only
-.016 .849
V=Soil, only
.000 .996
W=Soil, Drainage Mix (eg- verm,perlite added)
-.153 .060
Y=Soil, Fertility & Drainage Mix
-.223** .006
_-Not Available
-.122 .132

** Correlation is significant at the 0.01 level
* Correlation is significant at the 0.05 level
Number of records=153

Again, however, we will draw no firm conclusions yet about the relative efficacy of the three medium types that survived this preliminary elimination round, because there may be other hidden confounding influences that only appear to make these three media types seem important. For example, it could be that growers who happen to use a particular type of medium also tend to use more efficient lighting. Only subsequent regression modeling can successfully untangle these potentially confusing patterns of influence and allow us to see what is really going on.


Home | Analyses | Production | Lighting | About Weights | Tracking | FAQ

Copyright ©1997-2004 pH