Complex experiments

The types of experiment that are used in agriculture are often quite complex. This page describes a few real-life examples.

Rice yield and fertiliser

Aim
Researchers wanted to determine the effect of two inorganic fertilisers, ammonium sulphate and ammonium phosphate on the yield of paddy rice. A seconary aim was to assess whether the fertilisers have different effects on different varieties of rice.
Treatments
Two factors were varied — the variety of rice used (Red Aus, Kashiful and Dudkalama) and the fertiliser applied (ammonium sulphate at 30.4 lb per plot, ammonium phosphate at 40.4 lb per plot and a control level with no fertiliser added)
Design
Six blocks of land were used and each was split into nine plots, each measuring (12' by 39'). Within each block, the nine treatments (combinations of rice variety and fertiliser) were allocated at random.
Response
The yield was recorded from each plot. (Its units are "chattaks" where 1 chattak is approximately 58 grams.)
Data
The table below shows the data that were collected.
Treatment Block
Variety Fertiliser 1 2 3 4 5 6
Red Aus Ammophos
Ammosulph
Control
  112  
168
106
  128  
116
 84
  118  
144
 68
  128  
100
156
   92  
100
156
  152  
 80
128 
Kashiful Ammophos
Ammosulph
Control
112
 61
 97
 81
 98
 86
108
 58
 92
 96
 86
 80
 53
 65
 99
 48
 98
 66
Dudkalama Ammophos
Ammosulph
Control
134
125
 62
112
106
 60
116
110
 99
114
102
 90
101
 56
 58
128
110
 87
Experimental units
The experimental units are the 54 plots and the 9 plots within each block may be related — some blocks may have more fertile soil than others.

Calliandra feeding trial

Preliminary work showed that fresh leaves of Calliandra calothyrsus provide a fodder supplement for dairy cattle. Nutritional analysis suggests that 3 kg of Calliandra should be nutritionally equivalent to 1 kg of concentrate.

Aim
The experment was conducted in Embu, Kenya, to determine whether Calliandra will substitute (in terms of milk yield) for dairy concentrate at the rate of 3 to 1 when added to the basal diet of cows in Embu.
Treatments
1. Basal diet alone
2. Basal + 3 kg Calliandra
3. Basal + 1 kg concentrate
4. Basal + 3 kg Calliandra + 1 kg concentrate
Design
Twelve Farmers, each with a lactating cow, are used in the study. Each cow is allocated to a sequence of treatments — they are on one treatment for a period, and then change to the next. The order of the treatments is varied between cows so that each treatment is used with the same number of cows in each period. A treatment period is 14 days with milk yield assessed during the second seven only.
Response
Average daily milk yield (kg) over days 8-14.
Data
The table below shows the data that were collected.
  Period
Cow 1 2 3 4
1   13.88 (T1)     13.30 (T2)     12.80 (T3)     13.25 (T4)  
2 10.29 (T1)  9.33 (T3)  9.91 (T4)  8.65 (T2)
3  8.40 (T1)  9.30 (T4)  9.80 (T2) 10.30 (T3)
4 10.70 (T2) 10.19 (T1) 10.17 (T4) 10.10 (T3)
5 15.31 (T2) 15.07 (T3) 13.06 (T1) 13.67 (T4)
6  8.13 (T2)  9.06 (T4)  8.36 (T3)  8.19 (T1)
7 10.50 (T3)  9.86 (T1) 10.50 (T2) 11.12 (T4)
8  8.10 (T3)  9.40 (T2) 11.07 (T4) 10.50 (T1)
9 11.30 (T3) 12.57 (T4)  9.37 (T1) 10.70 (T2)
10 10.06 (T4)  8.36 (T1)  9.50 (T3)  9.15 (T2)
11 12.05 (T4) 11.83 (T2) 10.64 (T1) 11.00 (T3)
12 15.05 (T4) 13.88 (T3) 13.72 (T1) 13.44 (T2)
Experimental units
An experimental unit is a cow over a single 2-week period. These experimental units have some structure — some cows are more productive than others, so the four units for each cow are related. There may be a steady increase or decrease in yield for all cows over the 2 months of the study but the design ensures that this does not affect the conclusions.