Run Number | Bubble size | Vortex | Distribution of bubbles | Foam | Evenness of flow | 1 | Large | In Middle | Evenly distributed in the bulk with a lot present near baffles | Foam present at top | Even in the middle | 2 | Larger | Present | | | More turbulent | 3 | Smaller Bubble size | No vortex seen | | Lot less foam | It was noticed that one air bubbler was bubbling more air than the air. Discuss why this happened. | 4 | Small bubble size | Present | Much less bubble formation | Present at top | Turbulent | 5 | Medium sized bubbles. Bigger than run 3. | No vortex seen | Less bubble formation | Less foam | Bubbles were seen evenly generated from both bubblers | 6 | Medium sized | Present | Evenly distributed in the bulk with a lot present near baffles | Present at top | Turbulent | 7 | Large | No vortex seen | | No Foam | Bubbles were evenly formed from both bubblers |

5.3 Analysis of Variance in parameters based on pooled data
Variance in data was observed in each set of experiments due to a number of factors. These factors include stirrer speed, oxygen content, and the different sizes in air stones.
From the ANOVA illustrated in Figure 1, the sources X1, X2, X3, X4 are oxygen flow rate, stone size, stirrer speed, and groups respectively. It is seen that the largest variance occurs at the stirrer speed parameter, with a F value of 24.71 as compared to oxygen flow rate at 1.82, stone size at 1.98 and groups at 0.15. This shows that the stirrer speed plays the most significant role in the variation between the sets of data.
One of the reasons that might have caused this outcome was the formation of a vortex at high stirrer speeds. Although a high stirrer speed is needed for even mixing, the vortex formed creates a vacuum around an area this disrupts the mixing of the oxygen.
Another factor that might have caused a variation in data is the size of the bubbles produced during the runs. A larger bubble size suggests a larger surface area of reaction which leads to a more thorough mixing of the oxygen content in water which results in a higher K value.
Group variation at F= 0.15 is seen to be the smallest. This suggests that there is a good consistency of the calculated error variations between the groups.
It was also seen that on run 3, there were more bubbles forming on the one of the bubblers compared to the other. We deduced that this might have been the cause of the inner tubing being congested as it was rectified on the next runs.