7mm Analysis
Figure 1 shows the distributions of thermal spread by side and device. The figure shows that the mean is greater on both sides for the Voyant device (\({M}_{\text{right}}=2.54\), \({M}_{\text{left}}=2.32\)) than it is for the LigaSure device (\({M}_{\text{right}}=1.58\), \({M}_{\text{left}}=1.49\)). In addition, the lines connecting the means for each side are roughly parallel. If an interaction were present, such that the devices differed more for one side than the other, these lines would not be parallel. No such evidence for an interaction is present in the figure.
Table 1 gives the results of the two-way factorial ANOVA. The ANOVA confirms that that the interaction between side and device is not significant (\(p=0.73\)). The main effect of device is statistically significant (p < 0.001), which indicates a significant difference in between the Voyant and LigaSure devices. The main effect of side is not significant (p = 0.06).
Table 1
ANOVA Results (Cut-Depth: 7mm)
Term | SS: Type III | DF | F Statistic | P Value |
Device | 20.20 | 1 | 205.85 | < 0.001 |
Side | 0.35 | 1 | 3.53 | 0.063 |
Interaction | 0.01 | 1 | 0.12 | 0.734 |
Residuals | 10.60 | 108 | | |
4mm Analysis
Figure 2 shows the distributions of thermal spreads of the LigaSure and Voyant devices. The mean LigaSure thermal spread is slightly greater on the right side than on the left (\({M}_{\text{right}}=1.56\), \({M}_{\text{left}}=1.52\)), while the left side is slightly greater than the right side for Voyant (\({M}_{\text{right}}=2.26\), \({M}_{\text{left}}=2.35\)). The lines connect the sides between the two devices cross, suggesting the possibility of an interaction, although the slopes for each line do not appear to be very distinct.
Table 2 shows the result of the ANOVA for the cut-depth of 4mm. Despite the evidence in the figures above that there may be an interaction between the factors of side and device, the interaction effect is not significant in the ANOVA results (\(p=0.23\)). The main effect of device is once again significant at the \(\alpha =0.05\) level (\(p<0.001\)), but the main effect of side is not (\(p=0.68\)).
Table 2
ANOVA Results (Cut-Depth: 4mm)
Term | SS: Type III | DF | F Statistic | P Value |
Device | 16.28 | 1 | 182.37 | < 0.001 |
Side | 0.02 | 1 | 0.17 | 0.682 |
Interaction | 0.13 | 1 | 1.43 | 0.235 |
Residuals | 9.64 | 108 | | |
Combined Analysis
The third analysis combines the left and right thermal spread measure into a single “total” measure of thermal spread, or the sum of the left and right measures. This analysis treats cut-depth (4mm and 7mm) as an independent variable along with the device type.
Figures 3 shows that the thermal spread for the Voyant device is greater than the thermal spread of the LigaSure device for both cut-depths. While the LigaSure device has a consistent distribution at both cut-depths (\({M}_{\text{4mm}}=3.08\), \({M}_{\text{7mm}}=3.07\)), the Voyant device has a larger mean thermal spread at 7mm (\({M}_{\text{7mm}}=4.77\)) than it does at 4mm (\({M}_{\text{4mm}}=4.61\)).
Table 3 shows the ANOVA results for the total thermal spread show that there is no interaction between the factors of cut-depth and device (p = 0.27). The type of device again has a significant main effect (p < 0.001), while the cut-depth does not (p = 0.34).
Table 3
ANOVA Results (Total Thermal Spread)
Term | SS: Type III | DF | F Statistic | P Value |
Device | 72.76 | 1 | 415.41 | < 0.001 |
Cut-Depth | 0.16 | 1 | 0.90 | 0.344 |
Interaction | 0.21 | 1 | 1.21 | 0.274 |
Residuals | 18.92 | 108 | | |