Common use of Plotting Clause in Contracts

Plotting. In summary plots, we visualized mean values and %95 confidence interval values for our data using the ggplot2 package (▇▇▇▇▇▇▇ & ▇▇▇▇▇▇▇, 2007). When reading summary plots, we are mainly interested in whether or not confidence intervals overlap or not. Our confidence intervals were computed following ▇▇▇▇▇ (2008) and his correction of ▇▇▇▇▇▇▇▇▇ (2005). The reason for using these computed CIs instead of just standard errors is to include uncertainty due to sampling between different groups observed. We also assessed the variance in the difference between the two conditions. ▇▇▇▇▇▇▇▇▇ (2005) recommends using each group’s standard deviation in calculating the CIs. Moreover, we multiplied our intermediate number with 1.98 to achieve %95 CIs. In posterior plots, we visualized the mean of Bayesian model coefficients. We included %50 and %90 posterior intervals, and the probability of each coefficient to be smaller than 0.1 or bigger than 0.1, which are Region of Practical Equivalence Region borders ▇▇▇▇▇▇▇▇ and ▇▇▇▇▇▇▇ (2018). This ROPE region indicates no practical effect of a coefficient. If a distribution is completely outside this area, we can say we have definitive evidence for an effect. If it covers the practical equivalence area, we can say that according to our data, there seems to be no evidence for an effect. On occasions in which only a part of the distribution resides in the area, we explicitly quantify our degree of belief towards an effect. In this thesis, we always fit the yes responses to our stimuli. Negative values indicate a decreasing effect on the average number of yes responses. In contrast, positive values indicate an increase in the average number of yes responses.

Plotting. In summary plots, we visualized mean values and %95 confidence interval values for our data using the ggplot2 package (▇▇▇▇▇▇▇ & ▇▇▇▇▇▇▇, 2007). When reading summary plots, we are mainly interested in whether or not confidence intervals overlap or not. Our confidence intervals were computed following ▇▇▇▇▇ (2008) and his correction of ▇▇▇▇▇▇▇▇▇ (2005). The reason for using these computed CIs instead of just standard errors is to include uncertainty due to sampling between different groups observed. We also assessed the variance in the difference between the two conditions. ▇▇▇▇▇▇▇▇▇ (2005) recommends using each group’s standard deviation in calculating the CIs. Moreover, we multiplied our intermediate number with 1.98 to achieve %95 CIs. In posterior plots, we visualized the mean of Bayesian model coefficients. We included %50 and %90 posterior intervals, and the probability of each coefficient to be smaller than 0.1 −0.1 or bigger than 0.1, which are Region of Practical Equivalence Region borders ▇▇▇▇▇▇▇▇ Kruschke and ▇▇▇▇▇▇▇ Liddell (2018). This ROPE region indicates no practical effect of a coefficient. If a distribution is completely outside this area, we can say we have definitive evidence for an effect. If it covers the practical equivalence area, we can say that according to our data, there seems to be no evidence for an effect. On occasions in which only a part of the distribution resides in the area, we explicitly quantify our degree of belief towards an effect. In this thesis, we always fit the yes responses to our stimuli. Negative values indicate a decreasing effect on the average number of yes responses. In contrast, positive values indicate an increase in the average number of yes responses.