In terms of the success scores, we can see that the range of scores was very similar for both the workers and the wishers, but the workers contained slightly higher levels of success than the wishers. The workers had a very slightly higher median than the wishers, indicating marginally greater pre-intervention success but only marginally. Within the boxes, there is a thick horizontal line, which shows the median. It’s clear that the middle 50% of scores are more or less the same for both groups. Looking at the resulting boxplots above, notice that there is a tinted box, which represents the IQR (i.e., the middle 50% of scores). Again, this suggests very consistent evidence that the population value is greater than zero (i.e. antiSTATic works).īoxplot of success before each of the two interventions Look at how much of the confidence intervals are above zero across the 10 studies: even in studies for which the confidence interval includes zero (implying that the population effect might be zero) the majority of the bar is greater than zero. Again, this implies great consistency in the studies: they all throw up (potential) population effects of a similar size. The confidence intervals overlap with each other substantially in all studies, suggesting that all studies have sampled the same population. I would hope that some of you have changed your mind to option B: 10 out of 10 studies show a positive effect of antiSTATic (none of the means are below zero), and even though sometimes this positive effect is not always ‘significant’, it is consistently positive.
Now you’ve looked at the confidence intervals, which of the earlier statements best reflects your view of Dr Weeping’s potion?
Twenty-one heavy smokers were put on a treadmill at the fastest setting.