Answer:
Let's analyze the options:
"The y -axis starts at zero and uses consistent intervals."
This is the way that a graph is usually read, so this is not misleading at all.
"The y -axis starts at zero but has different intervals."
If the intervals change and is not very noticeable, then a quick view of the graph may be misleading, because you have the values of y, but do not know the actual relation between x and y.
"Different bar widths are used on the same graph."
Different bar widths may be used when you want to include more than one measure un one bar, (for example in a graph where each day has his own bar, except for the weekend that has only one bar for the two days)
This may be misleading because when you use wide bars you may not be clear enough, for example in the prior example, we have that Sunday and Saturday shared a bar, so you can not know if the values of Y are related to one particular day or not.
"The y -axis doesn't start at zero."
Ok, if the y-axis does not start at zero may be misleading because you do not have an actual, in this type of graphs very large values of Y may seem small.
"Different bar heights are used on the same graph."
You almost always will see different bar heights, so this is not misleading
"The intervals on the y -axis are too large."
If the intervals are too large, then a we need a large change on y to notice it in the graph. This means that this type of graph does not accurately describe small changes, so this can be used to "hide" big changes in the value of y, so this can be misleading.
