What is considered a a wide 95% confidence interval, what is considered a small 95% confidence interval? I am revising confidence intervals, and the books that I have found mentioned a wide confidence interval and a small confidence interval but do not define what is "wider/large" or "small."
Hence the question: what is considered a narrow 95% confidence interval, and what is considered a large confidence interval?
 A: Whether an interval is wide or narrow really depends on the context of the question and data behind the construction of the interval.
If we are measuring height in feet, is a width of 10 feet wide or narrow?  Well if we are measuring the heights of human beings, then that is extremely wide (what you you do with a confidence interval that said the mean height of people in a given population is between 1 and 11 feet?).  But if we are measuring the heights of 20-50 story buildings then a width of 10 feet is moderate and if we are measuring the heights of mountains then a C.I. width of only 10 feet would be amazingly narrow.
Whether an interval is considered narrow or wide could be rephrased as "does it give enough information for me to make a decision/answer the question of interest?"
If the meteorologist says that tomorrow's temperature will be between 60 and 70 degrees F, then I know that I should plan on wearing a jacket, but probably do not need my heavy coat, so that interval is narrow enough.  But if the forecast for tomorrow is between 40 and 90 degrees F, then I don't know weather to wear my heavy coat, or the lightest/coolest clothing that I have, so that interval is too wide.  But a width of 50 degrees F would be narrow enough if I am looking at the temperature of a furnace for melting steel as long as the lower bound is above the melting point of steel.
