Rachel has a large number of songs on her music player. The song lengths, in minutes, are summarized in the boxplot below.
Rachel decided to listen to ten songs each morning. She set the music player to select the songs at random. The boxplots below summarize song lengths for Monday, Tuesday, and Wednesday of that week.
Which of the following statements is true?
(A) On Thursday, song lengths will have a greater median and less variability than song lengths from Wednesday.
(B) Something must be wrong with the music player, because the boxplots for Monday, Tuesday, and Wednesday all look very different.
(C) Something must be wrong with the music player, because the boxplots for Monday, Tuesday, and Wednesday are different than the population boxplot.
(D) The song lengths varied less on Wednesday than the song lengths on Monday or Tuesday.
The correct answer to this question is Option (D). Both the range (the distance between the minimum and the maximum) and the interquartile range (IQR) (the length of the “box” that includes about 50% of the data points) are shorter on Wednesday than on Monday or Tuesday, which implies that the song lengths varied less on Wednesday. 71% of students answered this question correctly.
Option (A) is incorrect because it is based on a pattern in the way the medians and variability change across the random samples. In reality, one random sample is independent from another, so there is no reason to expect a pattern.
Options (B) and (C) both represent misunderstandings of sampling variability. When multiple samples are drawn from the same population, we do not expect the samples to look exactly the same as the population or exactly the same as each other. In this question, sampling variability is illustrated through boxplots: although each sample was drawn from the same population, the boxplots look somewhat different. We can also talk about the sampling variability in terms of numerical summaries. For example, the median song length varies from one sample to another.