Home‎ > ‎Statistics prompts‎ > ‎

Averages inquiry

There is no set of data that satisfies this condition: median > mode > mean > range.
   
The inquiry starts with students trying to understand the statement and, particularly, the constraint placed on the data set. Questions and comments that have arisen in the classroom are:
  • What do median, mode, mean, and range mean? 
  • What is a set of data? 
  • Why are there ‘greater than’ signs in the sentence? 
  • What does it all mean? 
  • How many numbers are there in a set of data? 
  • Do the median, mode, mean, and range have to be in that order? 
  • If the mean is higher than the mode, you have to start with the mean, then work out the mode. 
  • It must be possible.
  • When we tried this, it worked for the median and mode, but went wrong on the mean. 
  • You can’t do it because the range is too small. 
The prompt leads students into trying to construct sets of data that fulfil a certain condition. The teacher might orchestrate a discussion on how many numbers are appropriate in the data set and guide students towards using five or seven to simplify the start of the inquiry. If students are unfamiliar with the terms in the prompt, or need to revise them, they have in the past chosen a regulatory card that gives them time to practice finding the median and mode and calculating the mean and range before exploring the statement itself. However, it is possible to develop fluency in these procedures during the inquiry.
  
In the search for a set of data described in the prompt, students begin to create sets that satisfy different inequalities. So, for example, students found this set that has mode > mean > range > median:
3   5   6   6   10   10   10
Thus, a period of exploration can lead to a question about how many permutations of median, mode, mean, and range exist. In 
one inquiry, a group of students found the 24 permutations and 
The 'block' that starts with median
median mean mode range 
median mean range mode 
 median  mode mean range 
 median  mode range mean 
 median  range mode mean 
medianrangemean mode 
explained the result by using four ‘blocks’ based on the one they demonstrated to the class, which started with the median (see box).
   
Another class working collectively found that, along with the constraint in the prompt, there are four other permutations for which it is impossible to create a data set:
Median > mode > range > mean
Median > range > mode > mean
Mean > mode > median > range
Mean > mode > range > median
This final result, if confirmed by other classes, leads onto other questions: Why are the five permutations impossible to achieve? Do they have something in common? Do they remain impossible to create if we use more numbers in the data set?
  
Update
James Handscombe (Principal of Harris Westminster Sixth Form) has found a counter-example to the proposition in the prompt: 
  
6    23    23    24    25    26    27 
median 24 > mode 23 > mean 22 > range 21 
  
James is confident that a data set of seven numbers exists for each of the 24 permutations. One question that follows concerns the smallest set of data that satisfies all the permutations - is it 7, 6, or 5 numbers?

You can follow James on twitter @JamesHandscombe
   

Resources
Prompt sheet
Promethean flipchart    download
Promethean flipchart (alternative prompt)   download
Smartboard notebook   download

 
Venn diagrams and the prompt
Craig Barton (the TES adviser for secondary maths) has created tasks involving Venn diagrams and a wide selection of mathematical topics. The one for averages and range would make an effective means of assessing students' conceptual understanding that develops from the prompt. The task builds up to diagrams that require students to place numbers in three sets in such a way as to satisfy the constraints.

You can follow Craig on twitter @tesMaths. A full list of the excellent Venn diagram tasks can be viewed here.
  

The prompt as a problem    
Two months after the averages inquiry was posted on the site, NRICH published Unequal Averages – a problem solving task for students based on the well-established “M, M, and M” problem. While the task is similar to the inquiry presented on this page, the differences reflect the distinction between problem solving and inquiry. Problem solving has a pre-determined beginning (the question) and end (the solution). In the NRICH activity, the set of inequalities is defined and so is the size of the sets to be used (four, five, and six numbers). It is also clear that the problem setter knows the solutions, including those inequalities that cannot be satisfied. The problem solver is left with an 'open middle' in which to explain how we can arrive at the required end-point from the given starting point. In contrast, an inquiry has the potential to have an open beginning, middle, and end. This is more motivational for students, as well as giving them the opportunity to pose mathematical questions and determine appropriate solution sets. The NRICH activity does, however, suggest one way to run a structured inquiry. The teacher could restrict the prompt to the three averages and limit the size of the set of numbers in the initial inquiry.

Inquiry as a frame for learning
Ollie Rutherford who is on the Teach First programme at Haverstock School (inner London, UK) used the alternative averages prompt to initiate the first inquiry with his year 7 class. The picture below shows the students' initial questions and observations.
After the initial phase of the inquiry, Ollie used the pack of six regulatory cards to give students the opportunity to direct the inquiry. He recorded their choices on the board:
Card Frequancy
 Find more examples. 6
 Practise a procedure. 6
 Change the prompt. 4
 Inquire with another student. 2
 Ask the teacher or a student to explain. 3
 Decide on the aim of the inquiry. 1
   
As the aim of the inquiry, the class decided to find sets of data that satisfied the six inequalities possible with mean, median and mode:
median > mean > mode
median > mode > mean
mean > median > mode
mean > mode > median
mode > median > mean
mode > mean > median
The activity covered the first three cards and Ollie attended to those students who required more of an explanation than had emerged in the question and observation phase. At the end of the lesson, students presented their sets of data. Usually a quiet student, Billy was excited to create his own mathematics and enthusiastic to show his data set to the class. He explained how {1 1 9 11 12} satisfied the condition that median > mean > mode. Then Ahmad used negative numbers {-12 -2 1 1 2 3 4 5 6} to create a set of date for which median > mode > mean.
  
Ollie reports on how the inquiry framed students' learning: "Only a few students in my year 7 group had come across the three different types of averages and range as a measure of spread before. This inquiry was particularly useful for two reasons. First, it helped frame students’ understanding of averages: I found it a powerful way to introduce averages as it heightened students’ interest in the topic. After discussing the prompt, I taught them averages in the way that I normally do, but I then referred back to the prompt, which gave students a sense of closure and deeper understanding of the topic. Second, starting with the prompt sparked students’ curiosity, which assisted with their motivation to carry out the work and gave them more agency in doing so."

You can follow Ollie on twitter @Oliver_Kourosh.

Classroom inquiry
These are the initial questions and observations of a year 9 (grade 8) class about the prompt:
   
  
When invited to decide on the direction of the inquiry by selecting a regulatory card, the majority of students opted to "find more examples". The first lesson ended with one student presenting a set of data that satisfies the condition, although he used nine numbers instead of the seven agreed upon by the class:
    
31   32   53   53   56   57   58   59   60
median 56 > mode 53 > mean 51 > range 29
   
In the second lesson of the inquiry, students established four aims of their own:
  • Find sets of four, five, six, and seven numbers for which mean = median = mode = range.
  • Find sets of five numbers that satisfy inequalities involving permutations of mean, median, and mode only.
  • Find sets of seven numbers that satisfy inequalities involving permutations of mean, median, mode, and range.
  • Find a set of numbers that satisfy the same permutation of mean, median, mode and range as in the prompt, but with the inequality reversed to "less than".
After selecting an aim, the students explored in pairs or groups before explaining their findings to the rest of the class. 
   
Constructing an explanation
The inquiry can lead to some ideas that students find difficult to explain. Why is it that they cannot create a data set that fulfils the inequality in the prompt? In one discussion, the teacher and class tried to construct an explanation by focusing on one case. They used seven positive whole numbers in the data set. The class also tried to make as many of the median, mode, mean, and range as possible positive whole numbers.
  
Start with the median, let's say five. __  __  __  5  __  __  __
The mode has to be lower. We could arrange the numbers (with the mode shown by *) like this:  *   *   __  5  __  __  __ 
Or like this:  __  *   *   5  __  __  __ 
Or this:   *   *   *   5  __  __  __
If the mode is three, then the range is two. However, a contradiction arises because then the mode will become five when we fill in the remaining places. 3   3  __  5  __  __  5 
So we make the mode four, and the range three. However, again we reach a contradiction because whether we use fives or sixes to fill the final two places the mode is not four (or at least not on its own).3   4   4   5  __  __  6
So, we make all three numbers less than the median four. The greatest value is then seven and the mean must be below four, but above three. This is not possible whatever numbers we decide upon to fill the last two places and we have our final contradiction. 4   4   4   5  __  __  7
   
This did not constitute a proof of impossibility, even if we have exhausted all the possibilities in this one case. The class decided to inquire into whether changing the assumptions at the beginning - that is, only positive whole numbers in the data set, aiming at positive whole numbers for the median, mode, mean and range, using seven numbers, and starting with five as the median - led to a different outcome. Students chose one assumption to disregard. However, they could still not create a set of data to meet the conditions in the prompt. The class concluded that it is impossible.