MDM 4U - Student Page
How Faithful is Old Faithful? – An Activity in Statistical
Thinking
Imagine that you have just arrived at Yellowstone National Park, the home of geyser basins,
thermal mud pots, hot springs, acid lakes, and a multitude of fascinating animals and plants.
Furthermore, you have just missed the most recent eruption of Old Faithful Geyser, which has
periodically been spewing streams of hot water high into the sky for centuries.
How long would we have to wait until the next eruption of Old Faithful?
Before reading further, write down your best estimate.
If you have ever visited a geyser, you might have some basis for making an estimate. Some
geysers are dormant between eruptions for many hours or days. Other geysers erupt almost
continuously.
A strategy that can help to make a prediction is to use past data on Old Faithful.
One day’s worth of data:
Old Faithful erupts approximately 20 times per day. The data show the numbers of minutes
between the time when Old Faithful stopped erupting to when it first began to erupt again.
Day 1:
51
82
58
81
49
92
50
88
62
93
56
89
51
79
58
82
52
88
If you had some friends who were planning to visit Yellowstone National Park, how long should
you tell them to expect to wait between eruptions of Old Faithful. Make an estimate. Give
some justification for your prediction.
Construct a graph.
Of course, one day’s worth of data does not give much basis for a prediction. Two more days
of Old Faithful eruption data, picked at random from a larger data set, follow:
Day 2:
86
78
71
77
76
94
75
50
83
82
72
77
75
65
79
72
78
77
Day 3:
65
89
49
88
51
78
85
65
75
77
69
92
68
87
61
81
55
93
Construct graphical representations of the second and third days’ data, similar to the
representations of the first day’s data. Compare the data for the three days. At this point what
would you predict for your friends? How long should they expect to wait for the next eruption
of Old Faithful?
Share and discuss the graphical representations in your group.
MDM 4U – Teacher Notes
How Faithful is Old Faithful? – An Activity in Statistical
Thinking
In this activity data from Old Faithful geysers are introduced to highlight the
important role that variation should play in our teaching of statistics. Our past
teaching may have over-emphasized the role of centers to the neglect of issues of
spread and variability.
ADDITIONAL QUESTIONS TO PROMOTE STUDENTS’ STATISTICAL THINKING
Before students are given data:
How much data would you need for a prediction? One wait time? Two wait times? One day’s
worth of wait times? Two day’s worth of data? A year’s worth of data?
When they are given data:
How does your prediction using data compare with your first prediction without the data?
After they draw a graph of the first day’s worth of data:
What patterns if any do you notice?
After they draw graphs of two or more days’ data:
How do your predictions compare with your previous ones?
When they share graphs:
Compare and contrast the information revealed by each graph. What is gained or lost with the
various graphical representations? How are the graphs related to one another?
What are the advantages and disadvantages of each of the following types of graphical
representations?
Stem and Leaf Plots, Histograms, Box and Whisker Plots, Dot Plots and Bar Plots, 2
Variable Plots.
When they are ready to interpret the information:
What other data would be helpful, for example, duration time of eruptions, to enable you to
further understand the wait time between blasts? What does the oscillating pattern tell you
about how this geyser works? What other information on this geyser system would help you
interpret this pattern?
When they are ready to make a final prediction:
Will this prediction be true for the whole year? Will seasonal variation occur? Will this
prediction be true over several years? Will yearly variation occur? What, if any, limitations
should you put on your prediction? Is your prediction of oscillation valid? In one day, for
example, the alternating pattern may be long, long, short, and so on. What graphical
representation could you use to verify whether the alternating pattern of short and long wait
times is generally true?
When they are ready to communicate the information to others:
What graphical representation would best communicate your prediction? What other informatin
besides your prediction should be communicated?
At the end of the inquiry:
What have you learned about making predictions? About variation? How well do your
predicted times compare with the actual times? If you continued with this problem what would
you investigate next? Can you find an explanation for the pattern? Does a relationship exist
between duration times of eruptions and wait times between eruptions?
More Old Faithful Data:
One strategy that could be used is to give each student a strip showing a day’s
worth of Old Faithful data; have them analyze, graph and make predictions from it;
then have students trade several times with other students; and repeat the process
with data for several other days.
86
65
91
89
54
51
52
86
79
87
78
65
53
79
72
63
71
82
50
60
80
82
78
78
75
69
52
89
84
89
84
84
57
84
87
84
73
58
69
71
78
55
98
49
70
87
47
76
80
54
48
69
81
81
75
77
64
83
48
88
73
76
84
62
75
85
93
74
62
49
77
76
80
49
78
51
93
59
57
83
77
58
54
71
81
92
53
94
49
82
79
78
50
80
87
50
60
79
86
108
71
50
80
75
88
57
65
85
87
89
68
85
86
57
53
50
79
88
55
50
54
84
84
65
77
45
86
78
77
88
78
77
81
62
87
83
85
57
50
75
74
93
75
78
56
68
52
57
74
93
53
82
51
84
83
77
72
72
73
81
81
76
83
80
59
56
85
72
96
73
60
69
82
71
53
78
50
78
60
61
81
89
61
77
50
78
80
92
74
54
82
76
89
74
87
82
66
51
93
75
80
57
50
68
80
79
93
74
54
85
49
48
87
79
54
65
78
79
88
87
49
74
77
81
90
75
80
81
53
58
76
79
81
57
50
61
91
65
54
66
73
65
60
73
80
82
80
72
72
90
84
81
53
78
96
84
60
76
92
62
50
52
81
78
75
62
74
55
86
57
48
48
83
58
43
79
87
88
59
77
78
87
76
93
49
87
89
93
Statistical Thinking involves an understanding of variation, its omnipresence and its
explanation.
Students must look for patterns in data, deal with the variation and make judgements and
predictions on the basis of the data.
Five Elements of Statistical Thinking:
! Recognition of the Need for Data
Many real situations cannot be judged without gathering and analyzing properly collected
data. One’s own experience may be unreliable and misleading for judgements and for
decision making. Research suggests that students think that their own judgements and
beliefs are more reliable than data.
!
Transnumeration (a coined word meaning “Numeracy transformation for facilitating
understanding)
A dynamic process: Collect data, construct multiple statistical representations, communicate
to others what the representations suggest.
!
Consideration of Variation
Notice that variation exists. Ask why patterns appear.
!
Reasoning with Statistical Models
Selecting or creating a model that optimally represents and communicates the nature of the
real problem and that focuses our reasoning about the data.
Focus on the patterns in distributions and patterns in the centres and spreads, not on
individual pieces.
!
Integrating the Statistical and Contextual
The statistical model must capture elements of the real situation, and the data tell a story.
Mathematics Teacher Vol. 95, No. 4 April 2002
Examples of Graphs: