Classroom Case Studies, Grades 3-5 Part C: Inferences and Predictions (30 minutes)

The NCTM (2000) data analysis and probability standards state that students should “develop and evaluate inferences and predictions that are based on data.” In grades 3-5 classrooms, students are expected to develop and evaluate inferences and predictions, to propose and justify conclusions and predictions that are based on data, and to design studies to further investigate their conclusions or predictions.

Inference and prediction are more advanced aspects of working with data, as they require some notion of the ideas of sampling and population. Students in grades 3-5 are only beginning to develop an understanding of sampling. They often trust their own intuition more than the information they are obtaining from the data. Children begin to develop an understanding of these statistical ideas through conversations as they consider what the data are telling us, what might account for these results, and whether this would be true in other similar situations. Children’s early experiences are often with census data — e.g., the population of their class. When they begin to wonder what might be true for other classes in their own and other schools, they begin to develop the skills of inference and prediction. It is not until the later middle grades and high school that students begin to learn ways of quantifying how certain one can be about statistical results. See Note 6, below.

When viewing the video segment, keep the following questions in mind:
• How does Ms. L’Esperance encourage students to make inferences and predictions?
• What are some of the students’ preliminary conclusions?
• How are the ideas of sampling and population embedded in this conversation?

Video Segment
In this video segment, Suzanne L’Esperance facilitates a whole-class discussion as the students consider potential conclusions to the original problem on how large to build the house. Students discuss the variance in their data, the limitations of their small sample, and the need for additional data.

Problem C1
Answer the questions you reflected on above as you watched the video:

 a. How does Ms. L’Esperance encourage students to make inferences and predictions? b. What are some of the students’ preliminary conclusions? c. How are the ideas of sampling and population embedded in this conversation?

Problem C2
Based on the family-size data gathered by the class and shown in the line plot below, how would you respond to Ms. L’Esperance’s initial question: What size should she tell her friend to build his homes? What reasons can you offer to support this response, and how are they related to the ideas you have studied in this course? Are your reasons based on the data collected, or did you also bring in some of your own judgements?

Problem C3
Children are expected to develop and evaluate inferences and predictions. Evaluate each of the responses below by commenting on the following:

 • why the response makes sense (or doesn’t) based on the data; and • the limitations of each response. In other words, what statistical ideas are the children not taking into account?

The children’s responses to the question of how big to build the homes were as follows:

 a. “He should build homes for four people.” b. “You can tell him to put in a couple of each, because some people live with two people, so he should put more fours and threes, but put some of the other kinds also.” c. “He should build them for three people and four people.” d. “I know some people that have six and eight people in their families, so he should build some larger houses too.”

Problem C4
In thinking about the data that were collected, in what ways might the students’ sample be biased? How might you facilitate a discussion with the students about bias in data? What questions would you pose? What issues would you raise?

Problem C5
According to the 2000 census, the average size of households in the United States is 2.62 people. How might your students respond to this information in light of their own data? What statistical ideas would you want to surface in this discussion?

Problem C6
If you were teaching this lesson on investigating family size, what questions could you ask students to encourage them to focus on each of these central elements of statistical analysis?

 • Defining the population • Defining an appropriate sample • Collecting data from that sample • Describing the sample • Making reasonable inferences relating the sample and the population

Problem C7
A student commented that the class should “wait until we get more information” before making a recommendation to Ms. L’Esperance’s friend. How could you extend this conversation to bring out more predictions and then formalize these notions into stated conjectures that could then be investigated further? What questions would you ask? What are some conjectures that might result? How could these be investigated?

Principles and Standards for School Mathematics (Reston, VA: National Council of Teachers of Mathematics, 2000). Standards on Data Analysis and Probability: Grades 3-5, 176-181.
Reproduced with permission from the publisher. Copyright © 2000 by the National Council of Teachers of Mathematics. All rights reserved.

Notes

Note 6
You might want to review the statistical ideas of bias in measurement, samples, and populations, and the meaning of “conjecture,” possibly providing one or two examples.
Session 1, Part D: Bias in Sampling

Solutions

Problem C1
Here are some possible answers:

 a. Ms. L’Esperance encourages students to make inferences and predictions by focusing their attention on the problem context and asking them to make suggestions regarding what she should tell her friend about how big to build his homes. b. Many of the children concluded that Ms. L’Esperance should tell her friend to build homes for four people. However, other children took into account the variance in the data and concluded that, while he should build some homes for two people, he should build the most homes for three or four people. c. When the teacher asks the students to consider the number of data points collected (the sample size), she implicitly encouraged them to consider ideas of sampling and population.

Problem C2
The data make a strong case that homes should be built for families of size two, three, four, and five. You may agree with the students that four is an appropriate conclusion, but you probably also realize that this sample is very small and that more data should be gathered.

Problem C3

 a. This response makes sense in that it is based on the mode; the limitation is that it does not take into account the variation in the data. b. This response takes into account the variation in the data. c. This response is based on the two values with the greatest number of responses, so the student does consider variation in a narrow sense but does not take into account the limited sample. d. In this context, this response doesn’t make sense; the student has gone beyond the actual data involved and is considering issues of sampling and population.

Problem C4
The sample is biased in that, as children, they all live in households that contain at least two people; thus, households in which one person lives are not considered. Some questions a teacher might pose include, “Why doesn’t our line plot show any families of size one?” and “Does anyone in your neighborhood live in a household with only one person?”

Problem C5
The students are likely to wonder why the average size of households is so much smaller than what their data indicated. You would want students to think about how their sample was collected and the bias or limitations inherent in their sample.

Problem C6
Here are some questions you might ask:

 • What should we tell my friend about where this information came from and the part of our city in which he should build homes of this size? • If my friend decides to build houses in another city, should they be the same size as the houses we think he should build here?

Problem C7
Two conjectures that might result are, “The typical family size in our area is four people” and “You will not find families in our area that have 10 people.” These could be formulated as new questions to be investigated: “What is the typical family size in our area?” and “What is the range of family size in our area?” The students could investigate this question in several ways. They might want to survey students in other classes and grade levels in their school on family size, they might want to have each student survey 10 neighbors, or they might want to locate census data for their community.