Tables, Graphs & Scale

Learning objectives

• Construct a data table whose headings identify variables, units, raw observations, and calculated summaries.
• Choose among bar charts, line graphs, and scatter plots based on variable type and scientific purpose.
• Assign independent/explanatory variables and dependent/response variables appropriately to axes.
• Select and defend readable axis ranges and intervals without visually exaggerating or hiding patterns.
• Distinguish interpolation within observed data from extrapolation beyond the tested range.
• Critique a graph for missing context, misleading scale, unsupported connecting lines, or claims that exceed the evidence.

Materials

• Science notebook
• Graph paper and ruler, or accessible spreadsheet/graphing tool
• Paper-helicopter raw data and summaries from Day 06
• Dataset Lab through Science Studio when useful

1. A graph is an argument about data

The same numbers can look calm, dramatic, clustered, or nearly flat depending on graph type, axis range, interval, aspect ratio, and which observations are shown. A scientific graph should make the evidence easier to inspect—not manipulate the reader's impression.

2. Build tables that preserve the evidence trail

Table headings should name the variable and unit. Raw observations should remain distinguishable from calculated values such as mean and range.

• Place units in headings rather than repeatedly in every numerical cell when one unit applies to the whole column.
• Use consistent decimal conventions that reflect the measurement method.
• Do not silently replace raw values with rounded summaries.
• Mark missing observations explicitly rather than entering a zero unless zero was actually measured.

3. Choose the display from the variables

For rotor length versus mean fall time, both variables are numerical. A scatter plot is a strong default because it preserves the tested x-values without requiring you to imply that every value between them was observed.

4. Axes carry scientific meaning

In a controlled experiment, the deliberately changed independent variable normally belongs on the x-axis and the measured dependent variable on the y-axis. Axis labels need both variable names and units.

A title should describe what was plotted, not announce a conclusion such as “Long Rotors Are Best.”

5. Scale, range, and visual integrity

A truncated axis is not automatically dishonest. Sometimes a narrow range is necessary to inspect small but meaningful differences. The problem occurs when scale choices hide context or create an impression disproportionate to the actual numerical difference.

• Use evenly spaced numerical intervals unless a specialized scale is explicitly identified and justified.
• Avoid unexplained breaks in an axis.
• Choose enough range to show the data clearly without wasting most of the plotting area.
• Inspect whether the graph's proportions make tiny differences appear enormous.

6. Points, connecting lines, and trend lines

Connecting plotted points can imply that values between measured conditions are known or meaningfully continuous. A trend line summarizes a pattern; it does not transform association into proof of causation and does not guarantee predictions beyond the data.

7. Interpolation versus extrapolation

Interpolation estimates within the range of observed x-values. Extrapolation projects beyond that range. Extrapolation is generally riskier because the relationship may change outside the tested conditions.

8. Interactive Graph Integrity Lab

Use your Day 06 dataset in the Dataset Lab or another approved graphing tool. Produce two legitimate visualizations using different defensible axis ranges. Compare how the visual impression changes while the numerical data remain identical.

1. Verify all plotted values against the original table.
2. Check title, axis labels, units, and scale intervals.
3. Identify the strongest visible pattern.
4. Identify one feature the graph does not show—for example, individual-trial variation if only means are plotted.
5. Choose the version you would publish and explain why it communicates most honestly.

The tool creates the display. You remain responsible for the scientific choices and interpretation.

9. Graph forensics

Audit any scientific graph using this sequence:

1. Source: Where did the data come from?
2. Variables: What exactly is being compared?
3. Units: Are they present and consistent?
4. Scale: What range and intervals are used?
5. Selection: Are relevant observations, dates, groups, or uncertainties omitted?
6. Encoding: Do bar lengths, areas, icons, or lines visually represent the numbers faithfully?
7. Claim: Does the caption/conclusion say more than the graph demonstrates?

10. Mentor visualization defense

Bring your final graph and alternate visualization to the Science Mentor. Be ready to defend:

• why this graph type fits the variables;
• why the axis range and interval are fair;
• whether connecting points or fitting a trend is justified;
• what information is lost by graphing means instead of all raw trials;
• what claim a careless reader might make that your evidence does not support.

Mentor review supports revision. It does not replace the student's graph, source data, or mastery assessment.

11. Choose your pathway

12. Exit check

1. When is a bar chart preferable to a scatter plot?
2. Why can a truncated axis be useful yet potentially misleading?
3. Why is a missing observation not equivalent to zero?
4. What claim can a trend line not establish by itself?
5. How does extrapolation differ from interpolation?
6. What should a reader inspect before trusting the visual impression of a graph?