Science Practice 4: Building and Reading Graphs
Building and reading graphs means matching the graph type to the structure of your data, then labeling and scaling axes so the shape you see actually represents what happened.
A graph is an argument about your data, and the type you pick makes that argument. Categorical data, like growth rates in three different plant species, belongs on a bar graph because the x axis has no in-between values. Two continuous variables, like temperature and enzyme activity, belong on a scatter plot, where position along both axes carries real meaning. One continuous variable's spread, like the heights of 200 pea plants, belongs on a histogram to show frequency, or a box plot to summarize median, quartiles, and range. Pick wrong and the graph lies about your data's structure even if every number is correct.
This is why connecting bar-graph categories with a line is a real error, not just sloppy style. A line signals continuity: it tells the reader that halfway between point A and point B is a real, meaningful value. Between 'control group' and 'treatment group' there is no halfway. The line manufactures a trend that the data never claimed.
Full credit on a graph means every axis has a variable name and units, the scale is even and spans the data without wasted space, and the independent variable sits on x. These habits connect directly to other science practices: once a graph is built correctly, you read its shape to find slope, rate, or correlation, and you can transform it, like replotting exponential growth on a logarithmic y axis, to turn a curve into a straight line that's easier to measure.
Key terms
- Categorical variable
- A variable with distinct group labels, like species or treatment, that has no meaningful value between categories.
- Continuous variable
- A variable that can take any value along a range, like time, temperature, or concentration.
- Histogram
- A graph showing how often values of one continuous variable fall into different ranges, or bins.
- Box plot
- A graph summarizing a data set's median, quartiles, and spread using a box and whisker lines.
- Logarithmic scale
- An axis scale where each equal step represents multiplying by a constant factor rather than adding a constant amount.
Watch it explained
- How it worksHow do you decide between a bar graph, a scatter plot, a histogram, and a box plot?
- Common mistakeWhy doesn't a line connecting two categorical bars mean anything?
- How it worksWalk me through choosing axes, scaling, and labeling so a graph earns full credit.
- Going furtherWhat happens to a curve's shape when you replot it on a logarithmic y axis?
Check yourself
End of topic.