Welcome to Graphical Analysis and Interpretation of Data
Have you ever heard the saying, "A picture is worth a thousand words"? In mathematics, a well-drawn graph is often worth a thousand numbers! Raw data—long lists of figures, survey responses, or measurements—can be overwhelming and difficult to understand. Graphical analysis transforms messy numbers into clear visual stories. Through graphs, we can spot patterns, identify trends, make predictions, and communicate mathematical findings effectively.
In this chapter, you will learn how to read, interpret, analyze, and critically evaluate various statistical graphs. Whether you are aiming for standard or extended mathematics in your IB MYP eAssessment, mastering graphical interpretation will help you excel across Criterion A (Knowing and understanding), Criterion C (Communicating), and Criterion D (Applying mathematics in real-life contexts).
Note: For cumulative frequency curves and box-and-whisker plots, please refer to the dedicated chapter "Cumulative Frequency and Box Plots". For sampling methods, see "Data Collection, Population Sampling and Representation".
1. The Foundation: Types of Data and Choosing the Right Graph
Before analyzing a graph, we must know what type of data is being represented. The type of data determines which graph is mathematically valid.
Categorical (Qualitative) Data: Data sorted into non-numerical categories or labels, such as eye colour, favourite sport, or vehicle brand.
Numerical (Quantitative) Data: Data that consists of numbers that can be measured or counted. This is split into two important types:
• Discrete Data: Data that can only take specific, separate values (often whole numbers obtained by counting), such as the number of goals scored in a match \(x \in \{0, 1, 2, 3, \dots\}\) or shoe sizes.
• Continuous Data: Data resulting from measurements that can take any value within a continuous range or interval, such as time \(t\) in seconds, height \(h\) in centimetres, or mass \(m\) in kilograms.
Key Takeaway: Always check your variables first. Discrete categories use graphs with distinct separations (like bar charts), while continuous measurements use continuous scales (like histograms and line graphs).
2. Frequency Graphs: Bar Charts vs. Histograms
Many students confuse bar charts and histograms. Although they look similar at first glance, they serve very different purposes.
Bar Charts (for Categorical or Discrete Data)
A bar chart displays data using rectangular bars where the length or height of each bar is proportional to the frequency (or percentage) of that category.
• Gaps between bars: There must always be spaces between the bars to show that the categories or discrete values are distinct and separate.
• Horizontal axis: Shows categories (e.g., Apple, Banana, Orange) or discrete scores (e.g., \(1, 2, 3, 4\)).
• Vertical axis: Shows the frequency (how often each item occurs).
Histograms with Equal Class Intervals (for Continuous Data)
A histogram represents continuous data grouped into intervals (known as class intervals or bins).
• No gaps between bars: Because continuous data flows without interruption (like time or distance), the bars touch each other, unless an interval has a frequency of zero.
• Horizontal axis: Represents a continuous numerical number line. For instance, an interval might be written as \(10 \le x < 20\) followed directly by \(20 \le x < 30\).
• Reading intervals: Pay close attention to boundary values! If an interval is \(10 \le x < 20\), a value of \(10\) is included in this bar, but a value of \(20\) belongs to the next bar.
Common Mistake to Avoid: Never leave gaps between the bars of a histogram when dealing with continuous data, and never merge the bars of a bar chart when dealing with distinct categories.
3. Stem-and-Leaf Displays
A stem-and-leaf diagram is an organized way to display discrete or continuous numerical data while preserving every single original data value. It looks like a histogram turned on its side, but with the actual digits visible!
Structure of a Stem-and-Leaf Diagram
• Stem: Represents the leading digit or digits (e.g., tens, hundreds, or whole units).
• Leaf: Represents the trailing digit (usually the final single digit, e.g., units or tenths). Leaves must always be written in ascending order from left to right, evenly spaced.
• The Key: A stem-and-leaf diagram is completely meaningless without a key! For example: Key: \(3 \mid 2 = 32\) marks or Key: \(4 \mid 5 = 4.5\text{ cm}\).
Back-to-Back Stem-and-Leaf Diagrams
When you want to compare two related sets of data (for example, test scores of Class A versus Class B), you can place them on either side of a single central stem:
Class A leaves are written to the left (ordered smallest to largest reading outward from the stem to the left).
Class B leaves are written to the right (ordered smallest to largest reading outward from the stem to the right).
Memory Trick: Think of a tree trunk (the stem) with branches on both sides (the leaves). Always read from the stem outward to avoid reversing your digits!
4. Bivariate Data: Scatter Diagrams and Correlation
When we investigate relationships between two numerical variables simultaneously, we work with bivariate data. We plot these pairs of values as coordinates \((x, y)\) on a scatter diagram (or scatter plot).
• Independent (Explanatory) Variable: Plotted on the horizontal \(x\)-axis. This is the variable being controlled or tested.
• Dependent (Response) Variable: Plotted on the vertical \(y\)-axis. This is the variable being measured to see if it changes in response.
Types of Correlation
Correlation describes the strength and direction of the statistical relationship between two variables:
• Positive Correlation: As \(x\) increases, \(y\) tends to increase. The points trend upwards from bottom-left to top-right. Real-life example: Study time and exam scores.
• Negative Correlation: As \(x\) increases, \(y\) tends to decrease. The points trend downwards from top-left to bottom-right. Real-life example: Outside temperature and heating fuel consumed.
• Zero / No Correlation: There is no apparent relationship between \(x\) and \(y\). The points are scattered randomly across the grid. Real-life example: Shoe size and mathematics ability.
Strength and Form of Correlation
• Strong Correlation: The data points lie very close to an imaginary straight line.
• Weak Correlation: The data points follow a general upward or downward trend, but are widely spread out.
• Linear vs. Non-linear: A relationship is linear if the points follow a straight-line pattern. If the pattern curves (for instance, following an exponential or quadratic curve), it is a non-linear relationship.
Vital Concept: Correlation Does Not Imply Causation!
Just because two variables show a strong correlation does not mean that one causes the other. For example, ice cream sales and sunscreen sales are strongly positively correlated. Does eating ice cream make people buy sunscreen? No! Both are influenced by a third lurking variable: hot summer weather. In Criterion D tasks, always evaluate whether a causal link makes logical sense in real life.
5. The Line of Best Fit and Making Predictions
When bivariate data shows a linear correlation, we can model the relationship by drawing a straight line of best fit (trend line).
How to Draw a Reliable Line of Best Fit by Eye
Step 1: Calculate the mean of the \(x\)-values, denoted as \(\bar{x}\), and the mean of the \(y\)-values, denoted as \(\bar{y}\).
Step 2: Plot the mean point \((\bar{x}, \bar{y})\) on your scatter diagram.
Step 3: Place a transparent ruler so that your straight line passes directly through the mean point \((\bar{x}, \bar{y})\).
Step 4: Adjust the angle so that the line follows the general direction of the data, with roughly an equal balance of points above and below the line along its entire length.
Interpolation vs. Extrapolation
Once your line of best fit is drawn, you can use it to estimate values. Understanding the difference between interpolation and extrapolation is essential for evaluating the validity of your conclusions (Criterion D):
• Interpolation: Making a prediction within the range of the given data points. For example, if your experiment collected data for temperatures between \(15^\circ\text{C}\) and \(35^\circ\text{C}\), estimating the outcome at \(25^\circ\text{C}\) is interpolation. Validity: Generally reliable because the trend has been directly observed across this domain.
• Extrapolation: Making a prediction outside the range of the given data points (either beyond the maximum or below the minimum). For example, predicting the outcome at \(60^\circ\text{C}\) when your data ends at \(35^\circ\text{C}\). Validity: Highly unreliable and speculative! In real-world contexts, patterns can flatten, reverse, or hit physical limits beyond the observed data range.
Key Takeaway: Always trust interpolated values more than extrapolated values. When asked in an exam to comment on the reliability of a prediction, identify whether it was obtained by interpolation or extrapolation.
6. Time Series and Trend Lines
A time series graph is a line graph where the horizontal axis represents time (such as hours, days, months, or years) and the vertical axis represents a measured variable.
• Individual data points are plotted and connected in sequence with straight line segments.
• General Trend: Indicates the long-term upward or downward movement over time, ignoring short-term fluctuations.
• Periodic / Seasonal Variations: Regular, repeating rises and falls that happen at specific intervals (for example, retail sales rising every December, or temperature rising during daytime and falling at night).
7. Critical Analysis: Misleading Graphs and Ethical Representation
In modern media and advertising, graphs are often deliberately manipulated to mislead audiences or exaggerate differences. In MYP mathematics, you are expected to look at graphs with a critical eye, identifying distortions and evaluating validity.
Common Ways Graphs Mislead
• Truncated (Broken) Vertical Axis: Starting the vertical axis at a non-zero value (e.g., starting at \(90\) instead of \(0\)) without clear labelling visually exaggerates tiny differences, making a modest change look enormous.
• Uneven or Inconsistent Scale Intervals: Changing the scale spacing along an axis distorts the visual shape of the trend.
• Pictograms Using 2D or 3D Scaling: If the height of an image or bar is doubled to represent a doubling of value, its area actually quadruples (\(2^2 = 4\)), and its apparent volume increases eight-fold (\(2^3 = 8\)). This dramatically misleads the human eye into perceiving a much larger increase than actually exists.
• Missing Labels, Units, or Keys: Graphs without clearly defined axes or scale units prevent the reader from making accurate, quantitative judgments.
Quick Review: How to Critique a Graph in an eAssessment (Criterion C & D)
When asked to evaluate a graph, check these four points:
1. Does the vertical axis start at zero? If not, does it use a clear break symbol?
2. Are the scales linear and evenly spaced on both axes?
3. Are all labels, units of measurement, and keys clearly provided?
4. Is the visual size of the representation strictly proportional to the numerical data?
8. Worked Example: Interpreting a Bivariate Data Set
Context: A science student investigates the relationship between the mass of fertilizer added, \(x\) (in grams), and the height of a sunflower after six weeks, \(y\) (in centimetres). The data yields a mean point of \((\bar{x}, \bar{y}) = (20, 65)\). The observed fertilizer amounts ranged from \(5\text{ g}\) to \(35\text{ g}\).
Question 1: Describing the Correlation
The points on the scatter plot trend steadily upwards from bottom-left to top-right in a close, linear fashion.
Answer: There is a strong positive linear correlation between the mass of fertilizer and sunflower height.
Question 2: Predicting with Interpolation
Using the line of best fit that passes through \((20, 65)\), the student predicts that a sunflower given \(25\text{ g}\) of fertilizer will reach a height of approximately \(74\text{ cm}\). Comment on the reliability of this prediction.
Answer: The prediction is reliable because \(25\text{ g}\) lies comfortably within the experimental range of \(5\text{ g}\) to \(35\text{ g}\). This is an example of interpolation, supported by a strong correlation.
Question 3: Evaluating Extrapolation
Another student uses the line of best fit to predict that applying \(100\text{ g}\) of fertilizer will produce a sunflower of height \(200\text{ cm}\). Evaluate whether this prediction is sensible in context.
Answer: This prediction is unreliable and invalid. Firstly, \(100\text{ g}\) is far beyond the tested range of \(35\text{ g}\), making it an extrapolation. Secondly, in a real-life biological context, excessive fertilizer causes chemical burns or toxicity, which would stunt growth or kill the plant rather than cause endless linear growth.
Chapter Summary
• Bar charts have gaps between bars and display categorical or discrete data.
• Histograms have touching bars with continuous numerical class intervals on the horizontal axis.
• Stem-and-leaf diagrams display raw data values systematically; back-to-back displays enable direct comparison of two distributions; keys are mandatory.
• Scatter diagrams display bivariate data to investigate relationships: positive, negative, or zero correlation; linear or non-linear patterns.
• A line of best fit drawn by eye should pass through the mean point \((\bar{x}, \bar{y})\) and balance points above and below.
• Interpolation (within data range) is generally reliable; extrapolation (outside data range) is speculative and often invalid.
• Correlation does not imply causation—always evaluate real-life context and confounding variables.
• Watch out for misleading graphs, especially non-zero baselines, inconsistent axes, and disproportionate pictorial scaling.