Start by Naming the Difference
In “Matched Pairs Versus Independent Samples,” you learned that a paired comparison focuses on the two linked measurements within each pair. The next step is to say exactly how those measurements will be compared. Here we use the convention \(d=\text{after}-\text{before}\). That choice gives every difference a consistent direction: a positive value means the measurement was larger after, and a negative value means it was smaller after.
The subtraction order is not a minor detail. If you reverse it, every difference changes sign, and the interpretation changes from “after compared with before” to “before compared with after.” Either order can be useful if it is clearly stated, but in this tutorial we define the difference as after minus before. State that convention before calculating or describing the differences.
The subscript \(i\) identifies a particular pair. For example, \(d_1\) is the difference for the first student, patient, device, or other unit; \(d_2\) is the difference for the second. The value \(d_i\) describes that pair, not the entire population. In later work, the collection of these differences will be used to make a paired comparison. For now, the priority is to define and interpret each difference correctly.
Keep the Direction Visible
A reliable way to avoid sign errors is to write the subtraction in words first: after minus before. Then substitute the two measurements in that order. Do not decide the sign based on which number is bigger or on whether the result seems desirable. The sign records the direction of numerical change, not whether the change is good or bad.
Always include the measurement’s units when reporting a difference. If a student’s score is recorded in points, the difference is in points. If a machine’s operating time is recorded in hours, the difference is in hours. Subtracting two measurements with the same units leaves a difference in those same units.
A positive difference does not automatically mean an improvement. An increase in a test score might be welcome, while an increase in a pain rating might not be. The study’s context tells you what the variable measures and whether an increase is desirable; the sign alone tells you only whether the after value is greater or less than the before value.
Worked Examples
Worked Example: Tutoring Scores Before and After
A fictional tutoring program records a quiz score, in points, for each student before and after several weeks of tutoring. Define each difference as after minus before, calculate it for each student, and interpret the signs.
| Student | Before tutoring (points) | After tutoring (points) | After minus before (points) |
|---|---|---|---|
| A | 62 | 71 | ? |
| B | 75 | 79 | ? |
| C | 81 | 77 | ? |
| D | 68 | 68 | ? |
Solution. For every student, subtract the before score from the after score. For student A, \(d_A=71-62=9\) points. The positive sign means the score was 9 points higher after tutoring. For student B, \(d_B=79-75=4\) points, also an increase.
For student C, \(d_C=77-81=-4\) points. The negative sign means the score was 4 points lower after tutoring. For student D, \(d_D=68-68=0\) points, so the recorded score was unchanged. The differences, in student order, are 9, 4, \(-4\), and 0 points.
Notice that “positive” does not mean every student improved: it describes only the pairs with positive differences. Also, the negative difference for student C is not an arithmetic mistake. Because the definition is after minus before, a lower after score produces a negative result.
Worked Example: A Decrease in a Health Rating
In a fictional clinic exercise, each participant rates their discomfort on a scale from 0 to 10 before and after a guided relaxation session. Larger ratings indicate more discomfort. For one participant, the rating is 7 before the session and 3 afterward. Define the difference as after minus before.
Solution. Let \(d\) be this participant’s after-minus-before difference in discomfort rating. Substitute the after rating first and the before rating second:
The difference is negative because the after rating is lower than the before rating. In context, this participant’s discomfort rating decreased by 4 points after the session. The negative sign follows from the stated subtraction order; it does not mean the participant experienced a negative amount of discomfort.
If someone instead calculated before minus after, the result would be \(7-3=4\) points. That number also describes a 4-point decrease when properly explained, but it does not follow the convention used here. Mixing the two subtraction orders within one set of pairs would make the signs difficult to interpret and would not produce a consistent difference variable.
Worked Example: Sleep Duration After a Schedule Change
A fictional group of workers records sleep duration, in hours, before and after changing to a new schedule. Use \(d=\text{after}-\text{before}\), calculate each person’s difference, and describe what the sign indicates. Do not assume that the schedule change caused the observed differences.
| Worker | Before (hours) | After (hours) | After minus before (hours) |
|---|---|---|---|
| 1 | 6.5 | 7.0 | ? |
| 2 | 8.0 | 7.5 | ? |
| 3 | 7.0 | 7.0 | ? |
Solution. For worker 1, \(d_1=7.0-6.5=0.5\) hour, so the recorded sleep duration was half an hour higher after the schedule change. For worker 2, \(d_2=7.5-8.0=-0.5\) hour, so the recorded duration was half an hour lower afterward. For worker 3, \(d_3=7.0-7.0=0\) hours, so there was no recorded change. The differences are \(0.5\), \(-0.5\), and 0 hours.
Each value is a within-worker comparison: the two measurements belong to the same person. The difference is not “the first worker’s after value minus the second worker’s before value.” Pair the measurements using the study’s design, then subtract within each pair. As discussed in “What Makes Data Paired,” the link between observations comes from the design, not from their row positions in a table.
These calculations describe the observed changes only. They do not establish that the schedule change caused them. Causal conclusions depend on the study design, including whether there was random assignment; as covered in “Random Assignment Versus Random Sampling Conditions,” that is a separate question from defining the difference.
What the Sign Does—and Does Not—Tell You
Once the order is fixed, the sign gives a direct comparison for each pair. A positive \(d_i\) indicates that the after measurement exceeds the before measurement. A negative \(d_i\) indicates that the after measurement is less than the before measurement. A zero indicates equal recorded measurements. These statements remain true regardless of the variable’s context.
The sign does not tell you how important a change is. A difference of 0.1 hour and a difference of 5 hours are both positive, but their sizes are not the same. The sign also does not tell you whether the change is typical of a larger population or whether it is statistically convincing. Those questions require additional summaries and inference beyond defining the difference variable.
A difference is measured on the same scale as the original variable, but it represents a change between two measurements. For example, if quiz scores are measured in points, a difference of 9 means a 9-point change, not a score of 9. Be careful not to confuse the original measurements with the result of subtracting them.
For matched pairs that are not literally measured before and after, the same principle applies: name which member of the pair is subtracted from which. You might define treatment minus comparison, for example, if that direction suits the question. But it is essential to state the order because “the difference” alone does not reveal which value was subtracted. Here, whenever the measurements are described as before and after, use the explicit convention after minus before.
Common Mistakes and AP Exam Tips
- Reversing the subtraction order. Writing before minus after while claiming to use after minus before reverses every sign. Show the order in words and use it consistently.
- Reporting only “the difference was 4.” A full interpretation includes the direction, units, and context: “The after score was 4 points higher than the before score.”
- Treating a negative difference as an impossible measurement. The original measurements may be nonnegative while their difference is negative. A negative value simply means after was lower than before under this convention.
- Calling every positive difference an improvement. Positive means numerically higher after. Whether that is beneficial depends on what the variable measures.
- Subtracting observations from different pairs. For paired data, match each before measurement with its own after measurement. Arbitrary cross-pair subtraction does not describe a within-pair change.
- Confusing a difference with an original value. A change of 3 points is not a final score of 3 points. Label the difference and its units clearly.
- Using the sign to make an inference claim. A positive observed difference does not by itself prove a population increase or show that a change is statistically significant. Those conclusions require the appropriate analysis and evidence.
For a clear AP response, define the subtraction order, state the unit, and translate the sign into a sentence about the measured variable. For example: “For each student, \(d=\text{after score}-\text{before score}\), measured in points. A positive \(d\) means the student’s score was higher after tutoring.” That definition makes the direction unambiguous before any further analysis.
Check Your Understanding
For each item, use the convention \(d=\text{after}-\text{before}\), and explain the sign in context.
- A runner’s recorded time for a route is 18 minutes before a training plan and 16 minutes afterward. What is \(d\), in minutes, and what does its sign indicate?
- A plant is 12 centimeters tall before a light change and 15 centimeters tall afterward. Calculate the difference and interpret it.
- A customer gives a product a satisfaction rating of 8 before a redesign and 8 afterward. What is the difference, and what does zero mean here?
- For a pair of measurements, before is 24 units and after is 19 units. Is the after-minus-before difference positive or negative? State its value and direction.
- Why must a response state the subtraction order instead of simply saying “calculate the differences”?