ACT Science rewards careful interpretation: identify what a question asks, locate the relevant evidence, and decide what that evidence supports. Memorizing more facts can help with a genuine background gap, but it will not replace reading the right graph, tracking an experiment’s variables, or distinguishing two scientists’ claims.
When Science is included in an enhanced ACT administration, it has 40 questions in 40 minutes. It does not contribute to the English–Math–Reading Composite. A Science score does contribute to a STEM score with Math. Whether you can choose Science yourself depends on your testing program; a school-day contract may determine the choice. Check the official ACT structure before practicing.
The strategy below focuses on the work a Science question requires. All examples are original teaching examples, not official ACT questions.
Start by identifying the evidence you need
A Science passage can feel difficult because it introduces an unfamiliar topic. The question may still require a straightforward comparison between two numbers.
Before reading every detail, classify the question:
| Question asks for… | Look first at… | Main danger |
|---|---|---|
| A recorded value | The named row, column, point, or bar | Reading the adjacent condition |
| A trend | Several values across the relevant range | Describing one point as a trend |
| An experimental comparison | Procedures and variables in both trials | Missing a second changed variable |
| A prediction | The stated relationship or model | Extending a pattern beyond the evidence |
| Agreement or disagreement | Each explanation’s actual claim | Importing your own opinion |
This is a navigation aid, not a rule to ignore the passage. If a symbol, procedure, or condition is unclear, read the material that defines it before deciding.
ACT describes Science as covering interpretation, scientific investigation, and evaluation of arguments and models. Its topics include biology, chemistry, physics, and Earth and space sciences. General introductory knowledge may be needed, but advanced knowledge is not required. That is more precise than saying “Science contains no science.” See the official Science description.
Read tables with the condition attached
Consider this original experiment. A student grows the same variety of bean plant in identical pots. Each group receives the same light, water, and soil. The student changes the amount of fertilizer and records average growth after 14 days.
| Fertilizer per pot | Average growth |
|---|---|
| 0 grams | 6 centimeters |
| 5 grams | 10 centimeters |
| 10 grams | 11 centimeters |
The data value is not simply “10.” It is 10 centimeters of average growth for plants receiving 5 grams of fertilizer after 14 days. Keeping the condition attached prevents many lookup errors.
Example 1: Compare two recorded results
How much greater was average growth in the 5-gram group than in the 0-gram group?
Subtract the recorded growth values: 10 − 6 = 4 centimeters. Do not subtract the fertilizer amounts. The question asks about the measured outcome, not the treatment.
Example 2: Describe the pattern accurately
Average growth increases as fertilizer increases across these three tested amounts. However, the increase becomes smaller: adding the first 5 grams corresponds to 4 additional centimeters, while adding the next 5 grams corresponds to 1 additional centimeter.
“Increasing fertilizer always produces the same increase in growth” is unsupported. So is “fertilizer has no effect,” because that ignores the observed differences. A precise answer describes the tested pattern without claiming more than the table shows.
Example 3: Know the limit of a prediction
The table does not establish what happens at 30 grams. Growth might increase, level off, or decrease. A question could ask for a prediction based on a stated model, but the model and its assumptions would matter.
Likewise, an exact value at 7 grams is not directly recorded. If a question explicitly tells you to assume a straight-line relationship between 5 and 10 grams, interpolation is reasonable. Without that instruction, do not quietly turn a possible estimate into an observed fact.
Separate the variable changed from the variable measured
For any experiment, make a compact mental summary:
- Changed: what the researcher deliberately varies.
- Measured: the outcome recorded.
- Held constant: conditions intended to remain the same.
- Compared: the groups or trials that answer the question.
In the plant experiment, fertilizer amount is changed and growth is measured. The 0-gram group provides a comparison without added fertilizer.
Now imagine the 10-gram group also receives twice as much water. The groups differ in two relevant ways, so the growth difference cannot be attributed to fertilizer alone from that comparison. The problem is not that the numbers become unusable; it is that the proposed causal explanation is less specific than the design can support.
Replication matters for a different reason. Measuring several plants in each condition can show whether results are consistent and reduce reliance on one unusual plant. Repetition does not automatically fix a design that changes fertilizer and water together.
When two experiments appear similar, compare their procedures before treating the results as interchangeable. Different temperatures, durations, instruments, or definitions of the measured outcome may explain why the tables differ.
Distinguish graph height from graph change
On a graph, the highest line is not necessarily the fastest-changing line.
Suppose Sample A rises from 20 to 24 units while Sample B rises from 5 to 15 units over the same time interval. Sample A ends higher, but Sample B increases more: 10 units compared with 4. If the question asks about greatest increase, the final heights alone are insufficient.
Use this brief graph routine:
- Read the horizontal and vertical labels, including units.
- Check the legend for the correct group.
- Identify the interval named in the question.
- Decide whether you need a value, a difference, or a rate of change.
- Compare the relevant evidence.
Watch scale spacing. A vertical axis that starts above zero may make a small difference look large. A display with unequal or logarithmic spacing requires you to interpret the actual labels rather than count squares mechanically.
These habits also help with units. A temperature in degrees and a change in temperature may use similar-looking numbers but answer different questions.
Compare competing explanations without choosing a favorite
Some passages present explanations that disagree. Your task may be to predict what each explanation implies, not to decide which sounds most convincing in everyday life.
Consider this original example:
- Explanation A: A pond’s seasonal algae increase is mainly caused by warmer water.
- Explanation B: The increase is mainly caused by additional nutrients entering the pond after storms.
A question asks which experiment most directly tests whether nutrients can increase algae when temperature is unchanged.
A useful design holds temperature constant and varies nutrient levels across otherwise comparable samples. That design isolates the factor highlighted by Explanation B. Raising both temperature and nutrients would make the interpretation harder.
This does not prove that nutrients explain every seasonal change in the actual pond. A controlled result and a universal explanation are different claims.
For these passages, make a small comparison:
| Feature | Explanation A | Explanation B |
|---|---|---|
| Main proposed driver | Temperature | Nutrient input |
| Useful measurement | Water temperature | Nutrient concentration |
| Prediction to examine | More algae under warmer conditions | More algae with added nutrients |
Add qualifications if the passage includes them. A model that says “only above a threshold” should not be summarized as “always.” Many wrong answers become attractive only after a reader drops a condition.
Use background knowledge to understand, then answer from evidence
A familiar term can save time. Knowing what a control group is or how mass differs from volume helps you read efficiently. But familiar subject matter can also tempt you to answer a different question.
If a passage defines an unusual hypothetical organism, use that definition. If a question says “according to Experiment 2,” locate Experiment 2. Do not substitute a general fact from class for a specific result the question asks you to interpret.
After practice, separate two problems: “I did not know what concentration meant” and “I read the concentration column when the question asked for temperature.” The first suggests a short concept review. The second calls for a reading and checking routine.
Calculators are not permitted on ACT Science. Practice simple differences, ratios, and comparisons without reaching for one, consistent with the ACT calculator policy.
Build pacing around completed evidence checks
Forty questions in 40 minutes gives an average of one minute per question, but individual questions need different amounts of work. Do not treat that average as a command to stop every question at exactly 60 seconds.
A productive first pass resolves clear lookups and comparisons while keeping track of questions that need a longer model or procedure analysis. If permitted by your test format, mark an uncertain item and return within the Science section.
Before changing an answer, identify the new evidence. “This option sounds more scientific” is not a reason. “The question names Trial 3, and I initially used Trial 2” is.
Use current enhanced materials for timing. An older form can still offer interpretation practice, but its length should not quietly become your current pacing benchmark.
Review the error that actually happened
A useful Science review entry contains the question’s demand, the evidence needed, the mistake, and a specific repair.
| Error | Repair to practice |
|---|---|
| Wrong data series | Name the legend label before comparing |
| Confused outcome and treatment | Write “changed” and “measured” beside the table |
| Unsupported extrapolation | Separate recorded values from predictions |
| Mixed two explanations | Build a one-line claim for each |
| Missed a condition | Restate the claim with its qualifier |
Then solve a fresh example. Remembering the correct answer to an old item is not enough to show that the repair transfers.
For one practice week, combine a short untimed data set, a focused experimental-design review, and a timed Science section followed by careful correction. Adjust the balance according to your errors. The goal is a repeatable way to find and evaluate evidence, not a larger pile of completed pages.
Use the ACT calculator linked below only with its supported form and counting mode when estimating a practice score. SHSPrep’s ACT program includes lessons, targeted practice, and two original full-length mock tests with raw performance results; these guides and the calculator can support your planning alongside lessons and targeted practice.