ATI TEAS Scientific Reasoning: Variables, Controls, Measurement, and Evidence

Scientific Reasoning is investigation work

ATI's Version 7 content outline names four Scientific Reasoning objectives: use basic scientific measurements and measurement tools, apply logic and evidence to a scientific explanation, predict relationships among events, objects, and processes, and apply the scientific method to interpret a scientific investigation. Terms such as "variable" and "hypothesis" matter because they help you do that work, not because definitions are the end goal.

ATI assigns 9 scored items to Scientific Reasoning, out of 44 scored Science items. That is a category total, not a promised count for variables, graphs, experimental design, or any single skill.

Version 7 is still ATI's current exam as of October 2026, with no successor announced, so these four objectives apply to 2027 test dates unless ATI publishes a dated change. The ATI TEAS Science study guide shows why this area deserves time next to the three content domains.

Start with a five-question worksheet

Before you judge any conclusion, answer:

1. What changed? The factor the researcher deliberately varied, or the groups being compared.
2. What was measured? The outcome, with its unit or how it was defined.
3. What was kept the same? Controlled conditions, matched groups, a fair comparison.
4. What is the evidence? The result, pattern, difference, or repeated observation.
5. What remains unknown? Limits, other explanations, and claims the design never tested.

The worksheet separates the investigation from its story. A paragraph may mention many details, but only some define the comparison or support the conclusion.

Variables describe roles in a test

In an experiment, the researcher deliberately changes one or more factors to see the effect on one or more measured responses. The factor deliberately changed is the independent variable. The measured response is the dependent variable. Controlled variables are conditions kept the same so they don't offer a competing explanation.

Those roles depend on the study. Temperature can be the independent variable in one experiment and a controlled condition in another. Name what the researcher did instead of giving a quantity a permanent label.

Controlled variable versus comparison group

These related terms are not interchangeable:

• A controlled variable is a condition held constant, such as water volume or observation time.
• A comparison group gives a baseline or alternative condition to judge results against.
• A control group is a comparison group that gets no treatment, or a standard condition, so the treatment's effect can be seen.

If two plant groups get different fertilizers and also different amounts of light, fertilizer is not the only thing that changed. Light becomes a confounding factor: its effect can't be separated from the fertilizer's.

A hypothesis must connect variables

A testable hypothesis predicts a relationship that observations can check. A useful pattern is:

If the independent variable changes in a stated way, then the dependent variable will change in a stated way, because of a proposed explanation.

For example:

If otherwise comparable seedlings get eight hours of light a day instead of four, then their mean height increase over 14 days will be greater, because they have more time for light-dependent processes.

That hypothesis names a light condition, a measurable height change, a time period, and a reason. Results can support it or fail to support it; they do not "prove" it for every setting.

A question is not a hypothesis. "Does light affect growth?" can guide a study, but it predicts no direction and defines no measurement.

Design determines what a result can mean

A strong basic investigation changes one focal factor, measures the outcome the same way every time, and makes groups comparable enough for the conclusion it wants to draw.

Check these features:

• Groups start under comparable conditions.
• The independent variable has defined levels.
• The dependent variable is measured the same way in every group.
• Important outside factors are controlled or recorded.
• There are enough subjects and trials for the claim.
• The conclusion stays within the organisms, conditions, and time studied.

Random assignment keeps the researcher's choices or the order of runs from lining up with the treatment, which is why statisticians treat it as necessary for defensible conclusions. Replication, running the same condition more than once, shows whether a result is consistent. Neither one repairs an outcome that was measured inconsistently. When design questions are where you lose points, practice Scientific Reasoning questions on experiment design in full ATI TEAS prep → and fill in the worksheet first.

Read measurements with their tools and units

ATI names basic measurements and measurement tools as their own objective. Match each property to a tool and a metric unit:

A reported measurement should match the instrument: you record every digit the markings give you plus one estimated digit, not a string of unsupported decimals.

Two terms are easy to confuse. Accuracy is how close a result is to the true value of what you measured. Precision is how closely repeated results agree with one another. Suppose three mass readings are 15.1 g, 15.1 g, and 15.2 g. They agree closely, so they are precise. Whether they are accurate depends on the true mass or a trusted standard, which you haven't been given.

Consistent results across repeated trials also support reliability, the idea ATI uses in a sample question in its public outline. Consistency alone doesn't make a measurement accurate: a balance that is off by a gram can give the same wrong reading every time.

Always carry units into tables and conclusions. A value of 12 is incomplete if it could mean seconds, centimeters, or grams.

Work through an original investigation

A student tests whether light color is associated with seedling growth. The student randomly assigns 24 similar seedlings to two boxes, 12 per box. One box gets blue light and the other red light, for the same hours each day. Plant type, soil, water, temperature, starting height, and study length are kept comparable. Height change is measured in centimeters after 14 days.

Apply the worksheet:

• What changed? Light color between the two groups.
• What was measured? Mean height increase in centimeters after 14 days.
• What was kept the same? The listed plant and environmental conditions.
• What is the evidence? The blue group's mean increase was 0.9 cm greater (5.0 − 4.1) under these conditions.
• What remains unknown? The spread within each group, other species, times beyond 14 days, and the mechanism. Each color was also tested in a single box, so a difference between the boxes themselves could contribute.

A justified conclusion: in this investigation, seedlings under blue light had a greater mean height increase than seedlings under red light. "Blue light makes every plant grow faster" goes far beyond the evidence. If you keep picking the broader claim, revisit missed variable and evidence questions with Fix It in full ATI TEAS prep →.

Improve the design before extending the claim

Useful next evidence would include individual measurements, repeated trials with more than one box per color, or more light colors. An average alone can hide whether all seedlings responded similarly or one unusual plant shifted the result.

Changing both color and daily light hours in the next trial would weaken the comparison, because two factors would vary at once. Fix one limitation at a time.

Evidence sets the size of the conclusion

Match the wording of a conclusion to what the design can show:

• Difference: one group had a higher or lower measured outcome under the tested conditions.
• Association: two variables changed together, but the design may not show cause.
• Causal support: a controlled change produced a difference while other explanations were limited.
• No demonstrated difference: the data didn't show a difference. That is not always proof that none exists.

Imagine a survey in which students who report more weekly study time also report higher quiz scores. That is an association. Nobody assigned study time, so prior knowledge, course access, sleep, or reporting accuracy could explain part of the pattern.

Evidence also has direction. A result can support "A was greater than B" without explaining why. Keep the observed pattern separate from the proposed mechanism.

Predict from a stated relationship

A prediction applies a pattern to a new condition. First identify the direction and range of the relationship.

If a graph shows reaction rate rising as temperature increases from 10 °C to 30 °C, predicting a higher rate at 25 °C than at 15 °C stays inside the data. Predicting a continued rise at 100 °C goes far beyond it and may ignore a change in the system.

• Interpolation estimates within the observed range.
• Extrapolation extends beyond it, and statisticians treat it as potentially dangerous whatever model is used.

A trend can also curve. Don't assume equal steps in one variable always give equal steps in the other unless the data or model shows it.

Read data before the explanation

Read tables and graphs in a fixed order:

1. Read the title and identify the scientific question.
2. Read both axes or column headings, the units, and the legend.
3. Find the comparison the question asks about.
4. Describe the visible pattern without explaining it.
5. Decide which conclusion the data supports.
6. Check whether the proposed explanation goes beyond the design.

That order stops an appealing theory from replacing the measured result. If the evidence contradicts the prediction, report the evidence accurately and rethink the hypothesis or design. The ATI TEAS graphs and data guide covers reading scales and legends in more depth, and the free ATI TEAS practice questions give you a mixed sample to try.

Finish every investigation with an unknown

A careful reader can name the strongest supported statement and one meaningful unknown: a mechanism, a population outside the sample, an unmeasured confounder, or whether the pattern repeats.

Take one graph or experiment and complete the five-question worksheet before you look at any answer choices. Then work mixed Science sets that include investigation questions in full ATI TEAS prep →, especially when a conclusion sounds broader than the data.