Identify variables and experimental design flaws. - GED Science

SECTION 1

Why Experimental Design Matters

For centuries, people relied on tradition, authority, and guesswork to explain the natural world. A doctor might prescribe a remedy because it had "always been used," without ever testing whether it actually worked. The development of the scientific method — with its emphasis on careful, controlled experiments — changed everything. Understanding how experiments are designed, and recognizing when that design is flawed, is one of the most important skills tested on the GED Science exam.

1025 • Ibn al-Haytham's Optics

The Arab scholar Ibn al-Haytham insisted that claims about light must be tested through repeatable experiments, laying early groundwork for controlled investigation.

1747 • Lind's Scurvy Trial

James Lind divided sailors into groups, giving each a different treatment for scurvy. By comparing outcomes across groups, he demonstrated that citrus fruits cured the disease — one of the first controlled experiments in medicine.

1882 • Koch's Postulates

Robert Koch established strict rules for proving that a specific germ causes a specific disease, demanding controlled testing with isolated variables.

1948 • First Randomized Clinical Trial

The British Medical Research Council tested streptomycin against tuberculosis using random assignment and control groups, setting the modern gold standard for experimental design.

Each of these milestones moved science toward the same goal: making sure that when we observe an effect, we can be confident about what caused it. On the GED, you will be given descriptions of experiments and asked to identify the variables involved and to spot any flaws in how the experiment was set up. Let's build that skill step by step.

SECTION 2

Core Principles: Variables and Controls

Every experiment revolves around variables — factors that can change and potentially affect the outcome. Understanding the different types of variables is the foundation for evaluating any experiment. There are three main categories you need to know, plus the concept of a control group.

Independent Variable

The factor the scientist deliberately changes or manipulates. Think of it as the "cause" being tested. There is usually only one independent variable per experiment. Example: the amount of fertilizer given to plants.

Dependent Variable

The factor that is measured or observed as a result. Think of it as the "effect." It depends on what happens with the independent variable. Example: the height the plants grow.

Controlled Variables (Constants)

All other factors that the scientist keeps the same across all groups so they do not interfere with results. Examples: same type of soil, same amount of water, same sunlight exposure.

Control Group

A group that receives no treatment or receives the standard treatment. It provides a baseline for comparison so the scientist can see if the independent variable actually made a difference.

✦

KEY TAKEAWAY

Think of an experiment like a recipe test. If you want to know whether adding extra sugar makes a cake taste better, the extra sugar is the independent variable (what you change), the taste rating is the dependent variable (what you measure), and the oven temperature, baking time, and flour amount are controlled variables (what you keep the same). A cake made with the original recipe is your control group.

SECTION 3

Visualizing Variables in an Experiment

This diagram shows how a well-designed experiment separates the independent variable (what is changed, in cyan) from the dependent variable (what is measured, in pink), while keeping all other controlled variables (amber) the same. The control group (green) receives no treatment and provides a baseline.

Notice in the diagram that the experiment has three groups receiving different amounts of fertilizer. Group A, with zero grams of fertilizer, serves as the control group. If Group C's plants grow taller than Group A's, the scientist can attribute that difference to the fertilizer — but only if all other factors were held constant. This is why controlled variables matter so much. If Group C also got more sunlight, the scientist could never be sure whether the extra growth came from the fertilizer or the sunlight.

SECTION 4

How to Identify Variables in Any Experiment

On the GED, you will read a passage describing an experiment, and you need to quickly identify the variables. Here is a reliable strategy that works every time, broken into three questions you can ask yourself.

The Three-Question Strategy

Question 1: What is the researcher trying to find out? The answer to this question often reveals both the independent and dependent variables. Look for phrases like "the effect of X on Y" — X is the independent variable, Y is the dependent variable.
Question 2: What did the researcher deliberately change between groups? This confirms the independent variable. Look for differences in treatment, dosage, temperature, or conditions across the experimental groups.
Question 3: What did the researcher measure or record as data? This confirms the dependent variable. Look for numbers being collected — heights, weights, times, counts, temperatures, survey scores.

💡

GED Test Tip

The GED often phrases variable questions as: "What is the independent variable in this study?" or "Which factor did the researcher manipulate?" These mean the same thing. Similarly, "What was measured?" and "What is the dependent variable?" are identical questions.

Everything that is not the independent or dependent variable — and that the researcher kept the same across all groups — is a controlled variable. On the GED, you may also be asked which variables should have been controlled but were not. That leads us to the next critical topic: design flaws.

SECTION 5

Common Experimental Design Flaws

A perfectly designed experiment isolates one independent variable while keeping everything else the same. In reality, experiments often have flaws that weaken their conclusions. The GED frequently asks you to identify these flaws. Here are the most common ones you will encounter.

The six most common experimental design flaws tested on the GED, with examples. A well-designed experiment (green, bottom) avoids all of these problems.

The flaw that appears most frequently on the GED is the confounding variable — an uncontrolled factor that changes alongside the independent variable, making it impossible to determine the true cause. Whenever you see that two or more things changed between groups, you have found a confounding variable. The second most common flaw is the absence of a control group. Without a control, there is no baseline for comparison, and the experiment cannot show whether the treatment actually caused the observed effect.

SECTION 6

Worked Example: Analyzing an Experiment

ℹ

Experiment Description

A biologist wants to know if a new type of fish food increases the growth rate of goldfish. She places 10 goldfish in a tank with the new food and 10 goldfish in a different, smaller tank with regular food. After 8 weeks, she measures the length of each fish. The fish in the new-food tank grew an average of 2.1 cm, while the fish in the regular-food tank grew an average of 1.4 cm. The biologist concludes that the new food increases growth rate.

Step-by-Step Analysis

1
Step 1 — Identify the Independent Variable

Ask: What did the researcher deliberately change between the two groups? One group received the new fish food, and the other received regular food. So the independent variable is the type of fish food.

Independent variable: type of fish food (new vs. regular)
2
Step 2 — Identify the Dependent Variable

Ask: What did the researcher measure? She measured the length of each fish after 8 weeks. So the dependent variable is the growth in fish length.

Dependent variable: fish growth (cm)
3
Step 3 — Identify the Control Group

The group receiving the regular food serves as the control group because it represents the standard or baseline condition.

Control group: 10 goldfish with regular food
4
Step 4 — Check for Controlled Variables

What should have been kept the same? Tank size, water temperature, light exposure, number of fish, species of fish, and feeding schedule should all be identical across groups.

Controlled variables should include: tank size, water temperature, light, feeding schedule
5
Step 5 — Identify the Design Flaw

Read the description carefully. The new-food group was in one tank, and the regular-food group was in a "different, smaller tank." This is a major problem. The tank size is a confounding variable — the fish in the smaller tank may have grown less because they were crowded, not because their food was different. The biologist changed two things at once (food type AND tank size), so she cannot be certain which factor caused the difference in growth.

Flaw: Tank size is a confounding variable. Both groups should have been kept in identical tanks.

SECTION 7

Strong vs. Weak Experimental Designs

On the GED, you may be asked to compare two experimental setups and determine which one is better designed. The table below summarizes the features that separate strong designs from weak ones.

Feature	Strong Design	Weak Design
Control group	Includes a group that receives no treatment or a placebo	All subjects receive the treatment; no baseline for comparison
Variables controlled	Only one factor differs between groups; all else is identical	Multiple factors differ between groups (confounding variables)
Sample size	Large enough to reduce the effect of individual variation	Too few subjects; results may be due to chance
Random assignment	Subjects randomly placed into groups to avoid bias	Subjects self-select or are assigned by a non-random method
Repeated trials	Experiment repeated multiple times to confirm results	Conducted only once; results may be a fluke
Measurement	Objective, quantifiable data collected (numbers, measurements)	Subjective or vague observations ("seemed better")

Features of strong versus weak experimental designs

✦

KEY TAKEAWAY

Think of a well-designed experiment like a fair race. Every runner should start at the same line, wear similar shoes, and run on the same track. The only difference is the runner's training (the independent variable). If one runner gets a head start or runs on a smoother surface, the race is unfair — and you cannot trust the results. Similarly, an experiment with a confounding variable is like an unfair race: the conclusion is unreliable.

SECTION 8

GED Test Strategies for Design Questions

The GED Science test frequently presents experimental scenarios in passages, data tables, or diagrams and then asks you to analyze the design. These questions can appear in multiple-choice, drag-and-drop, or short-answer format. Here is how to approach them efficiently.

Question Type	What to Look For
"Identify the independent/dependent variable"	Find what was changed (independent) and what was measured (dependent). Use the three-question strategy from Section 4.
"What should be the control group?"	Look for the group that receives no treatment or the standard treatment. It provides the baseline.
"What is a flaw in this experiment?"	Check for confounding variables (more than one thing changed), missing control group, small sample size, lack of random assignment, or no repeated trials.
"How could the experiment be improved?"	The answer usually involves fixing the flaw: add a control group, increase sample size, randomly assign subjects, control a confounding variable, or repeat the experiment.
"Does the data support the conclusion?"	Even if the data looks convincing, the conclusion is NOT supported if the experiment has a design flaw that undermines it.

Common GED question types about experimental design

⚠

Short-Answer Strategy

For the GED's written-response questions, you may be asked to evaluate an experiment in 3–7 sentences. A strong response follows this pattern: (1) State the flaw or identify the variable. (2) Explain why it matters by connecting it to the experiment. (3) Suggest how to fix it. Use specific details from the passage — the GED scores you on evidence-based reasoning, not length.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A researcher wants to test whether listening to classical music improves test scores. She has 40 students take a math test. Group 1 listens to classical music while studying for two hours. Group 2 studies for two hours in silence. Both groups take the same test the next day. What is the independent variable in this experiment?

A. The math test scores
B. Whether or not classical music was played during studying
C. The amount of time spent studying
D. The number of students in each group

Show answer (B)

The correct answer is B. The independent variable is the factor the researcher deliberately changed between the two groups — in this case, whether classical music was played. Choice A (test scores) is the dependent variable, because it is what was measured as a result. Choice C (study time) is a controlled variable — both groups studied for the same two hours. Choice D (number of students) is also a controlled variable, not something being tested.

Question 2: BASIC CALCULATION

A scientist tests a new cough medicine. She gives the medicine to 50 patients and records how long their coughs last. She reports that the average cough lasted 4 days, compared to the typical 7 days. What is the biggest flaw in this experimental design?

A. The sample size is too small
B. The dependent variable was not measured
C. There is no control group receiving a placebo
D. The independent variable was not identified

Show answer (C)

The correct answer is C. The scientist did not include a control group — a group of patients with coughs who received either no treatment or a placebo. Without a control group, the "typical 7 days" is just a general estimate, and there is no way to confirm the medicine actually shortened the cough. Choice A is incorrect because 50 patients is a reasonable sample size. Choice B is wrong because the dependent variable (cough duration) was indeed measured. Choice D is wrong because the independent variable (the cough medicine) is clearly present, even though the study lacked a proper comparison.

Question 3: INTERMEDIATE

A teacher wants to know whether using flashcards improves vocabulary test scores. She assigns her morning class (25 students) to use flashcards for one week. Her afternoon class (25 students) studies vocabulary without flashcards. Both classes take the same test. The morning class scores an average of 88%, while the afternoon class scores 79%. Which of the following is a confounding variable that weakens the teacher's conclusion?

A. Both classes took the same vocabulary test
B. The morning and afternoon classes may have different types of students
C. The teacher measured test scores as the dependent variable
D. Each class had 25 students

Show answer (B)

The correct answer is B. The morning class and afternoon class may differ in important ways: for example, students who take morning classes may be more motivated, more rested, or higher-performing in general. Because students were not randomly assigned to groups, the difference in class composition is a confounding variable that could explain the higher scores, rather than the flashcards themselves. Choice A describes a controlled variable (same test), which is actually good design. Choice C describes the dependent variable, which is correctly chosen. Choice D describes equal sample sizes, which is also good practice.

Question 4: APPLIED

2 pts

A farmer wants to test whether a new organic pesticide reduces the number of insects on tomato plants. He sprays the pesticide on all 20 of his tomato plants and counts the insects after two weeks. He finds that there are fewer insects than he remembers from last year. In 3–5 sentences, identify two flaws in the farmer's experimental design and explain how each flaw weakens his conclusion.

Show answer

Rubric (2 points total): Point 1 (1 point): The student identifies the lack of a control group. The farmer sprayed all 20 plants with the pesticide, so there are no untreated plants to compare against. Without a control group, the farmer cannot determine whether the reduction in insects is due to the pesticide or to other factors such as weather changes, natural seasonal variation, or fewer insects in the area that year. Point 2 (1 point): The student identifies the use of memory-based comparison as a flaw. The farmer is comparing his current results to what he "remembers from last year," which is not a reliable or objective measurement. Conditions between years (temperature, rainfall, crop health) are different, so last year's insect count is not a valid baseline. A proper experiment would need a simultaneous control group of unsprayed plants measured at the same time. Sample strong response: The farmer's experiment has two major flaws. First, he has no control group because he sprayed all 20 plants, leaving no untreated plants for comparison. This means the decrease in insects might be due to weather, season, or other factors rather than the pesticide. Second, he is relying on his memory of last year's insect count instead of collecting objective data from a comparison group this year. Conditions vary from year to year, so his memory is not a reliable baseline.

Question 5: CRITICAL THINKING

3 pts

A pharmaceutical company tests a new pain reliever. They recruit 200 volunteers and randomly assign them to two groups of 100. Group A receives the new pain reliever. Group B receives a sugar pill (placebo). Neither the volunteers nor the doctors administering the pills know which group receives which pill. After one week, volunteers rate their pain on a 1–10 scale. Results: • Group A (new drug): average pain rating = 3.2 • Group B (placebo): average pain rating = 4.8 In 5–7 sentences, evaluate the design of this experiment. Identify at least two features that make it well-designed. Then explain whether the data supports the conclusion that the new drug reduces pain, and identify one additional step the company should take to strengthen their findings.

Show answer

Rubric (3 points total): Point 1 (1 point): The student identifies at least two strengths of the experimental design. Valid strengths include: (a) the use of a control group receiving a placebo, which provides a baseline for comparison; (b) random assignment of volunteers to groups, which reduces selection bias; (c) double-blinding (neither volunteers nor doctors know who gets the drug), which eliminates both placebo effect and researcher bias; (d) a large sample size (100 per group), which increases reliability. Point 2 (1 point): The student correctly evaluates whether the data supports the conclusion. The data does support the conclusion that the new drug reduces pain: Group A's average pain rating (3.2) is lower than Group B's (4.8). Because the experiment used random assignment, a control group, and double-blinding, confounding variables have been minimized, making it reasonable to attribute the difference to the drug. Point 3 (1 point): The student identifies a reasonable additional step. Acceptable answers include: (a) repeat the experiment (replication) to confirm the results are consistent; (b) conduct a longer study to check for lasting effects or side effects; (c) test with a more diverse population (different ages, genders, medical conditions); (d) perform statistical analysis to determine whether the 1.6-point difference is statistically significant (not just due to chance). Sample strong response: This experiment is well-designed in several important ways. First, it uses a control group that receives a placebo, providing a clear baseline for comparison. Second, volunteers were randomly assigned to groups, which prevents selection bias. Third, the study is double-blind, meaning neither the patients nor the doctors know who received the real drug, which eliminates the placebo effect and researcher bias. The data supports the conclusion that the drug reduces pain because Group A's average pain rating of 3.2 is 1.6 points lower than Group B's rating of 4.8, and the controlled design makes confounding variables unlikely. However, to strengthen these findings, the company should repeat the experiment with a different group of volunteers to confirm that the results are reproducible and not due to chance. They should also perform a statistical significance test to verify the 1.6-point difference is meaningful.

SUMMARY

Lesson Summary

Summary

Every experiment is built around three types of variables: the independent variable (what the scientist changes), the dependent variable (what is measured), and controlled variables (what is kept the same). A control group receives no treatment or the standard treatment, providing a baseline for comparison.

The most common experimental design flaws on the GED include confounding variables (more than one factor changed between groups), missing control groups, small sample sizes, lack of random assignment, and no repeated trials. When you encounter an experiment on the test, use the three-question strategy: What was changed? What was measured? What was kept the same? Then check whether the design allows for a fair comparison. A well-designed experiment changes only one variable at a time, includes a control group, uses a large and randomly assigned sample, and is repeated to confirm results.