What you'll learn
Scientists answer questions by designing investigations. The GED does not ask you to run experiments, but it does ask you to think like the person who designed one. You will read a short description of an experiment and decide what was being tested, what was changed, what was measured, and whether the design was fair.
In this lesson you will learn to:
- recognize a testable hypothesis;
- identify independent, dependent and controlled variables;
- explain the role of a control group;
- judge reliability (repeated trials, sample size) and spot flaws such as confounding variables;
- choose the best way to improve or extend an investigation.
How it shows up on the GED: Investigation design is one of the most heavily tested science practices. Expect questions such as "What is the independent variable?", "Which group is the control?", "Which hypothesis was tested?", "Which change would improve the investigation?", and "What is a flaw in this study?"
Key terms
- investigation (experiment) — a planned test to answer a scientific question (testing fertilizer on tomato plants).
- hypothesis — a testable prediction, often "If…, then…" ("If more salt is added, then water freezes at a lower temperature").
- independent variable — the factor the experimenter changes on purpose (amount of salt).
- dependent variable — the factor measured as the result (freezing temperature).
- controlled variable (constant) — a factor kept the same for all groups (water volume, cup type).
- control group — the group that does not get the treatment, used for comparison (plants with no fertilizer).
- experimental group — a group that gets the treatment (plants with 5, 10 or 15 g fertilizer).
- trial — one run of the experiment.
- sample size — how many subjects or items are tested (10 runners per group).
- reliable — gives similar results when repeated.
- valid — actually tests what it claims to test (a fair test).
- confounding variable — an uncontrolled factor that changes along with the independent variable and could explain the result (outdoor temperature on race day).
- random assignment — placing subjects in groups by chance so the groups start out alike.
- placebo — a fake treatment (a sugar pill) given to a control group so they do not know they are untreated.
- blind study — participants do not know which group they are in; in a double-blind study, the researchers measuring results do not know either.
- outlier — a result far from the others under the same conditions (51 s when other trials were 32 and 34 s).
1. Questions and hypotheses
An investigation begins with a question ("Does fertilizer affect tomato yield?"). A hypothesis is a testable prediction that answers it.
A good hypothesis:
- names the independent variable and the dependent variable;
- predicts a direction (more, less, faster, lower);
- can be checked with measurements.
| Hypothesis | Testable? | Why |
|---|---|---|
| If plants get more light, then they will grow taller. | Yes | names light (changed) and height (measured) |
| Plants like music. | No | "like" cannot be measured |
| If salt is added, the water will freeze faster. | Yes, but only if time is measured | the design must measure what the hypothesis predicts |
GED tip: When asked which hypothesis an experiment tests, match the hypothesis to the design. If the experimenter measured temperature, the hypothesis must be about temperature, not speed.
2. Variables and control groups
Ask three questions about any experiment:
- What did the scientist change on purpose? → independent variable.
- What did the scientist measure as the result? → dependent variable.
- What did the scientist keep the same? → controlled variables.
Example: A student drops a ball from heights of 0.5 m, 1.0 m and 1.5 m and measures how high it bounces, using the same ball and the same floor.
- Independent: drop height. Dependent: bounce height. Controlled: ball, floor.
On a graph, the independent variable usually goes on the x-axis and the dependent variable on the y-axis. In a table, the independent variable is usually the first column and the results come later.
Control group vs. controlled variables. These sound similar but are different:
- A control group is a group that gets no treatment (0 g fertilizer) or the normal condition. It gives a baseline for comparison.
- Controlled variables are conditions kept the same for all groups (water, soil, sunlight).
Without a control group, you cannot tell whether the treatment did anything. If every plant in a fertilizer test had received fertilizer, the scientist could not show that fertilizer increased the harvest compared with no fertilizer.
3. Fair tests and confounding variables
A fair test changes only one independent variable at a time. If two things change at once, you cannot tell which one caused the result.
Example of a confound: A company gives its sports drink to runners on a cool morning and water to other runners on a hot afternoon. The drink group is faster, but the temperature also changed. The heat, not the drink, could explain the difference.
Tools that make tests fair:
- Same conditions for all groups (same time, place, equipment).
- Random assignment, so groups start out alike. When people choose their own group, the groups may differ in fitness, health or habits.
- Placebo for the control group in medical or nutrition studies, so expectations do not affect results.
- Blinding: participants (and ideally the people measuring) do not know who got the treatment.
Experiments vs. observational studies. In an experiment, the scientist assigns the treatment. In an observational study, the scientist only records what people already do (a survey of coffee drinkers). Observational studies can show that two things are related, but they cannot prove that one causes the other, because the groups may differ in other ways.
4. Reliability: trials, sample size and outliers
Reliability means getting similar results if the investigation is repeated. Two ways to improve it:
- Repeated trials: run each condition several times and average the results.
- Larger sample size: test more plants, people or items. With only a few subjects, one unusual individual can change the average a lot.
Outliers. Sometimes one trial is far from the others. In this test of how many seconds a tablet takes to dissolve in water:
| Temperature (°C) | Trial 1 | Trial 2 | Trial 3 |
|---|---|---|---|
| 30 | 32 s | 34 s | 51 s |
The 51 s result is far from 32 and 34 s and breaks the overall pattern. A good scientist repeats the trial or looks for an error (a timing mistake, a different tablet). Simply ignoring data without a reason is not good practice, but recognizing a likely error is.
Precision of measurement matters too: timing with a stopwatch is more precise than "counting in your head", and measuring with a graduated cylinder is more precise than "about half a cup".
5. Spotting flaws and improving designs
GED questions often ask "Which is a flaw?" or "Which change would best improve the investigation?" Check this list:
| Flaw | Fix |
|---|---|
| more than one variable changed | change only the independent variable |
| no control group | add a group with no treatment |
| very small sample | test more subjects |
| only one trial | repeat trials and average |
| groups chosen by participants | random assignment |
| participants know their group | placebo and blinding |
| measures the wrong thing | measure what the hypothesis predicts |
Choosing the best improvement: The best answer fixes the biggest problem described in the passage. An answer that changes a controlled variable, removes the control group, or adds a new variable makes the design worse, even if it sounds scientific.
Finding the mean of repeated trials
When an investigation has repeated trials, scientists report the mean (average):
\text{mean} = \frac{\text{sum of the values}}{\text{number of values}}Example: Trials of 61 s, 59 s and 60 s:
\text{mean} = \frac{61 + 59 + 60}{3} = \frac{180}{3} = 60\text{ s}You are expected to know this formula. Remember that one outlier can pull the mean far up or down.
Worked example 1: Identifying variables
Stimulus: "A student wanted to know whether the length of a paper airplane's wings affects how far it flies. She made four airplanes from the same type of paper with wing lengths of 6, 8, 10 and 12 cm. She threw each one from the same spot with the same launcher and measured the distance it flew."
Question: Which is a controlled variable? A. wing length B. distance flown C. type of paper D. the student's hypothesis
Think-aloud:
- Changed on purpose: wing length → independent, so not A.
- Measured: distance flown → dependent, so not B.
- Kept the same: "the same type of paper", "the same spot", "the same launcher" → controlled.
- D is not a variable at all.
Answer: C, type of paper.
Worked example 2: Choosing a testable hypothesis
Stimulus: "A researcher grows yeast in sugar water at 20 °C, 30 °C and 40 °C and measures the volume of gas the yeast produces in one hour."
Question: Which hypothesis does this experiment test? A. Yeast is a living thing. B. If the temperature of sugar water increases, then the volume of gas produced by yeast in one hour will change. C. Yeast makes bread taste better. D. If more sugar is added, then more gas will be produced.
Think-aloud:
- Independent variable: temperature (20, 30, 40 °C). Dependent: gas volume.
- B links temperature to gas volume. ✓
- D is about sugar amount, which was not changed. A and C cannot be tested with this design.
Answer: B.
Worked example 3: Finding the flaw
Stimulus: "To test a new potting soil, a gardener planted 5 seeds in the new soil and placed them on a sunny windowsill. She planted 5 seeds in regular soil and placed them in a shady corner. After 3 weeks, the plants in new soil were taller. She concluded the new soil is better."
Question: What is the main flaw? A. She used seeds instead of small plants. B. The two groups received different amounts of sunlight. C. She measured height instead of mass. D. She waited 3 weeks.
Think-aloud:
- Intended independent variable: soil type.
- But sunlight also differed: windowsill vs. shady corner. That is a confounding variable.
- A, C and D are reasonable choices and were the same for both groups.
Answer: B. Fix: put both groups in the same location (and use more seeds).
Worked example 4: Repeated trials and the mean
Stimulus: A student times how long it takes a toy car to roll down a ramp: Trial 1: 2.4 s, Trial 2: 2.6 s, Trial 3: 2.5 s, Trial 4: 4.1 s. She notes that during Trial 4 the car hit the side of the ramp.
Question: What is the best value to report for the time down the ramp? A. 2.9 s, the mean of all four trials B. 2.5 s, the mean of trials 1–3 C. 4.1 s, the longest time D. 2.4 s, the shortest time
Think-aloud:
- Trial 4 is an outlier with a known cause (the car hit the side), so it does not represent a normal run.
- Mean of trials 1–3: \dfrac{2.4 + 2.6 + 2.5}{3} = \dfrac{7.5}{3} = 2.5 s.
- Mean of all four: \dfrac{11.6}{4} = 2.9 s, pulled up by the error.
- Choosing only the longest or shortest trial ignores the other data.
Answer: B, 2.5 s. Better still, she should run another trial to replace Trial 4.
Common mistakes to avoid
- Swapping independent and dependent. Wrong: "Tomato mass is the independent variable." Right: the experimenter changes the independent variable (fertilizer) and measures the dependent one (tomato mass).
- Confusing control group with controlled variables. Wrong: "Water is the control group." Right: water amount is a controlled variable; the 0 g fertilizer group is the control group.
- Thinking the best result is the control. Wrong: "Group 4 grew the most, so it's the control." Right: the control is the group with no treatment.
- Choosing a fix that adds a variable. Wrong: "Add sugar to some cups to improve the salt experiment." Right: improve reliability by repeating trials or enlarging the sample.
- Missing the confound. Wrong: "The drink worked because that group was faster." Right: check whether anything else (time, temperature, who was chosen) also differed.
- Mismatching the hypothesis. Wrong: choosing a hypothesis about speed when the experiment measured temperature. Right: the hypothesis must match what was measured.
GED test strategy
- Label the experiment as you read: write IV (what's changed), DV (what's measured), and C (what's kept the same) on your erasable note board.
- Find the control group by looking for "no treatment", "0", "placebo" or "normal conditions".
- Look for words of sameness: same, identical, each, all. They point to controlled variables.
- Ask "What else changed?" for any flaw question. A second difference between groups is the most common flaw.
- For "best improvement" questions, pick the choice that fixes the biggest problem without breaking anything else. Reject choices that remove the control or change two things at once.
- Remember the reliability fixes: more trials, larger sample, average the results.
- Pace yourself: about 1 to 1.5 minutes per question.
Check yourself
Fertilizer and Tomato Plants
Maria, a community gardener, wants to find out whether the amount of fertilizer affects how many tomatoes her plants produce. She sets up 20 tomato plants of the same variety in identical pots filled with the same soil. She divides the plants into four groups of 5 plants each.
Each week, Group 1 gets no fertilizer, Group 2 gets 5 g, Group 3 gets 10 g, and Group 4 gets 15 g. All plants get 500 mL of water each day, and all are placed in the same sunny spot. After 10 weeks, Maria weighs all the tomatoes from each plant and calculates the average mass of tomatoes per plant for each group.
| Group | Fertilizer per week (g) | Average mass of tomatoes per plant (kg) |
|---|---|---|
| 1 | 0 | 1.2 |
| 2 | 5 | 1.8 |
| 3 | 10 | 2.3 |
| 4 | 15 | 2.4 |
What is the dependent variable in Maria's investigation?
Hint
The dependent variable is the result that is measured at the end. It depends on what the experimenter changed.
Step-by-step solution
The dependent variable is what the experimenter measures to see the effect. The passage says that after 10 weeks, Maria "weighs all the tomatoes from each plant and calculates the average mass of tomatoes per plant." The table's last column, "Average mass of tomatoes per plant (kg)", holds those results.
Step by step:
- What did Maria change on purpose? The fertilizer amount (0, 5, 10, 15 g). That is the independent variable, so A is wrong.
- What did she measure as the outcome? The mass of tomatoes. That is the dependent variable.
- Water (500 mL) and the number of plants per group (5) were kept the same, so they are controlled variables.
The correct answer is B.
Tip: Ask "What did they change?" (independent) and "What did they measure?" (dependent). The dependent variable is usually the last column of a results table.
Check yourself
Fertilizer and Tomato Plants
Maria, a community gardener, wants to find out whether the amount of fertilizer affects how many tomatoes her plants produce. She sets up 20 tomato plants of the same variety in identical pots filled with the same soil. She divides the plants into four groups of 5 plants each.
Each week, Group 1 gets no fertilizer, Group 2 gets 5 g, Group 3 gets 10 g, and Group 4 gets 15 g. All plants get 500 mL of water each day, and all are placed in the same sunny spot. After 10 weeks, Maria weighs all the tomatoes from each plant and calculates the average mass of tomatoes per plant for each group.
| Group | Fertilizer per week (g) | Average mass of tomatoes per plant (kg) |
|---|---|---|
| 1 | 0 | 1.2 |
| 2 | 5 | 1.8 |
| 3 | 10 | 2.3 |
| 4 | 15 | 2.4 |
Which group serves as the control group in Maria's investigation?
Hint
A control group does not receive the treatment being tested.
Step-by-step solution
A control group gets none of the treatment being tested, so it shows what happens without it. Maria is testing fertilizer, and the passage says "Group 1 gets no fertilizer." The table confirms Group 1 received 0 g per week.
Step by step:
- The treatment being tested is fertilizer.
- Find the group with none: Group 1 (0 g).
- Groups 2, 3 and 4 all received fertilizer, so they are experimental groups.
- Group 4 had the largest result (2.4 kg), but the best result is not what makes a group the control.
The correct answer is A, Group 1.
Tip: The control group is the "normal" or "nothing added" group used for comparison. Without it, Maria could not tell whether fertilizer helped at all.
Summary
- A testable hypothesis links the independent variable to the dependent variable and can be checked by measurement.
- Independent = changed on purpose; dependent = measured; controlled = kept the same. The control group gets no treatment.
- A fair test changes only one variable; watch for confounding variables such as time, temperature or self-chosen groups.
- Reliability improves with repeated trials, larger samples and averaging; investigate outliers instead of trusting them.
- The best improvement fixes the biggest flaw without adding a new variable or removing the control.