On this page · 5 sections
Section II carries 50% of your AP Biology score, the same weight as all 60 multiple-choice questions. You get 90 minutes for six questions. Questions 1 and 2 are long, worth 8 to 10 points each, and both hand you an experiment to interpret. Question 2 is the one that requires you to construct a graph. Questions 3 through 6 are short, worth 4 points each, and cover scientific investigation, conceptual analysis, analysis of a model or visual representation, and data analysis.
Scoring is additive and mechanical. A reader works down a rubric of separate scorable elements and marks each one earned or not earned. There is no deduction for clumsy writing, no credit for a strong opening, and no partial credit inside a single point. Extra correct biology is free only where the task asks for more than one item. When a part asks for one example, the reader scores the first response you give, so piling on extras can only cost you. A contradiction costs you outright: if one sentence says the proton gradient collapses and the next says ATP synthase keeps spinning at full rate, the reader cannot award that point to either statement.
Two things follow from how the rubric works. Each scorable claim has to sit somewhere a reader can find it in about ten seconds, so write one idea per sentence and label every part. And every point is tied to a task verb. On the rubric, an "identify" point and an "explain" point are both worth 1, but the explain point is never awarded for a bare name. You are paid per correct, findable, verb-appropriate claim.
You have the formula sheet and a calculator for this section. Check the current question counts and timing against the official course and exam description before test day, since the College Board adjusts them from time to time.
The steps
- 1
Budget the clock before you write a word
Write a target finish time next to each question the moment the section starts. Give each long question 24 minutes and each short question 9 minutes. That is 84 minutes, which leaves six at the end for the final sweep. Then obey the clock. A short question is capped at 4 points, so a minute spent past your budget on Question 5 is worth less than the first minute spent on an untouched Question 6. If you stall, write your best partial answer, mark the part with a star, and move on. Points come back to you at the end; time does not.
- 2
Circle the task verb and count the tasks
Each lettered part contains a verb that sets the depth of the answer, and a number that sets how many scorable pieces you owe. Identify or state means name it in a phrase, with no reasoning scored. Describe means give the characteristics or the sequence of what happens. Explain means give the cause: answer how or why, not what. Justify means make a claim and then tie it to specific evidence from the stem or the data. Predict means state a direction, such as increase, decrease, or no change, and usually supply a reason. Calculate means write the setup with the numbers substituted in, then the answer with units, rounded to match the precision of the data, usually two or three significant figures. Rubrics often award one point for the correct setup and another for the correct answer, so never hand in the final number alone. Then count: "describe one change to the protocol and explain how it would affect the results" is two separate points, and answering only the first half caps you at one.
- 3
Rebuild the experiment in the margin first
Before you answer part (a) of a long question, read the entire stem and write five things in the margin: the independent variable, the dependent variable, the control group, what was held constant, and the sample size. Add the units on each axis or column of the data. This takes about 90 seconds and it pays twice. Long questions often award a point directly for naming a variable, a control, or a testable hypothesis. The later parts almost always change one variable and ask what happens, and you cannot predict that if you never sorted out which variable was which. If you cannot name the control group, you are not ready to start writing.
- 4
Answer in labeled parts, one claim per sentence
Start each answer with the part letter, and answer the parts in the order given. Skip the introduction, and do not restate the question; neither earns a point and both burn clock. Put one scorable idea in each sentence so the reader is never hunting through a paragraph for the claim. Write in complete sentences: a bare list or a labeled diagram on its own is not scored as a response, though a diagram added beside a written answer can help. When a part asks for one example, give one. A second example does not add points, and if the second one is wrong while the first is right, you have handed the reader a contradiction.
- 5
Turn every explanation into a because-clause
For any explain, justify, or predict task, force yourself to write the word because, then name a molecule, a structure, or a process on the other side of it. "Blocking ATP synthase decreases ATP production" is a restatement and scores nothing on an explain point. "Blocking ATP synthase decreases ATP production because protons can no longer flow down their concentration gradient through the enzyme, and that flow is what drives the phosphorylation of ADP" gives the reader the mechanism they are told to look for. The same test works across the blueprint. The cell swells because the solution outside has the higher water potential, so water moves down that gradient through aquaporins. The signal is amplified because one activated receptor switches on many G proteins, and each of those activates an enzyme that makes many second messengers. Transcription stops because the repressor occupies the operator and RNA polymerase cannot move past it.
- 6
Build the graph against the rubric, not by eye
Graphing points are the cheapest points on the exam and the most commonly dropped. Put the independent variable on the x-axis and the dependent variable on the y-axis. Label both axes with the variable name and the units. Use even intervals that spread the data across most of the grid, not a scale that crushes every point into one corner. Choose a bar graph when the independent variable is categorical, such as wild type versus mutant, and a line or scatter plot when it is numerical, such as temperature or concentration. Plot the points accurately. If the data include standard error or standard deviation, draw the error bars. Add a key whenever you plot more than one data set.
- 7
Work the model, do not describe it
Question 5 hands you a diagram, a pathway, a phylogenetic tree, or a sketch of a model and asks you to use it. Before you write, translate the picture. Say what each arrow, axis, label, and node stands for. An arrow can mean converts to, activates, inhibits, or moves toward, and those are not interchangeable. Then answer with the model rather than about it. If a tree shows two species meeting at a recent node, the scorable claim is the shared derived trait and the recent common ancestor, not the shape of the branch. Question 4, conceptual analysis, works the same way without a picture. It names a change and wants a causal chain: tie the structure to its function, then the function to the outcome the question asks about.
- 8
Sweep for blanks, contradictions, and units
Save the last five minutes. Check three things in this order. Is every lettered part answered, including the starred ones you skipped? Does any answer contain two statements that cannot both be true? Does every number carry a unit, and does every part that told you to use the data actually cite a value, a trend, or an error bar from the figure? A part that says "using the data in Table 1" is scored on whether you referred to the data, and a perfect answer written from memory can miss the point.
Worked through
Here is a full long-question response, written the way it would earn points.
The prompt. Researchers grew yeast in aerated culture, then moved equal samples into vials of buffer at four glucose concentrations. Each vial was capped with air in the headspace and a carbon dioxide probe, and the rate of CO2 production was recorded over the first two minutes, while oxygen was still plentiful. Assume the yeast respire aerobically in these vials. Each treatment was run in five vials at 25 degrees Celsius. The reported values are means, with error bars of plus or minus 2 standard errors. At 0% glucose the rate was 0.02 mL/min plus or minus 0.01, at 1% it was 0.35 plus or minus 0.04, at 2% it was 0.62 plus or minus 0.05, and at 4% it was 0.68 plus or minus 0.06. (a) Identify the independent variable and describe the purpose of the 0% vials. (b) Construct a graph of the results. (c) Using the data, evaluate the claim that raising glucose from 2% to 4% increased the rate of CO2 production. (d) The researchers repeat the 4% treatment in a vial flushed with nitrogen so that no oxygen is present. Predict the effect on the number of ATP produced per glucose molecule and justify your prediction.
Margin notes first. IV = glucose concentration. DV = rate of CO2 production in mL/min. Control = 0% glucose vials. Constants = temperature, vial volume, yeast amount, time. n = 5 per treatment.
(a) The independent variable is the glucose concentration in the vial. The 0% vials are the control group. They show how much CO2 the yeast release when no glucose is supplied, so any increase in the other treatments can be attributed to the added glucose rather than to the yeast, the probe, or the vial.
(b) X-axis: glucose concentration in percent, scaled 0 to 4 at even intervals. Y-axis: rate of CO2 production in mL/min, scaled 0 to 0.8 in steps of 0.1. Glucose concentration is numerical, so plot the four means as points and connect them or fit a curve. Draw a vertical error bar through each point: plus or minus 0.01, 0.04, 0.05, and 0.06 respectively.
(c) The claim is not supported by these data. At 2% the interval runs from 0.57 to 0.67 mL/min, and at 4% it runs from 0.62 to 0.74 mL/min. The error bars overlap, so the difference between the two means could easily be produced by variation among vials, and you cannot conclude that the rate rose. The pattern instead suggests the rate is leveling off, which is what you expect once glucose uptake and the enzymes that process it are already running at their maximum rate. Past that ceiling, more glucose outside the cell cannot make the reactions inside go any faster.
(d) The ATP produced per glucose will drop sharply. With oxygen present, a glucose molecule yields roughly 30 ATP; with no oxygen, only the net 2 ATP from glycolysis are produced. Without oxygen there is no final electron acceptor for the electron transport chain, so electrons back up, the chain stops pumping protons, and the gradient across the inner membrane collapses. ATP synthase then has no proton flow to drive it and oxidative phosphorylation stops. The only ATP left comes from substrate-level phosphorylation in glycolysis, which keeps running because fermentation converts pyruvate to ethanol and CO2 and regenerates the NAD+ that glycolysis needs.
Notice what each answer does. Part (a) names the variable and states the function of the control rather than merely labeling it. Part (c) quotes actual numbers from the figure because the part said "using the data," and it reaches a verdict instead of describing the bars. Part (d) gives a direction, a number, and a chain of causes that runs from the missing oxygen to the missing gradient to the missing ATP. Nothing in the response has an introduction, and nothing is longer than it needs to be.
Now a calculate part, worked. A short question crosses two heterozygous purple-flowered plants and counts 200 offspring: 148 purple and 52 white. Use a chi-square test to evaluate the predicted 3:1 ratio at p = 0.05.
Null hypothesis: the offspring fit a 3:1 purple to white ratio, and any deviation is due to chance. Expected counts: 200 x 3/4 = 150 purple, and 200 x 1/4 = 50 white. Chi-square = (148 - 150)^2 / 150 + (52 - 50)^2 / 50 = 0.03 + 0.08 = 0.11. Degrees of freedom = number of categories - 1 = 2 - 1 = 1. The critical value at p = 0.05 with 1 degree of freedom is 3.84. Since 0.11 is less than 3.84, you fail to reject the null hypothesis. The data are consistent with a 3:1 ratio.
Write the substitution line out in full. If the only thing on the page is 0.11, the setup point is gone. And phrase the verdict as reject or fail to reject: a chi-square test never proves or accepts a null hypothesis.
Common mistakes
Answering an "explain" prompt with a name or a restatement. Writing that the mutation stops translation, or that the rate decreased, tells the reader what happened; the explain point is awarded for why it happened, in terms of a molecule, structure, or process.
Describing a figure instead of interpreting it. "The bar for the treated group is taller" is a reading of the graph, not an analysis. Say what the difference means biologically, and when the part says to use the data, cite a specific value, trend, or error bar.
Hedging by giving two answers. Listing every possible cause so one of them is right does not work, because a response that contains both a correct and an incorrect statement about the same task is not awarded the point. Choose the best answer and commit to it.
Losing free graphing points: axes without units, the independent variable on the y-axis, a scale with uneven intervals, missing error bars when the data give standard error, or a bar graph used for a numerical independent variable.
Confusing the control group with the constants. The control is the group that does not receive the treatment and sets the baseline for comparison. The constants are the conditions kept identical across all groups, such as temperature and pH.
Mishandling significance. Overlapping error bars mean you cannot conclude the groups differ, and a chi-square value below the critical value at p = 0.05 means you fail to reject the null hypothesis. You never accept or prove a null hypothesis, and two means that differ on paper are not automatically different.
Checklist
- Can you write your six finish times on the page within the first minute and then hold to them?
- Can you say, for any part, what the verb demands: identify, describe, explain, justify, predict, or calculate?
- Can you name the independent variable, the dependent variable, the control group, the constants, and the sample size from a stem you have read once?
- Can you convert any explanation into a because-clause that names a specific molecule, structure, or process?
- Can you construct a graph with both axes labeled and scaled, the correct plot type for the variable, and error bars, in under five minutes?
- Can you decide from overlapping or separated error bars whether two treatments differ, and say why?
- Can you state a null hypothesis, work out the degrees of freedom, compare the chi-square value with the critical value at p = 0.05, and phrase the result as reject or fail to reject?
- Can you say what every arrow, axis, and label in a diagram represents before you write a word about it?
- Can you finish a response with no unanswered lettered parts, no self-contradiction, and a unit on every number?