Advanced reading practice test 4

Reading Passage 1

The Science of Decision Architecture

APeople often assume that important decisions result from a careful comparison of available options. In practice, the way choices are presented can influence what people select, even when the options themselves remain unchanged. Researchers call this influence choice architecture: the arrangement of information, the order of alternatives and the effort required to act. It does not necessarily remove freedom, but it can make some actions easier to notice or complete than others.

BOne familiar example is the default option. When a form automatically selects one choice unless a person changes it, many people leave the selection as it is. This may happen because the default appears to be recommended, because changing it requires effort, or because people postpone decisions when they are uncertain. Defaults can therefore have substantial effects in areas such as savings, data privacy and organ donation. Their influence, however, depends on context; a default that works in one setting may be ignored in another.

CThe order in which information appears can also matter. Consumers may pay more attention to the first items on a list, while people comparing prices can be influenced by an initial figure that serves as an anchor. An anchor can shape later judgement even when it is not especially relevant. For example, a high initial price may make a reduced price seem attractive, although the reduced price could still be greater than the product's value to the buyer. Awareness of anchoring does not eliminate it, but it can encourage people to compare alternatives more deliberately.

DChoice architecture is not automatically manipulative. A well-designed form can help people complete a task accurately, and clear reminders can make it easier to meet a goal that a person already values. The ethical question is partly whether the design supports the user's interests or mainly serves the organisation that created it. A reminder to renew an important document may be helpful; a confusing cancellation process designed to keep customers paying raises a different concern.

EResearchers evaluating these interventions must distinguish between an immediate change in behaviour and a lasting improvement. A reminder may increase attendance at one appointment without changing a person's longer-term habits. Similarly, a simplified application may increase the number of people who begin a process but not the number who successfully complete it. Evaluation should therefore measure outcomes that matter, not merely the first action that is easiest to count.

FThere are also limits to what can be inferred from average results. An intervention may help people who are familiar with a system while confusing those who are less experienced or who use a different language. The same default can be convenient for one group and inappropriate for another. Researchers should examine who benefits, who may be disadvantaged and whether people understand the choice being offered.

GTransparency can improve the legitimacy of choice architecture, although disclosure alone does not guarantee fairness. Organisations can explain why a default was selected, provide a simple way to change it and avoid designs that hide important consequences. Independent review may be appropriate when decisions affect money, health or legal rights. These safeguards do not remove the need for judgement, but they make it easier to question whether a design is serving its stated purpose.

HThe central insight is that decisions are made within environments, not in isolation. People remain responsible for their choices, yet the structure around those choices can help or hinder careful judgement. Good decision architecture makes useful actions easier without concealing alternatives or exploiting predictable mistakes. Its success should be judged by the quality of decisions and the interests they serve, rather than by whether it simply produces more of the behaviour an organisation wants.

Reading Passage 2

The Return of Peatlands

APeatlands are landscapes in which waterlogged conditions slow the decomposition of plant material. Over long periods, partly decayed vegetation accumulates to form peat, storing large quantities of carbon in the ground. Peatlands occur in different climates, from cool northern regions to tropical areas. Although they cover a relatively small share of the Earth's land surface, their carbon stores can be substantial, making their condition important for climate policy as well as biodiversity.

BMany peatlands have been drained to create farmland, extract peat for fuel or horticulture, or make land easier to develop. Drainage allows oxygen to enter the soil, accelerating decomposition and releasing carbon dioxide. Dry peat can also become more vulnerable to fire, which may release large amounts of stored carbon and damage habitats. The consequences vary with local conditions, but drainage can transform a landscape that accumulated carbon over centuries into a source of emissions.

CRestoration often begins by raising the water table. In some places, managers block drainage channels so that water remains closer to the surface. The aim is not necessarily to create a permanent lake, but to recover the wet conditions that allow peat-forming vegetation to grow. Rewetting can reduce oxidation and fire risk, although it may also produce methane under some conditions. The overall climate effect depends on the balance between different gases over time, the previous condition of the site and the success of the intervention.

DRestoring vegetation is another important task. Certain mosses and other plants help build peat, while trees or grasses may dominate where the ground has been drained. Re-establishing the appropriate plant community can take time, and planting alone may fail if the water level remains unsuitable. Some sites recover naturally once hydrology improves; others require active management. Restoration plans therefore need to respond to the history and ecology of each location rather than applying one recipe everywhere.

EThe benefits extend beyond carbon storage. Healthy peatlands can support specialised species, regulate water movement and help maintain water quality by retaining some pollutants and sediments. Their ability to reduce flood peaks depends on the landscape, the condition of the peat and the scale of the rainfall event. It would be misleading to present every restored peatland as a complete flood defence, but restoring natural water processes can form part of a wider catchment strategy.

FRestoration can create economic and social tensions. Farmers may depend on drained land for grazing or crops, and communities may rely on peat extraction for employment. Rewetting can change which activities are practical, so projects may require compensation, alternative income or agreements about land management. If local people are involved only after plans have been designed, they may regard restoration as an external demand rather than a shared effort.

GMonitoring is essential because visible greening does not necessarily prove that a peatland is recovering its former functions. Researchers may measure water levels, greenhouse-gas exchange, vegetation and the movement of carbon through the system. Some changes occur quickly, while others take decades. Short monitoring periods can therefore miss long-term benefits or overlook unexpected effects. Clear baselines and repeated measurements help distinguish real recovery from seasonal variation.

HPeatland restoration is neither a simple return to an untouched past nor a universal solution to climate change. It is a long-term effort to recover ecological processes in landscapes that people continue to use. Success depends on water management, suitable vegetation, local cooperation and evidence gathered over time. Protecting intact peatlands is generally easier than rebuilding damaged ones, so preventing further loss remains as important as restoring places that have already been altered.

Reading Passage 3

When Algorithms Shape Public Services

APublic agencies increasingly use algorithms to help allocate resources, identify patterns and prioritise cases. A system may estimate which roads need repair, flag unusual financial claims or help determine where inspections are most urgently required. Supporters argue that these tools can process information at a scale that would be difficult for staff to manage manually. Yet an algorithm does not make a decision in a social vacuum: its design reflects choices about objectives, data and the consequences of different kinds of error.

BThe data used to build a system may reflect earlier decisions and inequalities. If an agency has historically inspected some neighbourhoods more often than others, its records may contain more detected violations from those places. A model trained on those records could then direct even more inspections to the same neighbourhoods, producing additional data that appear to confirm the original pattern. The system may reproduce a pattern of enforcement rather than identify the true distribution of the underlying problem.

CAccuracy alone is not enough to assess a system used in public services. Overall performance can hide different error rates between groups, and the consequences of an error may be unequal. Incorrectly flagging a case for review may create inconvenience, while denying a benefit or delaying essential support can have much greater consequences. Evaluation should therefore examine the type of decision being supported, the severity of possible mistakes and the groups most likely to experience them.

DSome systems are difficult to explain because their predictions depend on many interacting variables. Technical complexity does not necessarily mean that a system is unfair, just as a simple rule is not automatically fair. However, people affected by a decision need meaningful information about the reasons for it and how to challenge it. An explanation that merely repeats a technical score may be of little use to someone who needs to correct an inaccurate record or appeal an outcome.

EHuman oversight is often proposed as a safeguard, but the presence of a human reviewer does not guarantee independent judgement. Staff may rely heavily on an algorithm's recommendation, especially when they are busy or when the system is presented as more objective than human judgement. This tendency can be reduced through training, clear authority to disagree with the system and procedures that record why a recommendation was accepted or rejected.

FProcurement decisions can shape public accountability long before an algorithm is used. Agencies may purchase tools from private suppliers whose methods are protected as commercial information. Contracts should clarify who is responsible for testing, updating and correcting the system, and whether independent auditors can examine relevant evidence. Without these arrangements, a public agency may remain responsible for a decision while lacking the information needed to explain or improve the process behind it.

GAlgorithms can still offer genuine benefits when they are used for suitable tasks and monitored carefully. They may help staff notice patterns, reduce repetitive work or direct attention to cases that need closer examination. In many situations, the strongest arrangement is not full automation but a system in which computational analysis supports professional judgement. The right balance depends on the stakes, the quality of the evidence and the availability of meaningful review.

HThe question is not simply whether an algorithm performs better than a person on one test. Public agencies must also ask whether the system serves a legitimate purpose, whether its errors are acceptable, whether people can challenge its use and whether its performance changes over time. These questions require technical evaluation as well as public accountability. Used carefully, algorithms can support services; used without scrutiny, they can make old problems harder to see because their decisions appear to be the result of neutral calculation.