Model Essay
Source A
Source B (Asteria University press release excerpt)
Asteria University adopted ScholarSort to manage a growing number of scholarship applications. The university states that the algorithm improves consistency by applying the same scoring rules to all applicants and by reducing manual handling of documents. ScholarSort is described as a set of step-by-step procedures: it checks eligibility, converts grades into a common scale, assigns points for financial need, and adds a “context” score based on school-resourcing indicators and first-generation status. Applicants can submit an appeal if they believe data is incorrect or circumstances are not captured. The university notes that ScholarSort is not designed to “predict future success” but to rank applicants according to published criteria. It also notes that changing weights (for example, increasing need) will change who receives awards, and that such changes are policy decisions.
Source C (same as above)
Applications processed: 8,400 (2024) → 14,900 (2026); staff hours spent on initial screening fell by 41% after ScholarSort.
Appeals rose from 3% to 9% of applicants; 28% of appeals resulted in a score change.
“Missing data defaults” affected 12% of applications (most commonly incomplete household documents); these applicants’ average rank dropped by 18 percentile points.
Among applicants with the same academic score band, those from low-resourcing schools received scholarships at 1.6× the rate of those from high-resourcing schools (consistent with the context weighting).
Audit sampling found the tie-break rule (alphabetical by surname when composite scores match) decided 7% of final offers in one faculty.
Source D (criticism by student)
ScholarSort is presented as neutral because it is “just applying published weights,” but the ethical problem is that the weights and defaults quietly become the real gatekeepers. A missing document does not simply mean “unknown”; it becomes a penalty through default scoring, which often hits families with less time, less connectivity, or more complex paperwork. The university celebrates efficiency, yet the rising appeal rate suggests the system is producing more disputable outcomes, not fewer. Even small technical choices, like using alphabetical tie-breaks, create arbitrary winners and losers while still looking objective on paper. And the phrase “the committee approves” can become a shield: humans rubber-stamp what the ranking already decided. If the algorithm is shaping access to education, accountability must include transparency about how each component changes outcomes, and a way to challenge decisions without requiring applicants to become expert auditors of a scoring system.
(d) Examine whether ScholarSort should be considered an appropriate use of algorithms in education funding. Answer with reference to all the sources (A–D) and your own knowledge of the Digital Society course.
ScholarSort should be considered only a conditionally appropriate use of algorithms in education funding. The sources show that it solves a real administrative problem and can support more consistent decisions, but they also show that apparently technical choices such as defaults, tie-breaks and review thresholds can have major consequences for fairness. In education funding, where decisions affect access and social mobility, efficiency alone is not enough: appropriateness depends on whether the system is transparent, auditable and genuinely accountable to applicants.
On the positive side, ScholarSort appears to be a relatively clear, rules-based system rather than a black-box predictor. Source A presents it as a sequence of visible steps: “Input data”, “Standardize scores”, “Calculate composite score”, “Apply constraints”, “Rank applicants” and “Offers sent”. This supports Source B, which says it is “not designed to ‘predict future success’ but to rank applicants according to published criteria”. That distinction matters in Digital Society because a system that applies stated rules is easier to justify than one making probabilistic judgements about who is likely to succeed. The weighting is also explicit in Source A: “Composite = 0.55 Academic + 0.30 Financial need + 0.15 Context”. This makes the algorithm a policy tool rather than an autonomous decision-maker. If a university wants to prioritise merit, need or widening participation differently, it can change the weights deliberately. Source B is clear that such changes are “policy decisions”, which means responsibility still lies with the institution.
The system also appears effective in handling scale. Source C shows applications rising from “8,400 (2024)” to “14,900 (2026)”, while “staff hours spent on initial screening fell by 41%”. Source A reinforces this with the claim that it “Runs nightly; recalculates rankings when new documents arrive.” In practice, this kind of automation can be appropriate because universities face large volumes of applications and manual processing can be slow, inconsistent and vulnerable to human error. Applying the “same scoring rules to all applicants”, as Source B states, may improve procedural consistency. Standardising grades onto a “common 0–100 scale” is also potentially fairer when applicants come from different schools or qualifications.
There is also evidence that ScholarSort is being used to pursue equity, not just efficiency. The “context” score includes “school-resourcing indicators and first-generation status”, and Source C reports that among applicants in “the same academic score band”, those from “low-resourcing schools received scholarships at 1.6× the rate” of those from high-resourcing schools. This suggests that the context weighting is doing what it is intended to do: correcting for structural disadvantage rather than treating unequal applicants as if they had identical opportunities. From a Digital Society perspective, this is important because fairness is not always the same as identical treatment. Sometimes substantive fairness requires differential weighting to offset unequal starting conditions.
However, the evidence also shows significant weaknesses that make ScholarSort questionable in its current form. The most serious issue is missing-data defaults. Source A itself flags “missing data defaults” as one of the key “Accountability points”, and Source C shows that these defaults affected “12% of applications” and that these applicants’ “average rank dropped by 18 percentile points”. That is not a trivial technical matter; it is a substantial penalty. Source D argues persuasively that missing data does not just mean “unknown” but becomes “a penalty through default scoring”, often harming families with “less time, less connectivity, or more complex paperwork”. This links directly to the Digital Society idea of the digital divide. If applicants are disadvantaged not because they are less deserving but because they struggle to upload documents or navigate bureaucracy, then the system reproduces inequality under the appearance of objectivity.
The appeal data strengthens this criticism. Although Source B says applicants “can submit an appeal if they believe data is incorrect or circumstances are not captured”, Source C shows appeals rising from “3% to 9%” and that “28% of appeals resulted in a score change”. This suggests the appeals system is not merely a safeguard for rare exceptions; it is correcting a significant number of outcomes. That weakens the claim that the algorithm is reliably fair in the first instance. It may also shift the burden onto applicants, who must notice an error, understand the system well enough to challenge it, and have the time and resources to appeal. Source D is right to argue that accountability should not require students to become “expert auditors of a scoring system”. In terms of procedural justice, a process is less fair if redress exists only for those capable of navigating it.
Another problem is arbitrariness in the tie-break rule. Source A lists “tie-break rule” under “Apply constraints”, and Source C reveals that “alphabetical by surname when composite scores match” decided “7% of final offers in one faculty”. Even if this affects a minority of cases, it undermines legitimacy because surname order has no meaningful relationship to merit, need or context. An automated system may appear objective simply because it is consistent, but consistency is not the same as fairness. A consistently arbitrary rule is still arbitrary. Better alternatives might include random allocation, additional human review, or a secondary criterion linked to policy goals.
The human oversight described in the sources is also weaker than it first appears. Source A states, “The algorithm ranks; the committee approves,” and highlights a “human review threshold (top 5% + flagged cases)”. On paper this suggests a human-in-the-loop model, which is often presented in Digital Society as a safeguard for automated decision-making. But Source D warns that “the committee approves” can become “a shield” if humans merely “rubber-stamp what the ranking already decided”. This is a valid concern. Human review is only meaningful if reviewers have the authority, time and willingness to question the system, not if they simply validate its outputs. Otherwise responsibility becomes blurred: the university can blame the algorithm, while the algorithm is said to be only implementing university policy.
Overall, then, ScholarSort is appropriate in principle but only if stronger governance is attached to it. Its use of explicit weights, published criteria and context scoring gives it advantages over more opaque predictive systems. It clearly improves efficiency and can support equity aims. Yet the current evidence shows non-trivial fairness risks: defaults that penalise missing documents, a high and successful appeal rate, and arbitrary alphabetical tie-breaks. For an algorithm allocating education funding to be appropriate, the university should publish how each component affects rankings, redesign defaults so missing data triggers review rather than automatic penalty, test regularly for disparate impact, and ensure appeals are simple and accessible. Clear responsibility must remain with institutional decision-makers, because they chose the criteria, the weights, the defaults and the review process.
Therefore, ScholarSort should not be seen as automatically appropriate just because it is efficient or rule-based. It is appropriate only if the university treats it as a governed socio-technical system, not a neutral machine. In its present form, the sources suggest it is useful but not yet fully justified, because its accountability mechanisms lag behind its influence over life-changing decisions.
Tip:
ScholarSort should be considered only a conditionally appropriate use of algorithms in education funding.
This opening takes a clear, qualified position straight away. The word 'conditionally' sets up a balanced argument rather than a one-sided response, which gives the whole essay direction.
On the positive side,
This transition neatly signals a shift into the supporting case. It helps the reader follow the essay’s balanced structure from the outset.
Source A presents it as a sequence of visible steps: “Input data”, “Standardize scores”, “Calculate composite score”, “Apply constraints”, “Rank applicants” and “Offers sent”.
This is a well-chosen textual detail from the source itself. Quoting the process labels precisely allows the argument about transparency to rest on concrete evidence rather than summary.
That distinction matters in Digital Society because a system that applies stated rules is easier to justify than one making probabilistic judgements about who is likely to succeed.
Here the response moves beyond description into conceptual interpretation. It explains why the source detail matters by linking rules-based systems to a wider issue in automated decision-making.
This makes the algorithm a policy tool rather than an autonomous decision-maker.
This is a sharp evaluative move: the writer reframes what the algorithm is for, which strengthens the line of argument. The phrasing is precise and controlled, giving the point authority.
The system also appears effective in handling scale.
This topic sentence does useful organizing work by naming the next line of argument before the evidence arrives. It keeps the essay cumulative and easy to follow.
Source C shows applications rising from “8,400 (2024)” to “14,900 (2026)”, while “staff hours spent on initial screening fell by 41%”.
The numerical evidence is integrated smoothly and compared directly, which makes the efficiency claim convincing. Using both figures together shows scale and impact in one sentence.
In practice, this kind of automation can be appropriate because universities face large volumes of applications and manual processing can be slow, inconsistent and vulnerable to human error.
This sentence interprets the data instead of leaving it to stand alone. It broadens the point from this case to a realistic institutional context, showing applied understanding.
Applying the “same scoring rules to all applicants”, as Source B states, may improve procedural consistency.
This is a strong argumentative technique: the quotation is embedded into the writer’s own sentence and immediately used to make a measured claim. The modal 'may' keeps the judgment careful rather than overstated.
There is also evidence that ScholarSort is being used to pursue equity, not just efficiency.
This sentence marks a clear development in the argument by adding a second positive dimension. It shows the response is not repeating one point, but extending the case thoughtfully.
The “context” score includes “school-resourcing indicators and first-generation status”, and Source C reports that among applicants in “the same academic score band”, those from “low-resourcing schools received scholarships at 1.6× the rate” of those from high-resourcing schools.
This is detailed, selective use of source material from more than one source in a single sentence. The paired evidence supports the claim that the equity mechanism exists and is having a measurable effect.
This suggests that the context weighting is doing what it is intended to do: correcting for structural disadvantage rather than treating unequal applicants as if they had identical opportunities.
This is thoughtful close reasoning on effect and purpose. The contrast between 'structural disadvantage' and 'identical opportunities' shows a nuanced grasp of what fairness can mean.
From a Digital Society perspective, this is important because fairness is not always the same as identical treatment.
This is an effective conceptual pivot from source detail to course thinking. The concise formulation gives the paragraph a memorable analytical insight worth imitating.
However,
This brief transition is highly effective because it signals a decisive turn to the counterargument. The structure stays balanced and easy to track.
Source A itself flags “missing data defaults” as one of the key “Accountability points”, and Source C shows that these defaults affected “12% of applications” and that these applicants’ “average rank dropped by 18 percentile points”.
This is strong evidential selection because it combines the source’s own framing with outcome data. The writer uses precise figures to show that the issue is systemic rather than marginal.
That is not a trivial technical matter; it is a substantial penalty.
This is crisp analytical judgment. The short, emphatic contrast translates statistics into significance and makes clear why the reader should care.
If applicants are disadvantaged not because they are less deserving but because they struggle to upload documents or navigate bureaucracy, then the system reproduces inequality under the appearance of objectivity.
This sentence is a strong argumentative move because it traces the mechanism from design choice to social consequence. The phrase 'under the appearance of objectivity' is especially well judged and sophisticated.
The appeal data strengthens this criticism.
This topic sentence links the new paragraph clearly to the one before it, so the argument builds rather than resets. It shows control over development of the case.
Source C shows appeals rising from “3% to 9%” and that “28% of appeals resulted in a score change”.
The response selects the two most telling appeal statistics and places them together effectively. That pairing lets the evidence imply both frequency of concern and meaningful correction.
This suggests the appeals system is not merely a safeguard for rare exceptions; it is correcting a significant number of outcomes.
This is strong inferential analysis: the writer draws a warranted conclusion from the figures and reshapes how the appeal process should be understood.
Source D is right to argue that accountability should not require students to become “expert auditors of a scoring system”.
This is a confident evaluative move that engages critically with a source rather than just reporting it. The quotation is integrated to sharpen the ethical point about burden and access.
Another problem is arbitrariness in the tie-break rule.
This sentence clearly introduces a fresh strand of criticism, helping the essay cover multiple dimensions of appropriateness. The naming of the issue upfront keeps the reasoning well organized.
Source A lists “tie-break rule” under “Apply constraints”, and Source C reveals that “alphabetical by surname when composite scores match” decided “7% of final offers in one faculty”.
This is precise, relevant evidence because it identifies both the rule and its real effect on outcomes. The detail makes the fairness concern tangible rather than hypothetical.
Even if this affects a minority of cases, it undermines legitimacy because surname order has no meaningful relationship to merit, need or context.
This analysis is effective because it weighs scale against principle. The writer acknowledges the limited scope but shows clearly why the issue still matters.
A consistently arbitrary rule is still arbitrary.
This compact phrasing is memorable and forceful. It distils the paragraph’s reasoning into a clear judgment, showing control of style as well as argument.
The human oversight described in the sources is also weaker than it first appears.
This opening sentence prepares the reader for a more probing evaluation of a supposed safeguard. It shows the argument is developing into deeper scrutiny, not just listing flaws.
Source A states, “The algorithm ranks; the committee approves,” and highlights a “human review threshold (top 5% + flagged cases)”.
This is well-selected evidence because it captures the formal oversight mechanism in the source’s own words. It gives the writer a concrete basis for testing whether human involvement is meaningful.
Human review is only meaningful if reviewers have the authority, time and willingness to question the system, not if they simply validate its outputs.
This is strong analytical unpacking of what 'human-in-the-loop' actually requires. The sentence moves from label to substance, which is exactly the kind of critical reading needed here.
Otherwise responsibility becomes blurred: the university can blame the algorithm, while the algorithm is said to be only implementing university policy.
This is an effective explanatory move because it shows the accountability gap created by weak oversight. The balanced clause structure makes the logic especially clear and fluent.
Overall, then, ScholarSort is appropriate in principle but only if stronger governance is attached to it.
This concluding synthesis returns to the central judgment and refines it in light of the full discussion. It pulls the essay together by balancing principle with conditions.
a high and successful appeal rate, and arbitrary alphabetical tie-breaks
These concrete details anchor the judgement that fairness problems are not hypothetical but already visible in outcomes and procedure.
For an algorithm allocating education funding to be appropriate, the university should publish how each component affects rankings, redesign defaults so missing data triggers review rather than automatic penalty, test regularly for disparate impact, and ensure appeals are simple and accessible.
The sentence turns evaluative claims into specific operational changes, showing how transparency, default design and monitoring would directly address the identified risks.