英语通过平行结构提示相同语法角色，也会省略可从前文恢复的成分。若错误匹配平行项，比较对象和论证关系都会被扭曲。
比较级究竟比较什么？省略后如何知道缺了哪个动词或名词？判断依据是结构对应和语义可比性。
than 后常省略与前文重复的谓语，但保留成分决定比较项。do so, one/ones, so, not 可替代前述结构，减少重复。The revised model predicts outcomes more accurately than the original model does.
比较两个模型预测结果的准确度；does 代替 predicts outcomes。
The salary of teachers is lower than that of engineers.
that 代替 salary，使工资与工资比较，避免把工资与工程师比较。
The policy reduced costs but increased waiting times.
共享主语的两个谓语平行，构成效果与代价的对照。
The study aims to describe the pattern, explain its cause, and test its stability.
三个原形动词共享 aims to，语法和研究动作层级一致。
Some participants completed the survey; others did not.
did not 省略 complete the survey，否定的是完成行为。
The urban sample was twice as large as the rural sample.
twice as large as 比较样本规模，表示前者是后者两倍。
The estimate is more precise, though not necessarily more accurate.
第二个表语省略 the estimate is；精确度与准确度被刻意区分。
What matters is how the policy was implemented, not merely whether it was adopted.
两个间接疑问从句平行，比较“如何实施”与“是否采纳”两个分析层级。
Online courses are easier to scale than small seminars are, but scalability is not the same as educational quality. A recorded lecture can reach more students, whereas a seminar can respond more directly to individual misunderstandings. The relevant question is therefore not which format is universally better, but which objective each format serves more effectively.
段落先比较可扩展性，再拒绝把该维度等同于质量，最后用 not...but... 重设比较问题。正确阅读需要始终追踪比较维度。
训练语境：学术科研
修正：The population of City A is larger than City B.
应写 larger than that of City B 或 City A has a larger population than City B。原句把 population 与 city 比较，类别不平行。
训练语境：学术科研
恢复省略：The second estimate is more stable than the first.
the first 代替 the first estimate，完整比较是第二个估计比第一个估计稳定。不能补成 first model，除非前文另有依据。
训练语境：学术科研
诊断平行性：The project sought to reduce costs, improving access, and better quality.
三项形式不平行。可改为 to reduce costs, improve access, and increase quality，让三个动词共享不定式标记并处于同一行动层级。
训练语境：学术科研
解释：Women responded more positively than men. 有何省略歧义？
通常恢复为 than men responded，比较两组回应；某些动词环境也可能产生“女性对男性回应更积极”的宾语解释。补出 did 可减少歧义。
训练语境：学术科研
比较 three times larger than 与 three times as large as。
前者在实际使用中可能被理解为三倍或增加三倍，存在歧义；three times as large as 明确总规模为三倍，定量写作更稳妥。
训练语境：学术科研
The model predicts employment better than wages 有何问题？
可能比较模型预测就业与预测工资的能力，也可能比较该模型与 wages 这个荒谬主体。应写 predicts employment better than it predicts wages 明确比较维度。
训练语境：真实生活
解释 I ordered the small one。
one 替代前文可数单数名词，例如 coffee 或 bag；small 保留区别属性。没有前文先行类别时，听者无法确定 one 指什么。
训练语境：真实生活
恢复：Maya can drive, but Leo can't.
can't 省略 drive，因为情态助动词保留并承载否定。它不表示 Leo 不能做所有事情，只否定语境中的驾驶能力。
训练语境：真实生活
修正：She likes hiking, to swim, and bicycles.
可改为 hiking, swimming, and cycling，让三项都是活动名词；或用三个不定式。平行不仅要求意思相关，也要求句法角色一致。
训练语境：真实生活
解释：The train is faster than the bus, but not cheaper.
第二分句恢复为 the train is not cheaper than the bus。两个比较维度分别是速度和价格，转折表明优势不延伸到第二维度。
The new curriculum improved students' performance more in tasks requiring explanation than in tasks requiring memorization. This pattern is less surprising than it first appears, because the intervention gave students more opportunities to compare strategies, revise claims, and justify conclusions. The gains were therefore not simply larger than those in the control group; they were larger in precisely the domain the curriculum was designed to affect.
less... than... 修正读者直觉。more in A than in B 比较的是同一课程对两类任务的影响。less surprising than it first appears 中省略了比较基准：“比最初看起来的那样更不意外”。compare strategies, revise claims, and justify conclusions 三个动词结构平行。those in the control group 中 those 代替 gains。高阶语境：学术科研
第一句比较的对象是什么？不要把它误读成哪一种比较？
它比较的是“同一新课程在解释型任务与记忆型任务中的提升幅度”。不要误读成“新课程和旧课程的总体比较”，因为句中比较结构是 more in tasks... than in tasks...。
高阶语境：学术科研
分析 compare strategies, revise claims, and justify conclusions 的平行性。
三个成分都是原形动词 + 名词宾语，且共同受 gave students more opportunities to 支配。平行结构让读者看出干预提供的是一组同类认知活动。
高阶语境：真实生活
写一句比较：某个 app 在“规划复杂行程”上比在“查附近餐厅”上更有用。
可以写：The app is more useful for planning complex trips than for finding nearby restaurants. 比较对象是两个用途，结构保持 more useful for A than for B 的平行。
遇到比较或省略，先把缺失成分补全，再检查比较双方是否同类、共享的结构是否真的适用于每一项。