## Punctuation ### Semicolons **When to use them** - **Link two closely related independent clauses** (two complete sentences) when a full stop feels too separating and a comma would be incorrect. - **Separate complex list items** when the list items already contain commas. - **Clarify contrast or cause-effect** between two complete clauses, often with transition words. **How to use them** Two independent clauses (no conjunction) - Pattern: `Sentence A; Sentence B.` - Example: "The sample size was small; consequently, uncertainty was high." Two independent clauses with a "connector" adverb - Pattern: `Sentence A; however, Sentence B.` - Example: "The model fit the training data well; however, performance degraded on the validation set." Common connector adverbs that work well after a semicolon: - however, therefore, consequently, nevertheless, moreover, thus, instead, similarly Lists with internal commas - Pattern: `item A, with detail; item B, with detail; item C, with detail` - Example: "Participants were recruited from three sources: local schools, which provided years 10–11; an online panel, which provided adults aged 18–65; and a community group, which provided older adults." ### Em dash **When to use it** - **Add a brief, high-impact interruption** (an aside) without the formality of brackets. - **Emphasise a clarification** at the end of a sentence. - **Mark a sharp pivot** or apposition (renaming a noun phrase) when commas would be ambiguous. Use sparingly in formal writing; overuse can make prose feel conversational. **How to use it** Keep formatting consistent: either "word—word" or "word — word" depending on your style guide; many academic styles prefer no spaces. If the aside is long or contains multiple clauses, use parentheses or a separate sentence. Insert an aside - Pattern: `Main clause—aside—continues.` - Example: "The second trial—conducted under higher humidity—produced larger measurement error." Add an end emphasis - Pattern: `Main clause—emphasis.` - Example: "The intervention improved short-term scores—but not long-term retention." Replace a colon for emphasis (rare) - Example: "The conclusion is straightforward—parameter tuning improved accuracy but reduced robustness." ### Commas Use commas to prevent ambiguity, not to add pauses. **How to use them** - After introductory clauses: "In contrast, the control group showed no change." - Around non-essential clauses: "The samples, which were stored at 4°C, were tested after 24 hours." - Before coordinating conjunctions joining two full clauses: "The method is simple, and it performs well under noise." ## Phrasing options ### Common word/phrase swaps | Informal | Academic | | ------------------- | --------------------------------------------------------------------------- | | a lot of | many; numerous; a substantial number of; a considerable amount of | | lots of | many; numerous; a large proportion of | | kind of / sort of | to some extent; relatively; in part; (often delete) | | basically | essentially; broadly; in essence; (often delete) | | big / huge | substantial; marked; considerable; (or quantify) | | small / tiny | modest; limited; slight; (or quantify) | | good | effective; robust; appropriate; beneficial (state criterion) | | bad | ineffective; limited; problematic; suboptimal (state criterion) | | better than | more effective than; superior to; outperforms (on metric X) | | worse than | less effective than; inferior to; underperforms (on metric X) | | get (obtain) | obtain; acquire; receive | | get (become) | become; increase; decrease; develop; evolve into | | get (understand) | determine; identify; establish; infer | | show (strong) | demonstrate; establish; confirm (use carefully) | | show (medium) | indicate; suggest; imply | | prove | support; provide evidence for; be consistent with | | stuff / things | factors; variables; components; materials; considerations; outcomes | | deal with | address; consider; examine; respond to | | talk about | discuss; examine; analyse; consider | | look at | examine; investigate; assess; evaluate | | figure out | determine; ascertain; identify | | bring up | raise; highlight; introduce | | go up | increase; rise; grow | | go down | decrease; decline; fall | | goes up and down | fluctuates; varies; oscillates | | a bunch of | a set of; a range of; a series of | | loads of | numerous; substantial | | pretty / really | relatively; notably; substantially (or delete) | | very | highly; substantially; markedly (or quantify) | | super | highly; exceptionally (rarely needed) | | obviously / clearly | (delete) or "as shown by …" | | I think | this suggests; it is plausible that; the evidence indicates | | in my opinion | based on the evidence; this analysis suggests | | sort of true | partially supported; supported to a limited extent | | it depends | the effect varies depending on; contingent on | | always | typically; often; generally; in most cases | | never | rarely; seldom; not observed in this sample | | huge difference | a substantial difference; a marked difference; (quantify) | | no difference | no statistically significant difference; no substantial difference observed | | weird / strange | unexpected; anomalous; counterintuitive | ### Common sentence structures | Sentence pattern | Academic equivalents | | ---------------------------------- | -------------------------------------------------------------- | | "This essay is about X." | "This paper examines X." / "This essay discusses X." | | "I’m going to look at X and Y." | "This study investigates X and Y." | | "I’ll talk about X first, then Y." | "First, X is considered; subsequently, Y is analysed." | | "By X, I mean…" | "Here, X is defined as…" | | "X is when…" | "X refers to…" / "X denotes…" | | "X is basically…" | "X can be described as…" | | "This happens because…" | "This occurs because…" / "This can be attributed to…" | | "That’s why…" | "This explains why…" / "Therefore…" | | "So…" | "Therefore…" / "Consequently…" / "Thus…" | | "But…" | "However…" / "Nevertheless…" | | "On the other hand…" | "By contrast…" / "Conversely…" | | "Even though…" | "Although…" / "Despite…" | | "X causes Y." | "X leads to Y under Z conditions." / "X is associated with Y." | | "This proves X." | "These findings support X." / "This is consistent with X." | | "Research shows…" | "Previous studies have found…" / "The literature indicates…" | | "They say…" | "The authors report…" / "The study concludes…" | | "This matches other studies." | "This aligns with prior findings." | | "This goes against…" | "This contrasts with…" / "This is inconsistent with…" | | "It went up a lot." | "It increased by …%." / "A marked increase was observed." | | "It stayed about the same." | "No substantial change was observed." | | "Most of them…" | "The majority of…" | | "Only a few…" | "A small minority of…" | | "This means…" | "This suggests/indicates…" | | "This could be due to…" | "This may be attributed to…" / "A possible explanation is…" | | "I’m not sure." | "The evidence is inconclusive." | | "We can’t say for sure." | "Causal inference cannot be established from these data." | | "There are some issues with this." | "This approach has several limitations." | | "This isn’t perfect." | "This method is subject to error due to…" | | "Some people might say…" | "A potential objection is…" | | "Still…" | "Nonetheless…" / "Nevertheless…" | | "This graph shows…" | "Figure X illustrates…" / "Figure X shows…" | | "You can see that…" | "As shown in Figure X…" | | "The table has…" | "Table X summarises…" | ### Claim strength ladder | Strength | Use when… | Safer academic phrasing | |---|---|---| | High certainty | strong evidence, direct measurement, robust design | demonstrate; establish; confirm; show | | Medium | evidence points one way but not definitive | indicate; suggest; imply; consistent with | | Low / tentative | limited data, plausible mechanism, exploratory | may; might; could; it is possible that; it is plausible that | | No causal claim | correlational data only | associated with; linked to; correlated with | | Causal (qualified) | causal design or strong justification | leads to; results in; causes (under X conditions) | ### Transition and signposting alternatives | Purpose | Spoken / basic | Academic options | | ----------- | -------------- | ---------------------------------------- | | Add | and; also | additionally; furthermore; moreover | | Contrast | but | however; nevertheless; nonetheless | | Compare | like; same as | similarly; likewise; in a similar manner | | Cause | because | since; given that; due to | | Effect | so | therefore; thus; consequently | | Example | for example | for instance; specifically; notably | | Clarify | I mean | that is; in other words | | Limit scope | not about | beyond the scope; outside the remit | | Sequence | first/then | first; subsequently; finally | ### Specific examples | Informal | Academic | | -------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | A study said that exercise is good for your mood. | A meta-analysis reports that regular physical activity is associated with improved mood and reduced depressive symptoms, although effect sizes vary by population and intervention type (Smith et al., 2021). | | Another paper showed that students learn better with practice tests. | Retrieval practice has been shown to improve long-term retention relative to re-reading, consistent with prior experimental findings (Jones & Patel, 2019). | | The researchers found that caffeine makes people more alert. | The authors report increased vigilance following caffeine intake, with the largest effects observed during sustained-attention tasks (Lee et al., 2020). | | One study said that social media causes anxiety. | One observational study reported an association between high social media use and anxiety symptoms; however, the design does not establish causality (Garcia et al., 2018). | | The paper proves that pollution harms health. | The study provides evidence linking higher particulate exposure to worse respiratory outcomes, aligning with broader epidemiological literature (Nguyen et al., 2022). | | A study said that wearing masks stopped the virus spreading. | The authors found lower transmission rates in settings with mask use, though effectiveness depended on compliance and local context (Chen et al., 2021). | | The research shows the new drug works. | In a randomised controlled trial, the treatment group showed improved clinical outcomes relative to placebo, suggesting therapeutic efficacy under the study conditions (Ahmed et al., 2020). | | The new medicine is safe because no one had problems. | No serious adverse events were reported during the trial period; however, rare harms may not be detectable without larger samples or longer follow-up (Ahmed et al., 2020). | | The study says the program helped people stop smoking. | Participants receiving the intervention exhibited higher cessation rates than controls, indicating a measurable treatment effect (Brown & Kim, 2017). | | The results were basically the same as last year. | The outcomes were broadly consistent with the previous year, with only minor variation within expected uncertainty bounds. | | The data is messy so it’s hard to tell what’s going on. | The dataset contains substantial noise and heterogeneity, complicating inference and motivating the use of robust statistical checks. | | Some of the measurements seem wrong. | Several measurements appear anomalous, warranting verification for instrument error, recording mistakes, or out-of-range operating conditions. | | The graph clearly shows that X caused Y. | The graph shows a temporal co-occurrence of X and Y; however, causal attribution requires stronger design assumptions and controls. | | The survey proves what people think. | The survey provides self-reported attitudes at a specific time point, which may be influenced by response bias and question framing. | | People were happier after the change. | Reported wellbeing increased following the intervention period, although confounding factors cannot be ruled out. | | The interview answers were interesting. | The interview responses revealed recurring themes, providing qualitative insight into participant experiences and perceived constraints. | | The study says the sample represents everyone. | The authors argue the sample is broadly representative; nevertheless, selection effects may limit generalisability to other populations. | | We didn’t have enough people to be sure. | The small sample size reduces statistical power and increases uncertainty around estimated effects. | | The results might be a coincidence. | The observed pattern may reflect random variation, particularly given multiple comparisons and limited statistical power. | | There was a big difference between the groups. | The groups differed substantially on the primary outcome measure, with a large between-group effect. | | The difference wasn’t that big. | The between-group difference was modest and may be of limited practical significance despite statistical detectability. | | The change was important. | The change is practically significant because it exceeds the threshold typically considered meaningful for the outcome. | | The results were surprising. | The results were unexpected relative to the stated hypothesis, suggesting either unaccounted mechanisms or measurement artefacts. | | We think the reason is because of stress. | A plausible explanation is stress-related behavioural change, though alternative mechanisms remain possible. | | The model isn’t perfect but it’s good enough. | While the model is simplified, it captures the dominant dynamics sufficiently for the intended analysis. | | The simulation matches reality well. | The simulation reproduces key observed trends, supporting its validity for the examined scenarios. | | The model failed when conditions changed. | Model performance degraded under distribution shift, indicating limited robustness outside the calibration regime. | | The algorithm is fast. | The algorithm exhibits low runtime complexity and performs efficiently at the tested scale. | | The algorithm is slow when the dataset is large. | Runtime increases sharply with dataset size, indicating scalability constraints for large-scale deployment. | | The AI is a black box. | The model’s decision-making process is not readily interpretable, limiting transparency and error attribution. | | The new app is easy to use. | Usability testing indicates low task completion time and reduced error rates, suggesting a low cognitive burden for users. | | Users liked the redesign. | User feedback shows improved satisfaction ratings, indicating increased perceived usability and clarity. | | The website change increased sales. | Sales increased after the redesign, though attributing the change to the redesign alone requires controlling for seasonality and marketing effects. | | The new teaching method made grades better. | Students taught with the new method achieved higher assessment scores than the comparison group, consistent with an instructional benefit (Clark et al., 2019). | | The classroom was more engaged. | Observational measures suggest higher behavioural engagement, though measurement subjectivity should be considered. | | The experiment worked because we followed the steps. | Procedural adherence reduced variability and improved reproducibility across trials. | | We did the test three times to be safe. | Measurements were repeated to assess repeatability and reduce random error. | | The equipment wasn’t accurate. | Instrument precision limitations likely contributed to measurement uncertainty and wider confidence intervals. | | The results are reliable because we repeated them. | Repeated trials produced consistent estimates, supporting reliability and reducing the likelihood of random error driving the findings. | | The conclusion is obvious from the results. | The conclusion follows from the results under the stated assumptions, though alternative interpretations remain feasible. | | The paper argues that inequality is getting worse. | The authors argue that inequality has increased over time, supported by longitudinal income distribution data (Davis, 2016). | | The study says minimum wage helps families. | The study reports improved household income among minimum wage recipients, though employment effects were context-dependent (Hernandez & Liu, 2020). | | A study found crime went down after more street lights. | The authors report reduced reported crime following lighting improvements, consistent with deterrence-based explanations (Reed et al., 2018). | | The policy failed because people didn’t follow it. | Implementation fidelity was low, suggesting that non-compliance may have undermined the policy’s effectiveness. | | The rule change made things unfair. | The rule change altered incentives and produced uneven impacts across groups, raising distributional equity concerns. | | The article says remote work is better. | The article reports higher self-reported productivity under remote work for some roles, though outcomes depended on job type and home environment (Khan et al., 2021). | | The climate is changing because humans pollute. | Multiple lines of evidence attribute recent warming primarily to anthropogenic greenhouse gas emissions (IPCC, 2021). | | One paper said forests store lots of carbon. | The study estimates substantial carbon sequestration in forest biomass, highlighting forests’ role in carbon cycling (O’Neill et al., 2019). | | The results show biodiversity is dropping. | The findings indicate declining biodiversity indices over the measured period, consistent with habitat loss pressures (White et al., 2017). | | The new material is stronger than the old one. | Mechanical testing indicates higher tensile strength for the new material relative to the baseline, suggesting improved load capacity. | | The design is safer because it has more supports. | Additional supports redistribute loads and reduce peak stress concentrations, improving structural safety margins. | | The bridge cracked because the load was too high. | Failure likely occurred due to stress exceeding material limits, leading to crack initiation and propagation under load. | | The battery died quickly in cold weather. | Battery performance degraded at low temperature, consistent with reduced reaction kinetics and increased internal resistance. | | The circuit didn’t work because the voltage was too low. | The circuit failed to operate because supply voltage fell below the minimum required threshold for stable function. | | The sensor readings jumped around a lot. | The sensor output exhibited high variance, suggesting noise, interference, or unstable calibration. | | The robot kept missing the target. | The robot showed systematic positional error, indicating calibration drift or deficiencies in the control loop. | | The new filter removed most of the noise. | Filtering reduced high-frequency components and improved signal-to-noise ratio, enhancing measurement stability. | | The code has bugs so the results might be wrong. | Implementation defects may have biased results, underscoring the need for validation tests and reproducible pipelines. | | We changed the settings until it worked. | Parameters were tuned iteratively until performance criteria were met, though this introduces a risk of overfitting to the test cases. | | The user study shows people prefer option A. | In the user study, participants reported higher preference ratings for option A, indicating stronger perceived usability. | | Most people answered the same way. | Responses clustered around a common option, suggesting consensus or a framing effect. | | Some people didn’t answer honestly. | Social desirability bias may have influenced responses, leading to systematic under- or over-reporting. | | The article says young people are more stressed now. | The article reports higher self-reported stress among younger respondents, though cohort and reporting effects may contribute (Taylor, 2020). | | This supports the theory we learned in class. | This pattern is consistent with the proposed theoretical mechanism and reinforces the predicted relationship between variables. | | This goes against the theory, so the theory is wrong. | The findings diverge from the theory’s predictions, suggesting boundary conditions, missing variables, or measurement differences rather than immediate falsification. | | We should do more research to be sure. | Further research is needed to replicate the findings, evaluate alternative explanations, and test robustness across contexts. | | This method is cheap and easy. | The method is low-cost and operationally straightforward, reducing barriers to adoption in resource-constrained settings. | | The method is not practical in real life. | Practical deployment is constrained by cost, infrastructure requirements, and operational complexity, limiting real-world feasibility. | | The study is biased because it was funded by a company. | The funding source may present a conflict of interest, warranting careful scrutiny of methodology, reporting choices, and transparency. | | The study is trustworthy because it was peer reviewed. | Peer review increases credibility but does not guarantee validity; methodological quality and reproducibility remain essential considerations. | ## Structure ### Abstract | Sentence type | Purpose | Template with placeholders | | ---------------------------- | ----------------------------------------- | ------------------------------------------------------------------------------------------ | | Context + problem | Situate topic; name the gap/problem | "**X** is important for **Y**, yet **Z** remains a challenge." | | Aim / research question | State what you did | "This study **examines/investigates/evaluates** **X** to determine **Y**." | | Method (minimal) | How you did it (1 sentence) | "A **method/design** was used, involving **A**, **B**, and **C** (n = **N**)." | | Key results (specific) | Main findings with numbers if possible | "Results show **A** increased/decreased by **N%**, while **B** remained **…**." | | Interpretation / implication | What results mean (not a full discussion) | "These findings suggest **mechanism/explanation**." | | Conclusion / value | Why it matters | "Overall, the study indicates **implication**, supporting **recommendation/application**." | **Context + problem** - "**X** plays a central role in **Y**; however, **Z** limits **outcome**." - "Despite advances in **X**, **Z** persists, particularly in **context/population**." - "Current approaches to **X** often fail to **Y**, resulting in **Z**." **Aim / question** - "This paper **examines** **X** with respect to **Y**." - "The aim of this study is to **evaluate** whether **X** improves **Y**." - "This work addresses the question: **To what extent does X affect Y?**" **Method** - "A **cross-sectional/experimental/simulation** approach was employed using **dataset/tool**." - "Data were collected from **source** and analysed using **technique/model**." - "The analysis combines **A** with **B** to estimate **C**." **Results** - "**X** was associated with **Y** (Δ = **…**, p = **…**)." - "The intervention reduced **Y** by **…** relative to **baseline/control**." - "Across **N** cases, **Y** increased from **A** to **B**." **Interpretation** - "This suggests that **mechanism** is a primary driver of **outcome**." - "A likely explanation is **…**, given **…**." - "Taken together, the findings indicate **pattern** consistent with **theory**." **Conclusion / value** - "Overall, **X** appears to be **effective/limited** for **Y** under **conditions**." - "These results support **recommendation**, particularly for **context**." - "The study contributes evidence that **X** can **Y**, informing **decision/practice**." ### Introduction | Sentence type | Purpose | Template | | --------------------------- | ------------------------------- | ------------------------------------------------------------------------ | | Broad context | Establish topic area | "**X** has become increasingly important due to **Y**." | | Narrow to problem | Identify specific issue | "However, **Z** remains unresolved in **context**." | | Evidence / what is known | Brief background with citations | "Prior work suggests **A**, yet findings on **B** are mixed." | | Gap | What’s missing / why needed | "Existing studies rarely address **C** in **context**, leaving **gap**." | | Aim / question | What your report does | "This report **evaluates** **X** by **method** to determine **Y**." | | Approach overview | High-level method | "The analysis uses **data/model/experiment** and compares **A vs B**." | | Contribution / significance | Why your work matters | "This provides **insight/benchmark/framework** for **stakeholder**." | | Roadmap (optional) | Guide reader | "Section 2 outlines **…**; Section 3 presents **…**." | **Broad context** - "In recent years, **X** has emerged as a key factor in **Y**." - "**X** is widely used in **context**, affecting **stakeholders/outcomes**." - "Improving **X** is central to achieving **Y**." **Narrow to problem** - "Despite this, **problem Z** continues to constrain **outcome**." - "A persistent challenge is **Z**, particularly in **setting**." - "However, current implementations often fail to **Y**, resulting in **Z**." **What is known** - "Previous studies report **A**, indicating **B**." - "The literature broadly agrees that **A**; however, **B** remains contested." - "Evidence suggests **A** is associated with **B**, though results vary by **C**." **Gap** - "Nevertheless, there is limited evidence on **X** in **context**." - "Few studies have examined **X** while accounting for **Y**." - "This leaves an open question regarding **mechanism/condition**." **Aim / question** - "This study investigates whether **X** improves **Y**." - "The purpose of this report is to assess **X** using **method**." - "Accordingly, the research question is: **How does X affect Y in context Z?**" **Approach overview** - "To address this, **data** from **source** were analysed using **technique**." - "A **model/simulation/experiment** was conducted, varying **X** and measuring **Y**." - "The approach compares **A** and **B** under controlled **conditions**." **Contribution / significance** - "This analysis clarifies **relationship** between **X** and **Y**." - "The findings provide practical guidance for **stakeholder**." - "This work offers a baseline for future evaluation of **X**." **Roadmap** - "The remainder of this report is organised as follows: **…**" - "Section 2 describes **…**, followed by **…** in Section 3." - "The discussion concludes with implications for **…**." ### Conclusion | Sentence type | Purpose | Template | | ---------------------- | -------------------------- | ---------------------------------------------------------------------- | | Restate aim (brief) | Remind reader what you did | "This report examined **X** to determine **Y**." | | Main findings (ranked) | 2–4 key results | "The main findings were: (1) **A**; (2) **B**; (3) **C**." | | Answer the question | Direct conclusion | "Overall, **X** **does/does not** improve **Y** under **conditions**." | | Implications | Why it matters | "This implies **recommendation/impact** for **stakeholder**." | | Limitations | Bound claims | "However, the findings are limited by **L1** and **L2**." | | Future work | Next steps | "Future work should **A** and **B** to test **C**." | **Restate aim** - "This study evaluated **X** with respect to **Y**." - "The aim was to assess whether **X** influences **Y**." - "This report investigated **X** in **context Z**." **Main findings** - "Three findings stand out. First, **…** Second, **…** Finally, **…**" - "The results show **A**, while **B** suggests **…**." - "Across the dataset, **A** was observed, with **B** occurring primarily when **C**." **Final claim** | Evidence strength | Sentence structure examples | |---|---| | Strong support | "Overall, the results **demonstrate** that **X** improves **Y** by **N%** under **conditions**." | | Moderate support | "Overall, the results **suggest** that **X** is likely to improve **Y**, although effects vary by **Z**." | | Mixed / conditional | "In summary, **X** improves **Y** only when **condition**, whereas under **other condition** the effect is limited." | | No effect | "Overall, no substantial improvement in **Y** was observed following **X**." | | Opposite effect | "Contrary to expectations, **X** was associated with a reduction in **Y**." | | Inconclusive | "The evidence is inconclusive regarding the effect of **X** on **Y**, due to **limitations**." | **Implications** - "This indicates that **stakeholders** should prioritise **action** when aiming to **goal**." - "A key implication is that **mechanism** may be more influential than **factor**." - "These findings support the use of **X** as a strategy for **Y** in **context**." **Limitations** - "Interpretation is limited by **sample size/measurement uncertainty**." - "A limitation of this study is **…**, which may bias **…**." - "These results may not generalise to **population/context** because **…**." **Future work** - "Future studies should replicate this analysis using **larger/more diverse** samples." - "Further work could test **mechanism** by varying **X** while controlling **Y**." - "Subsequent research should compare **approach A** and **approach B** over **timeframe**." ## Mini Examples ### Abstract (6 sentences) 1. "**X** is important for **Y**, yet **Z** remains a persistent challenge in **context**." 2. "This study investigates **X** to determine its effect on **Y**." 3. "A **method** was used, involving **A**, **B**, and **C** (n = **N**)." 4. "Results indicate **A** increased by **N%**, while **B** decreased from **P** to **Q**." 5. "These findings suggest **mechanism/explanation**." 6. "Overall, the study supports **implication/recommendation** for **stakeholder**." ### Introduction (7 sentences) 1. "In recent years, **X** has become increasingly important due to **Y**." 2. "Despite this, **Z** continues to constrain **outcome**, particularly in **context**." 3. "Previous studies report **A**; however, findings regarding **B** remain mixed." 4. "Consequently, there is limited evidence on **C** in **context**, leaving **gap**." 5. "This report evaluates **X** to determine **Y**." 6. "To address this, **data/method** from **source** were analysed using **technique**." 7. "The remainder of this report is organised as follows: **…**" ### Conclusion (6 sentences, "mixed results") 1. "This report examined **X** to determine its effect on **Y**." 2. "The findings show that **A** occurs consistently when **condition 1** is met." 3. "However, under **condition 2**, the effect on **Y** is modest." 4. "Overall, **X** appears most effective for **Y** in **context**, provided that **condition** is satisfied." 5. "Interpretation is limited by **L1** and **L2**, which may influence **…**." 6. "Future work should test **A** over **longer periods/larger samples** to clarify **B**."