The "Source-based Paper": Dominating Paper 3
Paper 3 gives you four source-based questions with quantitative and qualitative findings, drawn from one of the topics of technology, motivation, or culture, to establish one claim — one that you have never seen before. Your job is to analyse these sources. The good news: the questions are almost always the same structure.
If you are taking Paper 3, it means you are a HL student. Unlike other papers, you don't need to memorise dozens of studies — but if you want a 7, you need to ace Paper 3. It carries a massive 30% weighting, the highest of any single assessment. Compare this to the 20% weighting for Paper 1, Paper 2, and your Internal Assessment (IA).
Total time: 1 hour 45 minutes
Marks
30 marks
Duration
1h 45 min
SL weight
—
HL weight
30%
Questions
4 compulsory questions on an unseen source-based study (Q1: 3 marks, Q2: 6 marks, Q3: 6 marks, Q4: 15 marks)
Four structured questions based on an unseen source-based study (with a resource booklet). Q1 identifies a research method, Q2 analyses quantitative data, Q3 evaluates credibility or methodology, and Q4 is a 15-mark synthesis essay. This paper tests your ability to read and critique sources you have never seen, drawing on research methods, statistics, sampling, ethics, and validity.
Prep strategy
Revise all qualitative research methods (interviews, observations, case studies), validity types, sampling strategies, and ethical considerations. Practise reading short studies and identifying methodological strengths and weaknesses.
Last reviewed: 28 April 2026
Specimen Paper Claim
"The use of technology in education may have a negative effect on the mental health of students."
The specimen paper uses this claim across all five sources. Source 1 (a bar chart on work-life balance and mental health) appears in the exam paper itself. Sources 2–5 are in the separate resource booklet. All four questions relate to this single claim.
The Four Questions: A Breakdown
Explain one issue that limits the interpretation of the data in Source 1.
Specimen Source 1 — Bar chart (OwlLite, n.d.)
Shows the effect of healthy vs unhealthy work-life balance on student mental health (anxiety and depression).
Unhealthy: Anxiety 56%, Depression 55% | Healthy: Anxiety 24%, Depression 21%
A relevant issue is explained.
Teacher's Tip: Think about what we cannot tell and what we are uncertain of in the graph. Look for missing information like sample sizes, unclear definitions of variables (e.g., what exactly defines a "healthy" work-life balance?), or a limited y-axis range that distorts the data. Show the examiner that you know what other data are important. A relevant issue is explained, not merely described. Only one issue is credited — don't list multiple.
Markscheme — Acceptable issues include:
- Title doesn't specify the role of technology (many other factors could explain the data)
- Limited y-axis range distorts the visual representation of differences
- Unclear how "healthy" and "unhealthy" work-life balance was operationalised
- Unclear what percentage of students have both anxiety and depression (comorbidity)
- Sample size unknown — affects generalisability / population validity
Sample Answer
A major limit in Source 1 is the lack of a clear definition for "healthy" and "unhealthy" work-life balance. Without knowing how these categories were operationalised — for example, whether they were self-reported or measured through objective criteria — we cannot be certain that the data accurately reflects the variable described in the title. This makes the results difficult to interpret in relation to the claim about technology and mental health.
Analyse the findings from Source 2 and state a conclusion linked to the claim that X.
Specimen Source 2 — Experiment (N=50, 25 per group)
Group 1: online task for up to 6 hours | Group 2: no computer use for the whole day
Mental health test at end (max score 30 = high mental health)
| Condition | Max | Median | Mean | SD | Min |
|---|---|---|---|---|---|
| High screen time | 21 | 10 | 11.6 | 3.87 | 7 |
| No screen time | 25 | 15 | 17.64 | 4.48 | 12 |
Box plot medians: High screen time (7, 9, 10, 14, 21) | No screen time (12, 14, 15, 21, 25)
Data is analysed in detail; conclusion is explicitly linked to the findings and the claim.
Teacher's Tip: In your analysis, use numbers and specific statistical terms. Discuss the Mean (lower/higher performance), Standard Deviation (consistency vs. variability), and Box Plot Overlap (no overlap often suggests a significant difference). Note whether data is skewed. Then, add a concluding statement for this study itself, and another concluding line regarding the specific claim in the question.
Sample Answer
Source 2 shows that the mean mental health score for the high screen time group (M = 11.6, SD = 3.87) was considerably lower than the no screen time group (M = 17.64, SD = 4.48), suggesting that students who used computers for up to 6 hours had poorer mental health outcomes. The medians (10 vs. 15) reinforce this pattern. The box plots do not overlap — the maximum of the high screen time group (21) equals the median of the no screen time group — indicating a meaningful difference between conditions. The slightly larger SD in the no screen time group suggests greater variability in that condition. The data also appears right-skewed in the high screen time group (mean > median), indicating a small number of higher scorers pulling the mean upward. In conclusion, students who engaged in 6 hours of screen time were more likely to report lower mental health scores, which supports the claim that technology use in education may negatively affect student mental health.
Answer the one provided. Three possible questions:
Specimen Source 3 — Qualitative focus groups
Focus groups with students from 4 different schools on how online learning during COVID-19 negatively affected mental health. Inductive content analysis on transcripts identified three themes: isolation from friends (anxiety, loneliness), extensive time with parents (arguing, feeling misunderstood), and online learning feeling "unreal" and "detached."
Detailed understanding of the research consideration; reference to the source is explicit.
1. Discuss how the researcher could improve the credibility of the findings.
- Qualitative: credibility = trustworthiness
- Quantitative: internal validity
Markscheme — Credibility improvements may include:
- Member checking — participants review and confirm the researcher's interpretations
- Triangulation — combining focus group data with other methods (e.g., interviews, surveys)
- External audit — colleagues or experts review the methods and findings
- Reflexivity — researcher is transparent about potential conflicts of interest or biases
- Multiple researchers collecting or interpreting data to reduce single-researcher bias
Sample Answer
Credibility refers to the trustworthiness of qualitative findings — the extent to which the data accurately represents participants' experiences. In Source 3, the researcher used inductive content analysis on focus group transcripts to identify three themes. However, credibility is limited by the absence of member checking: participants were not given the opportunity to review the researcher's interpretations of their responses, which reduces confidence that the themes identified truly reflect their intended meanings. To improve credibility, the researcher could have returned the coded themes to participants and asked whether they accurately captured their experiences. Additionally, using triangulation — for example, combining the focus group data with individual interviews or a survey — would allow the researcher to cross-check findings across methods, strengthening the trustworthiness of the conclusions.
2. Discuss how the researcher could transfer findings to other populations or contexts.
- Qualitative: transferability
- Quantitative: generalisability = sample representativeness
Sample Answer
Transferability refers to the extent to which qualitative findings can be applied to other contexts or populations. In Source 3, focus groups were conducted with students from 4 different schools, but the study does not describe the demographic characteristics of participants or the cultural context in which the schools are located. This limits the transferability of the findings to students in different cultural or socioeconomic settings. To improve transferability, the researcher could have provided thick descriptions — detailed accounts of the research context, participants, and setting — so that readers can judge for themselves whether the findings are applicable to their own context. Additionally, using purposive sampling to deliberately include students from diverse backgrounds (e.g., different age groups, urban vs. rural schools) would strengthen the range of contexts to which the findings could reasonably be transferred.
3. Discuss how the researcher could avoid bias of the findings.
- Qualitative: more on researcher's bias to reduce participant's bias
- Quantitative: control confounding variables to minimise bias
Sample Answer
Researcher bias occurs when the researcher's own expectations or assumptions influence the collection or interpretation of data. In Source 3, the focus groups were conducted on the topic of how online learning negatively affected mental health — this framing may have led the researcher to ask leading questions, inadvertently encouraging participants to respond in ways that confirmed the claim. To reduce this bias, the researcher could have used inter-rater reliability — having a second independent coder analyse the transcripts — and compared their coding to ensure themes were not shaped by a single researcher's perspective. Using a semi-structured interview protocol with pre-tested, neutral questions would also help minimise the influence of researcher expectations on participant responses.
Teacher's Tip: Always begin with a short definition of the term — this signals to the examiner that you understand the question. Then discuss 2–3 specific points with explicit reference to the source. The top markband requires detailed understanding with explicit source reference.
Discuss the validity of the claim illustrated by at least three sources.
Specimen Sources 4 & 5 (for Q4)
Source 4 — Correlation study: mental health and hours of homework on laptops. Correlation coefficients: Cognitive dulling (r = 0.25), Depression (r = 0.76), General anxiety (r = 0.80), OCD (r = 0.12).
Source 5 — Line graph: mental health scores vs hours of continuous screen time (0–10 hours) for gaming, streaming, and smartphone use. All three show declining mental health as screen time increases; gaming drops most sharply after 2 hours.
| Marks | Descriptor |
|---|---|
| 0 | Does not reach the standard described by the descriptors below. |
| 1–3 | Little understanding of the demands; mostly descriptive; little or no discussion of different points of view; superficial conclusion. |
| 4–6 | Some understanding; limited analysis; mostly descriptive; simplistic conclusion. |
| 7–9 | Understanding partially addressed; analysis lacks development; uses at least 2 sources; partial conclusion. |
| 10–12 | Demands understood and addressed; critical analysis but lacks development; uses 2+ sources; consistent conclusion. |
| 13–15 | Demands fully understood; well-developed critical analysis; uses 3+ sources effectively; different points of view evaluated; reasoned conclusion. |
Teacher's Tip: Synthesis is Key — don't just summarise. Compare the sources: is the evidence consistent or contradictory? Use the GRAVE Approach to structure your evaluation: Generalizability, Reliability, Alternative explanations, Validity, and Ethics. You must use at least three sources and evaluate different points of view to reach the top markband.
Sample Answer
The claim that the use of technology in education may have a negative effect on the mental health of students is supported by multiple sources, though the strength and nature of this relationship varies across the evidence. Source 2, an experiment with 50 students, found that those who completed an online task for up to 6 hours had a lower mean mental health score (M = 11.6) compared to those with no screen time (M = 17.64), with non-overlapping box plots suggesting a meaningful difference. This provides relatively strong evidence for the claim, as the experimental design allows for some degree of causal inference. However, the small sample size (N = 50) limits the generalisability of these findings to the broader student population.
Source 4, a correlation study, found strong positive correlations between laptop homework hours and both depression (r = 0.76) and general anxiety (r = 0.80), further supporting the claim. The high correlation coefficients suggest a robust relationship. However, correlation does not imply causation — it is possible that students with pre-existing mental health difficulties are more likely to spend longer hours on laptops, representing a confounding variable that limits the validity of the claim.
Source 5 adds nuance: the line graph shows that all three types of screen use (gaming, streaming, smartphone) are associated with declining mental health scores as hours increase, but the pattern differs — gaming shows a sharp decline after 2 hours, while streaming declines more gradually. This suggests that the type of technology use matters, not just the quantity, which complicates a straightforward reading of the claim.
In conclusion, the evidence across Sources 2, 4, and 5 collectively supports the claim that technology use is associated with poorer mental health outcomes in students. However, the relationship is only partially valid as stated: methodological limitations (small samples, correlational designs) and the differential effects of different types of screen use mean that a blanket claim about "technology in education" oversimplifies a more complex picture.