RESM711 Advanced Social Science Research Methods (UNISA) Exam Notes and Study Guide

RESM711 Advanced Social Science Research Methods is a core postgraduate research methods module in the University of South Africa’s Masters in Research Psychology stream, and it demands far more than memorising definitions. The module requires a working command of research design, epistemology, sampling, measurement, data analysis, ethics, and the logic of inference in social science research. Strong exam performance depends on understanding how these parts fit together in real studies, how to defend methodological choices, and how to critique research with precision.

1. The logic of advanced social science research

Advanced social science research is built on the idea that social reality is complex, layered, and partly constructed through meanings, institutions, power relations, and historical context. Unlike simple technical exercises, research in this field requires constant judgement about what counts as evidence, how evidence is produced, and what kinds of claims can honestly be made from that evidence. In RESM711, a recurring expectation is that the student can move beyond descriptive statements and explain why a method is appropriate for a specific research problem.

1.1 What makes social science research “advanced”

The word advanced does not merely refer to difficult vocabulary or more complicated statistics. It refers to a deeper level of methodological reasoning. At this level, the researcher must weigh competing assumptions about knowledge, reality, causation, and interpretation. For example, a study on student burnout at a South African university could be approached as a quantitative survey measuring stress scores, a qualitative interview study exploring lived experiences, or a mixed methods design that combines both. An advanced researcher must be able to justify which approach best addresses the research question, what trade-offs are accepted, and what limitations remain.

Advanced research also involves stronger attention to:

the relationship between theory and empirical evidence
the fit between question, design, and analysis
the danger of overclaiming from weak data
the role of context in shaping findings
the transparency of the entire research process

A common exam task is to evaluate a proposed study and identify whether the design, sampling, instrument, or analysis is logically aligned with the stated objective. That requires more than knowing the names of methods; it requires understanding the logic of methodological fit.

1.2 Ontology, epistemology, and methodology

Three terms repeatedly appear in advanced research methods and should be clearly distinguished.

Ontology concerns what is assumed to exist. In social science, the central ontological issue is whether social phenomena are understood as objective entities existing independently of people, or as realities shaped through social interaction, language, and historical processes.

Epistemology concerns how knowledge is known. It asks what counts as valid evidence, how researchers can justify claims, and whether knowledge is discovered, interpreted, or constructed.

Methodology is the broader strategy that connects epistemological assumptions to concrete methods. It explains why a researcher chooses a survey, interview, ethnography, experiment, content analysis, or a combination of these.

A useful comparison is shown below.

Concept	Core question	Example in a study of school discipline
Ontology	What exists?	Is discipline a measurable institutional practice, a negotiated social process, or both?
Epistemology	How can it be known?	Can discipline be measured through incident records, or must it be interpreted through participant meanings?
Methodology	How should it be studied?	Should the researcher use quantitative records analysis, qualitative interviews, or mixed methods?

Students often lose marks by using these terms interchangeably. In an exam answer, strong performance comes from showing how they relate: ontology influences epistemology, and epistemology shapes methodology.

1.3 Paradigms in social science research

A paradigm is a general worldview that guides how a researcher understands reality and knowledge. The most commonly examined paradigms include positivism, post-positivism, interpretivism, critical theory, and pragmatism.

Positivism

Positivism assumes that reality is objective, measurable, and governed by regularities that can be discovered through systematic observation. In a strict positivist view, the goal is to identify causal relationships using controlled, replicable procedures. Social phenomena are treated as variables that can be measured and compared.

Post-positivism

Post-positivism accepts that objective knowledge is possible but imperfect. It recognises that observation is fallible, bias is unavoidable, and findings are always tentative. This is one of the most common paradigmatic positions in contemporary quantitative social science because it retains a commitment to empirical testing while acknowledging uncertainty.

Interpretivism

Interpretivism argues that social reality is meaningful and must be understood from the perspectives of participants. Here the aim is not to discover universal laws but to interpret meanings, experiences, and contexts. Interviews, focus groups, and ethnography are often associated with this paradigm.

Critical theory

Critical theory focuses on power, inequality, ideology, and emancipation. It does not treat social research as neutral; instead, it asks whose interests are served by existing structures and how research might expose or challenge domination. This paradigm is relevant to studies of gender, race, class, labour, coloniality, and institutional injustice.

Pragmatism

Pragmatism is guided by the practical usefulness of knowledge. Rather than committing rigidly to one worldview, pragmatists choose methods that best answer the problem. This paradigm is often used to justify mixed methods research because it values what works for the research purpose.

A concise exam strategy is to link the paradigm to the research question:

If the question asks “How many?”, “How much?”, or “Does X affect Y?”, quantitative post-positivist reasoning is often appropriate.
If the question asks “How do people understand?” or “What does this experience mean?”, interpretivism is more appropriate.
If the question asks “How do structures of power produce this outcome?”, critical approaches are often relevant.
If the question asks a practical intervention question, pragmatism may be the best fit.

1.4 Theory, concepts, and variables

A research study is stronger when it moves logically from theory to concepts to variables. Theory explains relationships among ideas. Concepts are abstract mental categories such as “social support,” “academic resilience,” or “institutional trust.” Variables are measurable representations of concepts.

For example, in a study on graduate employability:

Theory might predict that social capital improves job opportunities.
Concepts might include networking, skills acquisition, and employability.
Variables might include number of professional contacts, internship experience, and employment status six months after graduation.

The key exam point is that variables are not the same as concepts. A concept is broader and often theoretical; a variable is the operational form used to measure it.

1.5 Research problems and research questions

A strong research problem identifies a gap, tension, contradiction, or practical issue that warrants investigation. It is not simply a topic. “Unemployment in South Africa” is a topic; “the factors shaping graduate unemployment among social science graduates from public universities in Gauteng” is closer to a research problem because it narrows the focus, identifies a population, and points toward possible explanations.

Research questions should be:

clear
specific
researchable
linked to theory
appropriate to the design

Types of questions include:

Descriptive questions — What is happening?
Comparative questions — How do groups differ?
Relational questions — Is there an association between variables?
Causal questions — Does one factor influence another?
Exploratory questions — What themes or meanings emerge?
Evaluative questions — How effective is an intervention or programme?

An exam answer that distinguishes these types shows methodological maturity. It also helps to remember that the type of question should determine the choice of design, not the other way round.

2. Research design, sampling, and measurement

A research design is the overall blueprint that shapes how data are collected, from whom, and with what instruments. It is the bridge between an abstract question and a workable study. In advanced social science research, design decisions are never merely administrative; they determine the quality of the evidence and the strength of the conclusions.

2.1 Common research designs

Cross-sectional design

Cross-sectional studies collect data at one point in time. They are useful for describing prevalence, comparing groups, and examining associations. For example, a survey of stress and coping among first-year UNISA postgraduate students could reveal patterns at a single time point. However, cross-sectional designs cannot establish temporal order, so causal claims must be limited.

Longitudinal design

Longitudinal studies collect data over time. They may be panel studies, cohort studies, or repeated cross-sectional studies. These designs are stronger for analysing change and temporal relationships. For instance, a study tracking academic motivation from first semester to final semester would better capture developmental trends than a one-time survey.

Experimental and quasi-experimental designs

Experimental designs involve manipulation of an independent variable and random assignment to conditions. They are the strongest design for causal inference because they control many alternative explanations. In social science, true experiments are sometimes difficult or unethical, so quasi-experimental designs are often used instead. These may compare groups that were not randomly assigned but are similar in relevant ways.

Case study design

A case study provides in-depth analysis of a bounded system such as an institution, community, programme, or event. It is especially useful when context matters and when the boundaries between phenomenon and context are blurred. A case study of postgraduate supervision practices in one faculty can generate rich, contextualised insight.

Comparative design

Comparative research examines differences and similarities across groups, settings, or countries. For example, comparing social support experiences between part-time and full-time postgraduate students can clarify how study conditions shape outcomes.

Mixed methods design

Mixed methods combines quantitative and qualitative approaches within a single study. This can mean using surveys followed by interviews, or interviews followed by surveys, or integrating both simultaneously. Mixed methods are powerful when the researcher wants both breadth and depth, or when one form of data is needed to explain the other.

2.2 Designing a good study

A good design begins with the research problem and continues through sampling, data collection, analysis, and interpretation. Design quality depends on internal coherence.

A useful checklist includes:

Is the question clear?
Does the design match the question?
Is the sampling strategy defensible?
Are the instruments valid and reliable, or credible and trustworthy?
Does the analysis answer the question?
Are ethical risks addressed?
Are the conclusions limited to what the evidence supports?

If any one of these parts is weak, the whole study suffers. For instance, if a researcher wants to study the effect of financial stress on academic performance but uses a vague instrument for financial stress and a very small convenience sample, the study may generate interesting discussion but weak inference.

2.3 Sampling strategies

Sampling is the process of selecting a subset of a population for study. The quality of the sample affects the validity and credibility of the findings.

Probability sampling

Probability sampling gives every member of the population a known chance of selection. It is often used in quantitative research because it supports generalisation.

Common types include:

Simple random sampling: every unit has an equal chance of selection.
Systematic sampling: every nth case is selected from a list.
Stratified sampling: the population is divided into strata, and samples are drawn from each stratum.
Cluster sampling: clusters such as schools or wards are selected first, then units within them.

Non-probability sampling

Non-probability sampling does not give every member a known chance of selection. It is often used in qualitative research and exploratory studies.

Common types include:

Purposive sampling: participants are chosen because they have relevant experience.
Convenience sampling: participants are selected based on accessibility.
Snowball sampling: existing participants refer others.
Quota sampling: the sample includes set numbers from categories.

The choice of sampling strategy depends on the goal of the study. If generalisation to a large population is essential, probability sampling is preferable. If depth, relevance, and information-rich cases are the priority, purposive sampling may be better.

2.4 Sample size and saturation

Sample size is often misunderstood as a purely numerical issue. In reality, it depends on design, analysis, variability, and purpose. Quantitative studies usually require larger samples to detect effects reliably and to estimate population parameters with precision. Qualitative studies typically require smaller samples, but depth and saturation matter more than size alone.

Saturation is reached when additional data no longer produce new themes or insights. It is a conceptual threshold rather than a fixed number. In a study of postgraduate supervision experiences, for example, interviewing 12 students may be sufficient if the responses become repetitive and no new themes emerge, but in a highly diverse population more participants may be needed.

2.5 Measurement, operationalisation, and scales

Measurement is the process of assigning numbers or categories to variables according to explicit rules. Operationalisation translates abstract concepts into measurable indicators.

For example:

Academic engagement could be operationalised through attendance, assignment submission, and self-reported engagement.
Psychological distress could be measured using a validated scale with Likert-type items.
Socioeconomic status might be operationalised through income, employment status, and household assets.

Measurement levels matter:

Nominal: categories without order, such as gender or faculty.
Ordinal: ordered categories, such as low/medium/high satisfaction.
Interval: equal intervals but no true zero, such as many standard test scores.
Ratio: equal intervals with a true zero, such as age or income.

The level of measurement influences which statistical analyses are appropriate.

2.6 Reliability and validity

Reliability refers to consistency. A reliable measure produces similar results under similar conditions. Validity refers to whether the instrument actually measures what it claims to measure.

Types of reliability include:

Test-retest reliability
Internal consistency
Inter-rater reliability

Types of validity include:

Face validity
Content validity
Construct validity
Criterion validity

A measure can be reliable without being valid. For example, a broken scale that is consistently two kilograms off is reliable but not valid. This distinction is frequently tested because it captures a foundational measurement principle.

2.7 Quantitative and qualitative trustworthiness

Quantitative researchers emphasise validity, reliability, and objectivity. Qualitative researchers often use the criteria of credibility, transferability, dependability, and confirmability.

Quantitative criterion	Qualitative parallel	Meaning
Validity	Credibility	Findings accurately represent the phenomenon
Reliability	Dependability	Findings are consistent and traceable
Generalisability	Transferability	Findings may apply in similar contexts
Objectivity	Confirmability	Findings are shaped by the data rather than researcher bias

In exams, it is useful to explain that the criteria differ because the underlying paradigms differ. The goal is not to force qualitative research into a quantitative mould, but to judge it using appropriate standards.

3. Data collection, analysis, and interpretation

Collecting data is only one part of research; the central intellectual task is turning raw information into meaningful evidence. Advanced social science research methods demand that the researcher understand not only how to gather data, but also how to clean, organise, analyse, and interpret it with rigour. The method chosen must fit the data type, the research question, and the overall epistemological stance.

3.1 Data collection methods

Surveys and questionnaires

Surveys are widely used to collect standardised data from many participants. They are efficient, especially for descriptive and relational studies. Questionnaires can include closed-ended items, rating scales, rankings, and sometimes a few open-ended questions.

Strengths:

efficient for large samples
easy to compare responses
suitable for statistical analysis

Limitations:

limited depth
risk of superficial responses
sensitivity to wording and order effects

A survey on student mental health, for example, could measure stress levels, coping behaviours, and support access across multiple campuses. However, if the aim is to understand the meanings students attach to their experiences, a questionnaire alone would be inadequate.

Interviews

Interviews allow the researcher to probe, clarify, and explore meanings in depth. They may be structured, semi-structured, or unstructured.

Structured interviews follow a fixed set of questions.
Semi-structured interviews use guiding questions but allow flexibility.
Unstructured interviews are more open and exploratory.

Interviews are particularly useful when the topic is sensitive, complex, or under-researched. In a study of postgraduate supervision, interviews can reveal subtle issues such as communication breakdown, power imbalance, and emotional strain.

Focus groups

Focus groups gather participants together to discuss a topic. They are valuable for exploring shared norms, differences, and group interaction. The group setting can reveal consensus or tension that may not emerge in individual interviews. However, dominant voices may suppress others, so facilitation skills matter greatly.

Observation

Observation involves systematically watching behaviours, events, or interactions. It may be participant or non-participant observation. In schools, community settings, or organisations, observation can reveal actual practices rather than self-reported accounts.

Document and content analysis

Documents, policy texts, media content, court records, meeting minutes, and online materials can all serve as research data. Content analysis may be quantitative, qualitative, or mixed. This method is important in social science because institutions often leave documentary traces that reveal priorities, assumptions, and power relations.

3.2 Ethical and practical issues in data collection

Data collection is shaped by access, consent, confidentiality, and the sensitivity of the subject matter. In a South African context, additional issues may include language, inequality, historical mistrust, and institutional gatekeeping. Ethical research requires:

informed consent
voluntary participation
protection from harm
anonymity or confidentiality where promised
respect for vulnerable groups
transparent handling of data

Ethical approval is not a formality. It is part of research integrity. A study on trauma, violence, or discrimination must anticipate emotional risks and provide appropriate support or referral information where needed.

3.3 Preparing data for analysis

Before analysis, data must be prepared carefully.

For quantitative data, preparation may include:

coding responses
checking for missing data
identifying outliers
recoding reversed items
creating composite scores
examining distributions

For qualitative data, preparation may include:

transcribing interviews verbatim
checking transcript accuracy
organising data by case or theme
anonymising names and identifying details
creating a coding framework

Poor preparation leads to poor analysis. For example, if Likert items are coded inconsistently or missing values are ignored, statistical results may become misleading. Similarly, if a transcription omits pauses, emotional cues, or speaker shifts that matter to interpretation, the analysis may lose meaning.

3.4 Quantitative analysis

Quantitative analysis is used to summarise patterns, test hypotheses, and examine relationships among variables. It typically proceeds from descriptive to inferential analysis.

Descriptive statistics

Descriptive statistics include frequencies, percentages, means, medians, modes, ranges, standard deviations, and graphs. They help describe the sample and the distribution of variables.

Example:
A study of 240 postgraduate students may find that:

62% are female
38% are male
mean stress score = 3.8 on a 5-point scale
standard deviation = 0.9

These figures describe the sample, but they do not yet explain relationships or causation.

Inferential statistics

Inferential statistics allow the researcher to make estimates or test hypotheses about a broader population. Examples include:

t-tests
chi-square tests
correlation
regression
analysis of variance
non-parametric alternatives
factor analysis
multilevel models

A study examining whether financial stress predicts academic performance might use correlation or regression. A study comparing three teaching approaches might use analysis of variance. A study on associations between categorical variables, such as gender and completion status, might use chi-square.

Regression logic

Regression is especially important because it estimates how much one or more predictors are associated with an outcome while holding other variables constant. In advanced research, understanding regression as a model of adjustment and association is more important than memorising formulas.

For example, if academic performance is the dependent variable and financial stress, study time, and employment status are predictors, regression can estimate the unique contribution of each predictor. The researcher must still avoid claiming causality unless the design supports it.

3.5 Qualitative analysis

Qualitative analysis seeks patterns of meaning rather than numerical relationships. It is interpretive, iterative, and reflexive.

Thematic analysis

Thematic analysis identifies recurring patterns across a dataset. A typical process includes:

familiarisation with the data
coding meaningful segments
grouping codes into categories
developing themes
reviewing and refining themes
naming and interpreting themes

Example themes in a study of postgraduate supervision might include:

uncertainty about expectations
uneven communication
emotional support and isolation
resource constraints
strategies for persistence

Grounded theory

Grounded theory aims to build theory from data through systematic coding and comparison. It is useful when existing theory does not adequately explain the phenomenon. The researcher compares incidents, codes, categories, and theoretical relationships.

Narrative analysis

Narrative analysis focuses on stories, sequences, and identity construction. It is suitable when the way a person tells the story is as important as the content.

Discourse analysis

Discourse analysis studies how language creates social reality, identity, and power. This method is often used to examine policy texts, media statements, or institutional communication.

3.6 Mixed methods integration

Mixed methods research is not simply the use of both interviews and surveys. The key issue is integration. Integration can happen at the design stage, during data collection, at the analysis stage, or in interpretation.

Common mixed methods designs include:

Sequential explanatory design: quantitative data first, qualitative data later to explain findings
Sequential exploratory design: qualitative data first, quantitative data later to test or extend themes
Concurrent triangulation design: both strands collected at the same time and compared
Embedded design: one method is primary and the other supportive

A practical example is a study of academic resilience:

survey 300 students to measure resilience, stress, and support
interview 20 students to understand coping strategies
integrate both strands to explain why some students persist despite hardship

Advanced exam answers should emphasise that mixed methods are most valuable when they answer a problem that neither quantitative nor qualitative methods could answer alone.

3.7 Interpretation and inference

Interpretation is the process of explaining what findings mean in relation to the research question, theory, and context. Good interpretation avoids three common errors:

Overgeneralisation — treating a small, biased sample as representative of everyone
Overcausation — claiming cause and effect from correlational data
Overinterpretation — reading more into the data than the data justify

Interpretation should connect findings to:

the literature
the theoretical framework
the study context
limitations and alternative explanations

For example, if a study finds that students with stronger peer support report lower stress, the interpretation should note whether support may reduce stress, whether less stressed students seek more support, or whether a third factor such as socio-economic resources influences both.

4. Ethics, quality, and reflexivity in social science research

Research ethics and quality assurance are not separate from methodology; they are part of it. In advanced social science work, the researcher is responsible not only for generating knowledge but also for doing so with integrity, fairness, and awareness of power. The study’s credibility depends heavily on how these concerns are handled from start to finish.

4.1 Core principles of research ethics

Several ethical principles recur across social science research:

Respect for persons: participants must be treated as autonomous individuals capable of making informed decisions.
Beneficence: the study should aim to maximise benefits and minimise harm.
Justice: the burdens and benefits of research should be distributed fairly.
Confidentiality: data should be protected from unauthorised disclosure.
Integrity: the researcher must avoid deception, fabrication, falsification, and plagiarism.

In a postgraduate research environment, ethics also includes proper supervision, truthful reporting, careful authorship practice, and responsible data management.

4.2 Informed consent and voluntariness

Informed consent means participants understand:

the purpose of the research
what participation involves
potential risks and benefits
their right to refuse or withdraw
how their data will be used

Consent must be voluntary. If participants feel pressured by authority, fear, or dependency, consent may be formally signed but ethically weak. This is particularly important in hierarchical contexts such as schools, workplaces, hospitals, and universities.

An example:
If a lecturer recruits students from her own class for a survey, the power relationship may influence willingness to participate. To reduce coercion, participation should be clearly separated from grading, and an independent person should handle consent where possible.

4.3 Privacy, confidentiality, and anonymity

These terms are related but not identical.

Privacy concerns the participant’s right to control access to themselves and their information.
Confidentiality means the researcher promises to protect identifiable information.
Anonymity means the researcher cannot link data back to a participant.

In many small qualitative studies, true anonymity is difficult because details in narratives may identify participants indirectly. In such cases, confidentiality becomes especially important, and identifying details must be altered carefully without distorting meaning.

4.4 Ethics in vulnerable and sensitive research

Research involving trauma, violence, children, prisoners, unemployed people, migrants, or people in precarious conditions requires heightened ethical care. Risks may include:

emotional distress
retraumatisation
stigma
reprisal
social or economic harm

A study on gender-based violence, for instance, should not force participants to disclose painful details unnecessarily. Questions should be carefully sequenced, support referrals considered, and data storage secured. The ethical issue is not only whether the topic is sensitive, but whether the research procedure is safe and respectful.

4.5 Research quality: validity, trustworthiness, and rigour

Quality in research is the degree to which findings are believable, well-supported, and appropriate to the purpose.

Quantitative rigour

Quantitative studies are judged by:

internal validity
external validity
construct validity
statistical conclusion validity
reliability
transparency of procedures

Threats to internal validity include:

history
maturation
selection bias
instrumentation changes
attrition
testing effects

For example, if a study measures students’ stress before and after exams, a reduction in stress may reflect the end of exams rather than the intervention being studied. The researcher must consider rival explanations.

Qualitative rigour

Qualitative studies are judged by:

credibility
transferability
dependability
confirmability
authentic representation

Strategies to improve qualitative rigour include:

prolonged engagement
triangulation
member checking
audit trails
reflexive journaling
peer debriefing

4.6 Reflexivity

Reflexivity is the practice of critically examining the researcher’s own position, assumptions, values, and influence on the research process. This is especially important in social science because the researcher is not a neutral machine. Their identity, background, training, and institutional location can shape what questions are asked, what participants disclose, and how findings are interpreted.

Reflexivity matters because:

it helps identify bias
it improves transparency
it deepens interpretation
it strengthens ethical awareness

A researcher studying educational inequality in South Africa may need to reflect on how their own schooling history, language background, and class position influence their interpretation of participant accounts. Reflexivity is not self-indulgence; it is methodological discipline.

4.7 Positionality and power

Positionality refers to the social and political location from which the researcher speaks. Power operates in research through access, language, institution, gender, race, class, and authority. Advanced social science methods require sensitivity to these issues because they influence both data production and interpretation.

For instance, in an interview study with junior employees, participants may provide socially desirable answers if the interviewer is perceived as linked to management. Similarly, in community-based research, mistrust may shape participation if people have experienced extractive or exploitative research in the past.

A strong exam response should show that ethics is not limited to paperwork. It is embedded in the structure of the relationship between researcher and participant.

5. Exam strategy, common pitfalls, and integrated application

Success in RESM711 depends on more than familiarity with the terminology. It requires the ability to build coherent methodological arguments, identify weaknesses in research designs, and apply concepts flexibly to new scenarios. Exam questions often combine theory with application, so answers should be structured, analytical, and precise.

5.1 How to answer advanced methods questions

A strong answer usually follows a disciplined structure:

Define the key concept accurately.
Explain the underlying logic or principle.
Apply the concept to a relevant example.
Discuss strengths and limitations.
Conclude with a clear judgement.

For example, if asked about sampling, do not merely list methods. Explain how sampling affects representativeness, bias, feasibility, and the scope of inference. If asked about mixed methods, explain why integration matters and when a mixed approach is superior to a single-method design.

5.2 Common exam question types

Definition and distinction questions

These questions test whether terms are understood clearly. Examples include:

Distinguish between reliability and validity.
Differentiate ontology from epistemology.
Compare qualitative and quantitative approaches.

The safest strategy is to define each term, then show the difference in purpose, logic, and example.

Applied design questions

These ask which method or design is suitable for a scenario. Example:
“A researcher wants to understand how unemployed graduates experience the transition from university to work.”

A good answer would likely recommend a qualitative or mixed methods design, possibly with purposive sampling and semi-structured interviews, because the question focuses on experience and meaning.

Critique questions

These ask for strengths, weaknesses, and limitations of a method or study. The strongest responses identify specific methodological risks rather than general complaints. For example, “the sample is too small” is weaker than “the use of convenience sampling in a highly diverse population limits transferability and increases selection bias.”

Comparison questions

These ask for evaluation across two or more methods or paradigms. Always compare along the same dimensions:

purpose
data type
sample logic
strengths
limitations
type of inference

5.3 Building a methodologically sound argument

When writing an exam answer or proposal, the argument should connect the following elements:

problem statement
theoretical orientation
research question
design
sampling
measurement or data generation
analysis
ethics
expected contribution

A coherent study on postgraduate mental health might look like this:

Problem: student distress is rising, but the mechanisms are poorly understood.
Question: how do postgraduate students experience and manage academic stress?
Paradigm: interpretivist or pragmatic.
Design: mixed methods or qualitative case study.
Sampling: purposive sampling of diverse postgraduate students.
Data collection: survey plus interviews.
Analysis: descriptive statistics plus thematic analysis.
Ethics: consent, confidentiality, emotional support.
Contribution: improves institutional understanding of support needs.

The key is not merely to choose methods, but to justify them in relation to the research problem.

5.4 Frequent mistakes to avoid

Students often lose marks for the following reasons:

Confusing method with methodology
A method is a specific technique; methodology is the overall strategy and logic behind choosing methods.
Using “qualitative” and “quantitative” as value judgements
Neither is inherently better. The right method depends on the question.
Claiming causation from correlation
Association does not prove cause and effect unless the design supports causal inference.
Ignoring limitations
Every study has limitations. Good scholarship acknowledges them honestly.
Vague operationalisation
Abstract concepts must be translated into clear measures or categories.
Weak alignment between question and design
A question about lived experience should not be answered only with numeric frequencies unless the design explicitly allows that.
Overlooking ethics
Ethical issues are not optional extras; they are central to quality and legitimacy.

5.5 A practical study example

Consider a hypothetical study titled: “Academic resilience among first-generation postgraduate students at a South African public university.”

A strong methodological plan might include:

Research problem: first-generation students may face distinct barriers, but the processes that support resilience are under-explored.
Research question: how do first-generation postgraduate students sustain academic resilience in the face of financial, social, and institutional pressures?
Paradigm: pragmatic, with interpretive sensitivity.
Design: sequential explanatory mixed methods.
Sampling: survey 180 students, then purposively interview 18 students representing high, medium, and low resilience scores.
Data collection: a structured questionnaire for resilience, stress, and support; follow-up semi-structured interviews.
Analysis: descriptive statistics, correlation, and thematic analysis.
Integration: use interview findings to explain why some students score high despite hardship.
Ethics: informed consent, confidentiality, and referral pathways for distress.

This example demonstrates how the parts of a study fit together. It also shows how advanced methods combine measurement, interpretation, and ethics rather than treating them as separate exercises.

5.6 Final revision priorities

For final revision, focus on the concepts most likely to be examined in integrated form:

paradigms and philosophical assumptions
types of research questions
design logic
sampling and representativeness
measurement, reliability, and validity
qualitative trustworthiness
statistical inference
thematic and mixed methods analysis
ethics and reflexivity
critique and justification of methodological choices

A useful way to revise is to practise answering scenario-based questions. For each scenario, identify:

the research problem
the best paradigm
the most suitable design
the sample strategy
the data collection method
the analysis approach
the ethical concerns
the likely limitations

This kind of structured practice mirrors the real demands of postgraduate research and helps build the habit of methodological reasoning.

5.7 Core takeaway for RESM711

The central lesson of advanced social science research methods is that good research is not defined by fashion, complexity, or the quantity of data collected. It is defined by coherence, transparency, and fit between purpose and procedure. A researcher must be able to explain why a particular method is appropriate, what kind of knowledge it produces, what its limits are, and how the findings should be interpreted responsibly. In RESM711, that methodological judgement is the difference between a superficial answer and a strong postgraduate answer.

Quick revision tables

Key distinctions at a glance

Topic	Essential idea	Common mistake
Ontology	What exists	Confusing it with methods
Epistemology	How knowledge is known	Treating it as the same as theory
Methodology	Why methods are chosen	Using it to mean only “research method”
Reliability	Consistency	Thinking it guarantees validity
Validity	Accuracy of measurement	Assuming all valid measures are reliable in the same way
Sampling	How participants are selected	Assuming convenience samples are representative
Correlation	Association	Mistaking it for causation
Trustworthiness	Qualitative quality	Applying only quantitative criteria

High-yield terms to know

paradigm
ontology
epistemology
methodology
operationalisation
validity
reliability
credibility
transferability
triangulation
saturation
reflexivity
positionality
causality
mixed methods
thematic analysis

Final exam habit

When answering any question, repeatedly ask:

What is the claim being made?
What evidence is needed?
What design can produce that evidence?
What are the limits of the design?
What would a critical examiner challenge?

That habit turns isolated knowledge into advanced methodological competence.