Psychology Dissertation Topics for UK Students: Methods, Measures and Data Routes

Questions to Settle Before You Choose a Psychology Dissertation Topic

Strong psychology dissertation topics share one trait: a student can finish them within a single academic year. Before you pick from the ideas below, settle these questions, because each one decides whether a topic survives your ethics committee and your timetable.

• Can I answer my question with a design my department’s ethics committee will approve before my data collection window closes?
• Do I have a realistic route to participants, or should I start from an existing dataset such as Understanding Society?
• Does a validated scale already measure my outcome, and is that scale free to use or does it need a licence?
• Does my course expect me to collect new data, or can I submit a systematic review or a secondary analysis?
• Which statistical test does this design need, and could my likely sample detect a realistic effect?

Psychology is a broad field, so the ideas here are grouped by the route you would take to answer them: online experiments, existing datasets, qualitative work, replication, reviews, applied settings, and clinical or health questions. Every area from social to neuropsychology appears somewhere in that structure. Animal research, work with detained patients and studies that need NHS approval are left out, because most undergraduate projects cannot take those routes (the ethics section explains why). Every topic is free to use and adapt.

If you are still deciding on a direction, the dissertation topic finder narrows a broad interest into a shortlist. The reasoning behind a good choice is covered in more detail in a guide on how to choose research topics, so read it if none of the ideas below fits yet.

Online Experiments and Surveys for Psychology Dissertations

Online designs suit most undergraduate and many Master’s projects, because platforms such as Gorilla, Pavlovia, jsPsych and Qualtrics remove the need for a lab. Recruit through your university participant pool, or through a paid panel such as Prolific if your department funds it. In every case, run a power analysis first, and a sample size calculator gives a quick estimate before your supervisor meeting.

1. How far does Cyberball-style exclusion in an online group chat lower belonging compared with exclusion in a ball-tossing game?
Cyberball, developed by Kipling Williams and colleagues, is the standard ostracism task, and its need-threat measures (belonging, self-esteem, control and meaningful existence) give you clear outcomes. A text-based group chat version tests whether the paradigm transfers to the way students actually experience exclusion. Randomise participants to included or excluded conditions within each format, and then compare formats with a two-way ANOVA. Plan a thorough debrief, because exclusion involves mild deception, and look at further social psychology ideas if you want a different paradigm built on the same logic.

2. Whether Allport’s optimal contact conditions, rather than the sheer amount of contact, predict intergroup anxiety between home and international students in the same seminar groups.
Allport’s conditions are equal status, common goals, cooperation and institutional support. Measure perceived contact quality and quantity alongside the Intergroup Anxiety Scale developed by Stephan and Stephan, then test whether quality predicts anxiety once quantity is in the regression. Seminar groups give you a defined setting, but recruit across several groups so one tutor’s style does not drive the result. A Master’s version could add cross-group friendship as a mediator.

3. To what extent does attentional control theory explain why test anxiety slows flanker-task responses more than it reduces accuracy when time pressure is added?
Attentional control theory, proposed by Eysenck, Derakshan, Santos and Calvo, predicts that anxiety first erodes processing efficiency, so response times should suffer before accuracy does. A two-by-two design (high versus low test anxiety, with and without a response deadline) puts that prediction under pressure. Build the flanker task in PsychoPy or jsPsych, and screen anxiety with a validated scale such as the Test Anxiety Inventory. Report reaction time and errors separately, since the theory turns on the gap between them, and browse more cognitive psychology research topics for related tasks to use as a second measure.

4. Does retrieval practice beat highlighting for a one-week delayed quiz on lecture content, and does Need for Cognition change the benefit?
The testing effect, shown by Roediger and Karpicke, makes retrieval practice a well-supported contrast with re-reading or highlighting. A short recorded lecture and a free-recall quiz keep the study affordable for undergraduates. Attrition between the first and second session is the main threat, so schedule the delayed test before recruiting and send reminders. Adding the Need for Cognition scale lets you ask for whom the benefit is larger, which moves a replication-style study towards a modest original contribution.

5. Do the Cognitive Reflection Test and the Berlin Numeracy Test predict sunk-cost errors independently of each other?
Frederick’s Cognitive Reflection Test and Cokely and colleagues’ Berlin Numeracy Test both tap analytic ability, so entering them together in a hierarchical regression shows which one carries the prediction. Write sunk-cost scenarios of the kind used by Arkes and Blumer in a student context, such as a non-refundable society trip. Because the original reflection items circulate widely online, use extended versions and ask participants whether they have met the problems before.

6. Whether HEXACO Honesty-Humility predicts intention to cheat on coursework better than the Short Dark Triad.
Honesty-Humility is the HEXACO factor most directly tied to rule-breaking, while the Short Dark Triad by Jones and Paulhus captures Machiavellianism, narcissism and psychopathy in a brief self-report. Self-report on misconduct invites social desirability, so ask about intention in anonymous vignettes instead of past behaviour. State clearly that responses cannot identify individuals or modules, and keep the sample at one institution only if participants cannot be identified from their answers.

7. How effectively does a two-week “three good things” diary raise wellbeing compared with a neutral-events diary, measured with the Warwick-Edinburgh Mental Wellbeing Scale?
The “three good things” exercise from Seligman and colleagues is a clear, low-risk intervention for students. Use a pre-post design with random assignment and a matched control task, and track completion rates, since diary studies lose participants quickly. Check the licence terms for the wellbeing scale before you build the survey around it. Other positive psychology dissertation ideas follow the same intervention-versus-control logic, so you can swap the exercise without changing the design.

8. How far does attachment anxiety on the ECR-R predict distress after an unanswered message that shows a read receipt?
The Experiences in Close Relationships Revised scale separates attachment anxiety from avoidance, and read receipts give a concrete, everyday trigger. Present short vignettes in which the sender is a close friend or a partner, and measure distress and hostile attribution after each. A within-subject design raises power, which helps with small samples. Pilot the vignettes for realism before the main study.

9. Which of Fear of Missing Out and trait impulsivity better predicts response to limited-stock messages in a simulated checkout?
The Fear of Missing Out scale developed by Przybylski and colleagues, paired with the Barratt Impulsiveness Scale, sets up a contest between two explanations. Randomise participants to a scarcity banner or no banner, and record purchase intention and a behavioural choice such as selecting a bonus item. Preregister the interaction you expect, since consumer designs invite many unplanned comparisons, and see how consumer psychology research handles realistic stimuli before you build your own.

10. Does cognitive offloading to a generative AI drafting tool reduce later recall of the arguments students produced, compared with drafting alone?
Risko and Gilbert’s work on cognitive offloading gives a theoretical frame: relying on an external aid can free resources but weaken memory for the offloaded content. Ask participants to draft a short argument with or without a chatbot, then test free recall after a delay, scored by two raters who do not know the condition. Record the tool version and the date of use, because model behaviour changes and your method must be reproducible. A simple two-group design works at Bachelor’s level.

11. Does telling participants they are chatting with an AI rather than a human counsellor change willingness to disclose, once general disclosure tendency on the Distress Disclosure Index is controlled?
Kahn and Hessling’s index measures a general tendency to disclose, so treat it as a covariate and measure state willingness separately after a text vignette. Keep vignettes to everyday stress, not trauma, so the study stays within standard university ethics review. A manipulation check asking who participants believed they were talking to is essential. Related designs in counselling psychology topics use the same vignette logic with different outcome measures.

12. Does eco-anxiety on the Hogg Eco-Anxiety Scale predict a real donation choice after accounting for general anxiety on the GAD-7?
Self-reported pro-environmental behaviour overlaps heavily with attitudes, so replace it with a behavioural choice, such as splitting a small prize between a climate charity and keeping it. Enter the GAD-7 first in a hierarchical regression to see whether eco-anxiety adds anything. A power analysis for a small incremental effect often shows that the sample you need is larger than expected, so decide early whether a paid panel is affordable.

13. Is the Rosenberg Self-Esteem Scale measurement invariant between UK-born and international students?
Comparing group means is only meaningful if the scale measures the same construct in both groups, which multi-group confirmatory factor analysis tests through configural, metric and scalar invariance. Use R with lavaan or JASP, and inspect the negatively worded items, which often behave differently across cultural groups. This is a Master’s-level measurement project, and cross-cultural psychology topics suggest other scales that deserve the same test.

14. How far do public and anonymous dictator-game choices differ in generosity, and does Social Value Orientation explain who responds to an audience?
Reputation effects are a staple of evolutionary accounts of cooperation, and the dictator game is a transparent way to test them. Pair it with the SVO Slider Measure by Murphy, Ackermann and Handgraaf, and compare giving when choices are visible to a small group with fully anonymous choices. Use small, genuine stakes and pay participants according to their choices, because deception about real payment is hard to defend to an ethics committee. Further evolutionary psychology dissertation ideas use related economic games, so the same software can serve a second study.

Psychology Research Topics Using Existing UK Datasets

Secondary analysis avoids recruitment, which is often the step that runs late. Understanding Society, the Millennium Cohort Study and the English Longitudinal Study of Ageing are all available through the UK Data Service, normally after registration and agreement to usage terms. Check your department’s ethics rules for secondary data, and read each study’s documentation before you write a research question, because variable availability differs by wave.

15. How much do conclusions about social media use and adolescent life satisfaction in Understanding Society change across reasonable analytic choices, shown with a specification curve analysis?
Orben and Przybylski applied specification curve analysis to three large datasets and reported a small negative association between digital technology use and adolescent wellbeing, explaining at most 0.4 percent of the variation. Their approach shows how the same data can support opposite headlines, depending on measures and covariates. A later paper by Orben and colleagues used Understanding Society and the Millennium Cohort Study to propose age- and sex-specific windows of sensitivity, which gives you a published result to extend or test. Your project would define a defensible set of choices (which social media item, which wellbeing measure, which covariates) and plot every result. R is the usual tool, so the project suits a Master’s student comfortable with scripting.

16. Does loneliness or objective social isolation better predict decline in delayed word recall across English Longitudinal Study of Ageing waves?
The study follows adults aged 50 and over and, as far as its documentation shows, carries both loneliness items and word-recall memory tests, so check which waves hold both. Cross-lagged or latent growth models handle change over time, and attrition needs multiple imputation or full-information maximum likelihood. Because participants who drop out tend to be in poorer health, discuss how attrition could bias your estimates. Questions about cognitive ageing continue in neuropsychology dissertation topics, which also list tasks suited to smaller primary studies.

17. To what extent do trajectories of parent-reported emotional symptoms on the Strengths and Difficulties Questionnaire differ by household income at age three in the Millennium Cohort Study?
The Millennium Cohort Study follows children born in 2000 and 2001, which gives repeated measures across childhood and adolescence. Growth mixture modelling can ask whether distinct trajectories exist, but it demands large samples and careful model selection, so a latent growth curve with income as a predictor is the safer Master’s-level choice. Report how you handled survey weights and the sampling design. Child-focused questions that rely on different measures appear in developmental psychology topics, and they are useful for comparing how other researchers framed outcomes.

18. Is the factor structure of the GHQ-12 in Understanding Society invariant across age groups and sexes?
The twelve-item General Health Questionnaire is widely used as a screen for psychological distress, and researchers still disagree about whether it reflects one factor or several, partly because of method effects from reverse-worded items. A multi-group confirmatory factor analysis with ordinal estimators tests whether distress means the same thing across groups. The project suits a student who wants methods depth, and quantitative psychology dissertation topics list further psychometric questions you could compare it with.

19. How many trials per participant are needed for stable N400 and P3 effects in the openly shared ERP CORE EEG dataset?
ERP CORE, from Emily Kappenman, Steven Luck and colleagues, is an open set of EEG recordings across several standard tasks, shared through the Open Science Framework. You can subsample trials, compute data quality and effect size at each count, and show where reliability levels off. The answer matters to every EEG student, and it needs no recruitment, but it does require comfort with EEGLAB, ERPLAB or MNE-Python. The project suits Master’s or PhD level.

Qualitative Psychology Project Ideas and Methods

Qualitative projects answer questions about meaning rather than effect size. Choose the method before the topic, because interpretative phenomenological analysis, reflexive thematic analysis, grounded theory and discourse analysis each rest on different assumptions and are marked against different criteria.

20. Making sense of impostor feelings: an interpretative phenomenological analysis of first-generation students’ first-year accounts
Clance and Imes described the impostor phenomenon in 1978, and IPA suits it because the method asks how people make sense of a specific experience. Smith, Flowers and Larkin’s guidance asks for a small, fairly homogeneous sample, so define “first-generation” tightly and recruit within one or two institutions. Expect to spend most of your time on analysis, not recruitment. IPA interprets each participant in depth before comparing cases, which is different from coding across a whole dataset.

21. How do hybrid workers describe the boundary between home and work, read through boundary theory using reflexive thematic analysis?
Ashforth, Kreiner and Fugate’s work on role transitions supplies the theoretical lens, and Braun and Clarke’s reflexive thematic analysis supplies the method. In reflexive analysis, themes are constructed through your interpretation instead of being found in the data, so write about your own position and avoid phrases such as “themes emerged”. Interview adults who work from home for part of the week, and keep the topic guide short enough to allow follow-up questions.

22. What do trainee counsellors describe as ruptures and repairs in online sessions, interpreted through Safran and Muran’s rupture resolution model?
Safran and Muran’s model describes withdrawal and confrontation ruptures and the repair processes that follow, which gives you concrete categories to explore without imposing them on participants. Trainees on supervised placements can discuss their own practice without disclosing client details, but the consent form must state the confidentiality boundary explicitly. A critical incident interview works well here, and psychotherapy dissertation topics include other process research questions that use similar interview formats.

23. How do competitive esports players explain and manage losing streaks? A constructivist grounded theory study
Charmaz’s constructivist grounded theory treats the theory as co-constructed by researcher and participants, so memo writing and theoretical sampling matter more than counting interviews. Players in organised leagues give you a bounded population for a Master’s project. Expect your supervisor to ask why grounded theory instead of thematic analysis, so prepare that justification before the proposal meeting.

24. How do commenters on public short-video posts talk about self-diagnosing ADHD, analysed with discursive psychology?
Edwards and Potter’s discursive psychology looks at what talk does, such as how people claim expertise or resist stigma, instead of what is “really” in their heads. Public comments are accessible data, but the British Psychological Society’s guidance on internet-mediated research asks you to consider whether posters would expect researchers to read their words. Seek advice from your ethics committee, anonymise quotations, and avoid verbatim text that a search engine could trace. The analytic stance takes time to learn, so this is a Master’s project.

25. Camouflaging and burnout among autistic university students: a survey followed by interviews, co-designed with autistic advisers
The Camouflaging Autistic Traits Questionnaire, developed by Hull and colleagues, gives a quantitative strand, while interviews explore how camouflaging feels from the inside. Involving autistic advisers in wording, recruitment and interpretation is widely recommended good practice, and your ethics application should say how you will do it. Allow extra time for adviser meetings and for accessible participant information. Other directions, including different measures and populations, are collected in autism dissertation topics, which helps if you need a backup design.

Replication and Open Science Dissertation Ideas

Replication projects suit students who want a clear method and a defensible contribution. The Open Science Framework hosts preregistrations and open data, and some departments accept preregistered or Registered Report-style proposals. Check your course rules first, since some require a novel hypothesis.

26. Can a small preregistered ego-depletion study with Bayes factors tell a null effect apart from an underpowered one, following the 23-laboratory replication by Hagger and colleagues?
That replication, published in 2016, reported an effect of d = 0.04 with a confidence interval spanning zero for a sequential-task protocol based on work by Sripada, Kessler and Jonides. A single student sample cannot rule out a small effect, so frame the aim as estimation, report Bayes factors in JASP, and be honest about what your sample can and cannot show. Pooling data from several students under one supervisor is one way to increase precision. The project suits Master’s level.

27. Do bilingual and monolingual adults differ in Simon-task conflict effects once proficiency and socioeconomic status are modelled as continuous variables, and what do Bayes factors show?
The bilingual advantage in executive control remains debated, and many studies compare crude groups. Measure language history with a tool such as the Language Experience and Proficiency Questionnaire, and treat proficiency as a continuous predictor. A null result is informative if you report the Bayes factor and pre-specify the smallest effect you would consider meaningful. Bachelor’s students can run the task online, while Master’s students might add a second executive function task.

28. Does a brief incremental theory of intelligence message change persistence on unsolvable anagrams among first-year undergraduates? A preregistered conceptual replication
Dweck’s implicit theories work predicts that believing ability can grow encourages persistence after difficulty. Unsolvable anagrams give a measurable persistence outcome, such as time before giving up, but participants meet a task they cannot complete, so the debrief must be thorough. Preregister your exclusion rules and your persistence measure before data collection, and report the manipulation check.

29. How much do the conclusions of a published study with open data depend on exclusion rules and scale scoring, as shown by a multiverse analysis?
Steegen and colleagues introduced multiverse analysis to show how many defensible data-processing choices exist for the same dataset. Choose a study whose raw data and materials are public on the Open Science Framework, list every reasonable choice, and plot the range of results. You are not accusing the original authors, so frame the work as a robustness test. The project needs R skills and fits Master’s level.

30. Why do robust cognitive-task effects such as the Stroop effect show weak test-retest reliability, and what do split-half and retest estimates show in your own data?
Hedge, Powell and Sumner described this as the reliability paradox: tasks that produce stable group effects can still be poor at ranking individuals. Collect the same task twice with a gap between sessions, and estimate the reliability of the difference score directly. The result matters for any individual-differences study that uses the task, which makes the project useful beyond your own dissertation.

Systematic Review and Meta-Analysis Topics

If your course allows reviews, they remove the need to recruit participants. They are not easier, though, because transparent searching, screening and risk-of-bias judgement take longer than most students expect. Register the protocol where your university allows it, for example on PROSPERO for reviews with a health outcome, and follow the PRISMA 2020 reporting guideline.

31. Do therapist-guided and unguided internet-delivered CBT programmes differ in outcomes for adolescent anxiety? A PRISMA 2020 systematic review
Search for existing reviews before you commit, because you must justify why another is needed, for example a narrower age range or a later search date. Use a recognised risk-of-bias tool for randomised trials and pre-specify the primary outcome. The project suits Master’s level, while an undergraduate version would normally narrow to a single outcome or a narrative synthesis.

32. How large is the effect of mobile mindfulness apps on perceived stress in university students, and how much does small-study bias change it?
A random-effects meta-analysis in R with the metafor package gives a pooled estimate, while funnel plots, Egger’s test and PET-PEESE show whether small studies with large effects inflate it. Report heterogeneity and explain how app type, comparator and follow-up length contribute to it. You will need a second screener for part of your records, which is easier to arrange before you start than after. The project suits Master’s or PhD level.

33. How do studies of generative AI chatbots measure psychological outcomes? A scoping review following Arksey and O’Malley and Levac and colleagues
A scoping review maps what exists instead of testing whether interventions work, so your research question concerns the range of outcomes, designs and populations. Keep your search date explicit and record which chatbot versions each study used, because the literature changes quickly. A tightly defined scoping review can suit a Bachelor’s project, but ask your supervisor first whether your course accepts a review.

34. How do adults describe returning to work after burnout? A meta-ethnography using Noblit and Hare, reported with ENTREQ
Meta-ethnography translates findings across qualitative studies instead of pooling numbers. Noblit and Hare’s approach asks you to compare concepts across papers, and the ENTREQ statement provides a reporting checklist for qualitative synthesis. It is demanding interpretative work that suits Master’s or PhD level.

Applied Psychology Dissertation Topics for Schools, Workplaces and Sport

Applied projects depend on access, so confirm a gatekeeper’s agreement in writing before you commit. Schools, clubs and employers each have their own approval steps, and these can take longer than a university ethics decision.

35. Do scores on the Basic Psychological Need Satisfaction and Frustration Scale explain differences in engagement between live and recorded lectures, using self-determination theory?
Self-determination theory, from Deci and Ryan, predicts that autonomy, competence and relatedness support engagement, and the scale by Chen and colleagues measures both satisfaction and frustration of those needs. Compare students’ ratings across formats on the same module, and add a short engagement measure. Reaching students in your own department is simple, and educational psychology topics offer more learning-focused questions that could replace or extend the engagement outcome.

36. Is performance on the Strange Stories task related to peer acceptance in middle childhood after controlling for receptive vocabulary on the British Picture Vocabulary Scale?
Happé’s Strange Stories task measures advanced theory of mind, but vocabulary can explain apparent differences, which is why you must control for language. Working with children requires an enhanced Disclosure and Barring Service check in England and Wales (the PVG scheme applies in Scotland and AccessNI in Northern Ireland), plus parental consent and school approval. The vocabulary test is commercially published, so check access and cost early. Most undergraduates cannot arrange this alone, so it suits a Master’s student or an undergraduate joining a supervisor’s existing school partnership, and school psychology research topics show other school-based designs that need less access.

37. Which recovery experience in Sonnentag and Fritz’s framework (psychological detachment, relaxation, mastery or control) best protects hybrid office workers from weekly exhaustion?
A weekly diary over several weeks gives repeated measures, which you analyse with multilevel models. Choose a free exhaustion measure, such as the Copenhagen Burnout Inventory, because some widely used burnout inventories are licensed for a fee. Recruit through professional networks instead of a single employer, so no manager can see who took part. Related industrial and organisational psychology ideas extend the same diary approach to other wellbeing outcomes.

38. Does team psychological safety, as defined by Edmondson, predict how often members admit mistakes during a simulated project task?
Edmondson’s scale measures shared beliefs about interpersonal risk, and a simple simulated task with built-in errors lets you observe disclosure instead of relying only on self-report. Use student project teams and randomly assign a briefing that frames mistakes as learning or as failure. Ethics review will focus on deception and on keeping team identities private. Questions about organisations beyond student teams appear in business psychology topics, which suit a Master’s extension.

39. Does the direction of competitive state anxiety, not just its intensity, predict performance in club-level sport?
The Competitive State Anxiety Inventory-2, associated with Martens and colleagues, has an added directional scale from Jones and colleagues that asks whether symptoms feel helpful or harmful. Collect scores shortly before competition, together with performance indicators such as shooting accuracy or match statistics agreed in advance with the coach. The main practical risk is timing, because you must be present before matches without disrupting routines. Different measures and sports are covered in sports psychology dissertation ideas, which is useful if your club has only a few fixtures.

40. Does psychological readiness on the ACL-RSI scale predict return to sport among club athletes recovering from knee ligament injury?
The Anterior Cruciate Ligament-Return to Sport after Injury scale, developed by Webster and colleagues, measures confidence and fear around returning to sport. Recruiting through clubs and physiotherapists is slower than using a participant pool, and health information is sensitive, so your data management plan must cover storage and who can see identifiable data. Prospective follow-up is stronger than a single survey but takes time, so a Master’s project is more realistic. Related questions about adjustment and recovery sit in rehabilitation psychology dissertation topics, and they can inform your outcome choice.

41. To what extent does rapport-based PEACE-style questioning elicit more correct detail from mock witnesses than a closed-question format?
PEACE is the model of investigative interviewing used by police in England and Wales, and its rapport-based style contrasts naturally with closed questioning. Show participants a staged event video, interview them in one of two formats, and code correct details, errors and confabulations without knowing the condition. Train your student interviewers, and add a check on how closely they followed each format. Other designs for witness memory appear in forensic psychology dissertation topics, which also discuss jury vignettes.

42. How do attitudes towards prisoners, measured with the Attitudes Toward Prisoners scale, relate to support for rehabilitation versus punishment in sentencing vignettes?
Melvin and colleagues’ scale gives a validated attitude measure, which you can pair with vignettes that vary offence type. Keep offences non-graphic, since vivid material raises ethics concerns and may distress participants. A Bachelor’s survey design works well. Offender-focused ideas with different measures are listed in criminal psychology topics, and they pair naturally with this attitude design.

Clinical, Counselling and Health Psychology Topics Without NHS Access

Clinical questions are possible with non-clinical or analogue samples, as long as you describe your sample honestly and do not claim your findings apply to patients. Screening for distress also creates a duty of care, so your ethics application needs a signposting plan for participants who score high.

43. Does repetitive negative thinking mediate the link between insomnia severity and next-day anxiety in a daily diary of students?
The Perseverative Thinking Questionnaire by Ehring and colleagues and the Insomnia Severity Index by Bastien and colleagues are validated and well documented. Daily diaries need multilevel mediation analysis, which is demanding, so plan a simpler two-wave version if your course has not taught multilevel modelling. Include signposting to university wellbeing services in the information sheet, because you are measuring sleep problems and anxiety.

44. How well do intolerance of uncertainty and health anxiety each predict distress after a standardised online symptom search?
Pair the Intolerance of Uncertainty Scale with the Short Health Anxiety Inventory from Salkovskis and colleagues, then ask participants to look up a benign symptom in a fixed set of pages, so everyone sees the same material. Measure distress before and after, and debrief by explaining that the scenario was mild and by pointing to reliable sources of health information. The Cyberchondria Severity Scale offers a trait outcome if you prefer a survey-only design.

45. Do students high in social anxiety show more safety behaviours and post-event processing after a speech task, as the Clark and Wells model predicts?
The Clark and Wells model proposes that self-focused attention and safety behaviours maintain social anxiety. Use the Social Phobia Inventory to select high and low groups from a non-clinical sample, then record safety behaviours and post-event rumination. A live speech task is distressing, so ethics review will want a debrief, a right to stop and a recorded plan for participants who become upset. Say plainly in your write-up that the results come from a non-clinical analogue sample. Treatment-focused alternatives are collected in clinical psychology dissertation topics, which helps if your department rejects live stressors.

46. Does self-compassion buffer the link between perceived academic failure and rumination, and which of its components carries the effect?
The short form of Neff’s Self-Compassion Scale yields subscale scores, so you can ask whether self-kindness, common humanity or mindfulness matters most. Pair it with a rumination measure and a standardised vignette about a failed assessment. Report subscale and total scores carefully, since the factor structure of the scale is itself debated.

47. Which COM-B components best predict intention to seek help for low mood among university students?
Michie, van Stralen and West’s COM-B model splits behaviour into capability, opportunity and motivation, and links to the Behaviour Change Wheel. Build a short questionnaire from Theoretical Domains Framework items, and use a vignette about low mood. Intention is not behaviour, so say so in your limitations, and consider a follow-up question about help sought a few weeks later. More behaviour-change designs appear in health psychology topics, including ones that use other models.

48. Is habitual sleep quality on the Pittsburgh Sleep Quality Index related to n-back accuracy at two load levels in healthy adults?
The index asks about the past month, so pair it with a sleep diary for the week before testing. Test at a similar time of day for all participants, and record caffeine use as a covariate. The n-back task is available in several open tools, which keeps costs down, and the index is copyrighted, so confirm permission to use it. Keep your discussion free of causal claims about sleep.

Worked Examples With Aims and Objectives

A topic becomes a proposal when it has an aim and objectives you can test. Four of the ideas above are written out below, and the research aims and objectives generator can draft the same structure for your own shortlist.

Topic 15: Specification curve analysis of social media use and life satisfaction

Aim: To establish how much the association between social media use and life satisfaction in Understanding Society’s youth data varies across defensible analytic choices.

1. To define, before analysis, a justified set of social media measures, life satisfaction measures and covariates.
2. To run every combination and record the standardised coefficient and confidence interval for each.
3. To report the share of specifications that are negative, null and positive, and identify which choices change the sign.

Topic 20: Impostor feelings among first-generation students

Aim: To understand how first-generation undergraduates make sense of impostor feelings during their first year at a UK university.

1. To conduct semi-structured interviews with a small, homogeneous sample recruited from one or two institutions.
2. To analyse each transcript in turn using interpretative phenomenological analysis and record personal experiential themes.
3. To compare cases to identify shared and divergent themes, alongside a reflexive account of the researcher’s position.

Topic 1: Online exclusion and belonging

Aim: To test whether exclusion in an online group chat lowers belonging more than exclusion in a Cyberball game.

1. To randomise participants to included or excluded conditions in one of two task formats.
2. To measure need satisfaction with a validated need-threat scale immediately afterwards.
3. To test the format-by-condition interaction and report effect sizes with confidence intervals.

Topic 31: Guided and unguided internet-delivered CBT for adolescent anxiety

Aim: To synthesise trial evidence comparing therapist-guided and unguided internet-delivered CBT for adolescent anxiety.

1. To search at least three databases using a registered protocol.
2. To screen records with a second reviewer for a defined proportion of the results.
3. To assess risk of bias and synthesise anxiety outcomes, using meta-analysis only where trials are comparable.

New Data, Secondary Data or a Review: Which Route Fits?

The route you choose shapes the risks you carry. Compare the options below before you commit, and then match your design to a test with the statistical test selector, which saves time when you write the analysis plan.

Route	What you need	Main risk	Often suits
Online experiment or survey	Platform, participant pool, ethics approval, power analysis	Small sample or attrition between sessions	Bachelor’s and Master’s
Secondary data analysis	UK Data Service registration, statistical software, time to read documentation	Variables not measured the way your question needs	Master’s and PhD
Qualitative interviews	A small, well-defined group and time for transcription	Analysis takes longer than collection	Bachelor’s to PhD
Systematic review or meta-analysis	Protocol, second screener, reporting checklist	Your course may not accept a review	Master’s and PhD
Applied fieldwork	Gatekeeper agreement, sometimes a DBS check	Delay before data collection starts	Master’s and PhD

Whichever route you take, write your prediction in testable form. A hypothesis builder helps you separate the directional prediction from the null, which makes the analysis plan easier to defend.

Ethics and Approvals for a Psychology Dissertation

Ethics approval comes before data collection, not after. Your university committee applies the BPS Code of Human Research Ethics, and for online studies it also uses the British Psychological Society’s guidance on internet-mediated research, which covers consent, data security and the public or private status of online material.

Studies involving NHS patients, staff or data work differently. The Health Research Authority states that standalone undergraduate health and social care research needing NHS ethics review or HRA approval cannot take place, and that Master’s students should complete its student research toolkit to check eligibility, with the supervisor acting as chief investigator. You can read the full position in the HRA student research guidance. In practice, most student projects use non-clinical samples, secondary data or reviews.

Position correct as of October 2026; check the Health Research Authority for the current student research criteria.

Other approvals depend on your participants. Working with children needs an enhanced Disclosure and Barring Service check in England and Wales, with equivalent schemes in Scotland and Northern Ireland. Adults who may lack capacity to consent fall under the Mental Capacity Act 2005 in England and Wales, which makes them rarely suitable for student projects. All data collection must also follow UK GDPR, so decide which personal data you need, where you will store it and when you will delete it.

On many BPS-accredited courses, passing an empirical project is a condition for eligibility for Graduate Basis for Chartered Membership. Confirm with your course team whether a review or a secondary analysis counts before you build your plan around one.

How Study Level Changes Your Research Topic

Bachelor’s. Choose one clear hypothesis, an accessible sample such as students or community members reached online, and a validated measure that is free to use. A two-group experiment or a correlational survey with one or two predictors is enough. Keep the analysis simple, because a clean, well-reported small project scores better than an ambitious one that stalls.

Master’s. You can support more complex designs, such as moderation and mediation, multi-wave data, qualitative methods like IPA, or a registered systematic review. Secondary data analysis also fits well at this level. Show that you can justify each methodological choice against alternatives, and that you understand the limits of your design.

PhD. Doctoral work needs an original contribution, such as a new measure, a theoretical test that could falsify a model, or a programme of linked studies. A Registered Report format or a multi-study design often suits this level better than a single data collection. Your topic should connect to a live debate in the literature, not only to a gap in it.

Common Mistakes When Choosing Psychology Research Topics

• Inventing a questionnaire. A validated scale usually already exists, and examiners expect you to use it and report its reliability in your sample.
• Ignoring scale licences. Some well-known measures are copyrighted and charge for use, so confirm access before you design around one.
• Making causal claims from correlational data. Unless you manipulate a variable or use a longitudinal design, word your conclusions as associations.
• Underestimating attrition. Two-session and diary studies lose participants at every step, so recruit more than your power analysis suggests.
• Choosing a clinical topic without a recruitment route. If you cannot name where your participants will come from, the topic is not yet feasible.
• Testing too many hypotheses. Each extra test raises the risk of false positives, so decide your primary analysis before you collect data.

Questions Students Ask About Psychology Dissertation Topics

Can I repeat a published study as my dissertation?

Yes, if your course allows it, and replications are often easier to defend than a first-time design. Obtain the original materials or contact the authors, preregister your analysis plan, and add a small extension so your project contributes something new. Check your department’s marking criteria, because some reward originality of hypothesis more heavily than rigour of method.

Do I need a hypothesis for a qualitative project?

No. Qualitative work uses a research question and an aim, and the method decides how you analyse the data. Explain why a qualitative approach fits your question, and include a reflexive account of how your own position may shape your interpretation.

Should I use SPSS, jamovi, JASP or R?

Use whichever your department teaches and supports, because your supervisor can then check your work. SPSS is common in UK departments, jamovi and JASP are free and menu-driven, and R gives you scripts that make your analysis reproducible. For Bayes factors and measurement invariance, JASP and R are the practical options.

Can I study something linked to my own experience?

Often yes, and personal insight can improve your questions. It also creates bias and emotional risk, so discuss it with your supervisor early, avoid recruiting close friends, and plan how you will look after your own wellbeing during analysis.

Turning Your Shortlist Into a Proposal

Pick two or three topics that survive the opening questions, then test each against your timetable, your access to participants or data, and the measures you can legally use. Turning each into a precise question is the next step, and the research question generator gives you wording to react to. A short guide on how to write a research question explains why the narrower version usually wins.

If you want a second opinion on a shortlist, psychology dissertation support is available for planning, feedback and structure. When a topic is chosen and you need help shaping it into a plan your supervisor will approve, research proposal support covers that stage. Treat every idea above as a starting point for your own work, and follow your institution’s policy on any academic support you use.