A clinical psychology dissertation needs a testable hypothesis built around one measurable relationship between named variables in a defined population — for example, whether health anxiety predicts avoidance of medical appointments in adults, not the vaguer aim of “exploring anxiety.” Below are worked research questions and hypothesis sets across common clinical sub-fields, the mistakes examiners flag most, and how to move from a hypothesis to a defensible design.
Why a clinical psychology hypothesis is harder to write than a general one
Clinical psychology sits closer to real diagnostic categories, real distress and real treatment claims than most branches of psychology, which raises the bar for a hypothesis in three ways. First, constructs need a named, validated measure attached before the hypothesis can be tested — “anxiety” on its own is not testable, but “scores on the GAD-7” is. Second, most undergraduate and taught master’s dissertations in this area cannot recruit clinical populations directly: NHS research ethics approval has not been available to standalone undergraduate research since 1 September 2021, so a UK clinical psychology dissertation is far more often built on a non-clinical or sub-clinical community sample, secondary/archival data, or a vignette and survey design than on patients recruited through a service. Third, the British Psychological Society’s Code of Human Research Ethics expects any research that touches distress, self-harm ideation or diagnostic labels to justify its sampling and disclosure procedures explicitly — a hypothesis that implies clinical recruitment without a feasible access route is a proposal a supervisor will send back before you write a word of the method.
The practical consequence: a strong clinical psychology hypothesis in a taught dissertation usually studies attitudes, beliefs, help-seeking behaviour, or symptom severity in the general population or in a specific non-clinical group (students, a particular age band, a particular occupation), and frames the clinical relevance explicitly in the discussion rather than pretending to have tested a clinical sample it never recruited.
The anatomy of a testable hypothesis
Every workable hypothesis in this field has four visible parts, and a marker checking your proposal or introduction chapter will look for all four:
- A named independent variable and dependent variable, each attached to a specific instrument or operational definition (not just a word like “stress” or “wellbeing”).
- A defined population that you can actually access within your ethics approval and timeline.
- A stated direction where the literature supports one (a directional or one-tailed hypothesis, e.g. “higher X predicts lower Y”) or an explicit non-directional hypothesis where the evidence is mixed or this is the first study of its kind in that population.
- A null hypothesis stated alongside it, even if you never write it out in the final report — it is what your statistical test actually evaluates.
Illustrative example of the difference: “This study explores the relationship between rumination and depression” is an aim, not a hypothesis. “Higher trait rumination, measured by the Ruminative Responses Scale, will be associated with higher depressive symptom severity, measured by the PHQ-9, in adults aged 18–65” is a testable, directional hypothesis with its null implied (there is no association between the two).
Worked hypothesis sets by clinical sub-field (illustrative)
The following are fully worked, illustrative examples — built to show structure, not lifted from a real study — across sub-fields that UK clinical psychology dissertations commonly cover without needing direct clinical-service access.
Health anxiety and healthcare avoidance
RQ: Does health anxiety predict avoidance of routine medical appointments in UK adults?
H1: Higher scores on the Short Health Anxiety Inventory will be associated with higher self-reported avoidance of routine GP appointments.
H0: There is no association between health anxiety scores and appointment avoidance.
Perceived stigma and help-seeking intentions
RQ: Does perceived public stigma towards depression predict intention to seek professional psychological help?
H1: Higher perceived stigma will be associated with lower help-seeking intentions, measured by the General Help-Seeking Questionnaire.
H0: Perceived stigma is not associated with help-seeking intentions.
Therapeutic alliance and early treatment engagement (secondary-data design)
RQ: Does early therapeutic alliance, rated after session two, predict attendance across a six-session low-intensity CBT course?
H1: Higher early alliance ratings will be associated with fewer missed sessions.
H0: Early alliance ratings are not associated with session attendance.
This design is normally only feasible using anonymised, aggregated service data already held by a placement provider or a published dataset — not primary data collection with current clients, which raises consent and dual-relationship issues a taught dissertation timeline cannot usually clear.
Emotion regulation strategy and self-harm ideation in a non-clinical sample
RQ: Is habitual use of expressive suppression, rather than cognitive reappraisal, associated with higher self-harm ideation in a non-clinical student sample?
H1: Higher expressive suppression scores on the Emotion Regulation Questionnaire will be associated with higher scores on a validated ideation screening measure.
H0: No association exists between suppression scores and ideation scores.
Any hypothesis touching self-harm ideation needs an explicit, pre-agreed safeguarding and signposting procedure in the ethics application before a supervisor will approve it — state this in the proposal itself, not only in the ethics form.
Perfectionism and burnout in trainee healthcare students
RQ: Does socially prescribed perfectionism predict burnout symptoms in trainee nursing and paramedic students?
H1: Higher socially prescribed perfectionism, measured by a validated multidimensional perfectionism scale, will be associated with higher burnout scores, measured by a validated burnout inventory (check its current licence and cost before committing — some widely used burnout scales are fee-based, others are freely available for student research).
H0: There is no association between socially prescribed perfectionism and burnout scores.
This is a convenient design for a clinical psychology dissertation because trainee healthcare students are a genuinely accessible, non-patient population, and the clinical relevance — burnout in future clinicians affecting patient care quality — can be argued clearly in the discussion without needing clinical access at the data-collection stage.
Mistakes examiners flag most in this section
- Writing an aim and calling it a hypothesis. “To explore the link between X and Y” cannot be tested statistically; it has no direction and no stated relationship.
- Stacking too many hypotheses. A taught dissertation typically supports one primary and one or two secondary hypotheses well — five or six spreads the sample thin and the discussion chapter cannot do justice to all of them.
- Claiming directionality without evidence. A directional hypothesis needs at least one cited study pointing that way; if the literature is genuinely mixed, a non-directional hypothesis is the honest and defensible choice.
- Implying clinical access the ethics application does not support. A hypothesis about “patients with a diagnosis of X” needs a realistic recruitment route through the NHS or a specialist service — rare at undergraduate level and slow even at master’s level — or it needs reframing to the population you can actually reach.
- Leaving the construct unoperationalised. Naming the instrument in the hypothesis itself (not just in the method chapter) shows a marker you designed the study backwards from a measurable outcome.
From hypothesis to design
Once the hypothesis is fixed, the design questions that follow are: which statistical test actually matches this hypothesis and this level of measurement, and which theoretical framework explains why you expect the relationship in the first place. Our guide to choosing the right statistical test for a psychology dissertation walks through exactly this decision once your hypothesis is fixed, and our comparison of theoretical frameworks for a psychology dissertation is worth reading before you finalise the wording, since the framework you choose is what justifies the direction you predict.
If you want to see how a hypothesis set like the ones above sits inside a complete dissertation, from introduction through to discussion, our annotated walkthrough of a psychology dissertation shows the full structure section by section. And if the instrument you want to build your hypothesis around needs checking for cost or licence restrictions first, see which psychology scales you can actually use as an undergraduate before committing to it in your proposal.
Writing this section with Tesify
Tesify helps you turn a rough research idea into a properly structured hypothesis set — IV, DV, direction, null — checked against what a realistic UK undergraduate or master’s ethics application can actually support, and drafts the surrounding introduction and rationale in your own voice rather than a generic template. Over 9,000 students have used Tesify across more than 15,000 dissertation chapters, and every dissertation on the platform is 100% written by you, with the tool helping you structure and check your own thinking rather than generating a submission on your behalf.
Frequently asked questions
Can I study a clinical population directly for my undergraduate dissertation?
Rarely. Since 1 September 2021 the Health Research Authority has not accepted standalone undergraduate research for NHS ethics review, so direct recruitment of NHS patients is not normally available at that level. Most clinical psychology dissertations instead use non-clinical or sub-clinical samples, secondary data, or vignette-based designs, and discuss clinical relevance explicitly rather than testing it directly.
What is the difference between a directional and non-directional hypothesis?
A directional (one-tailed) hypothesis predicts the direction of a relationship, such as “higher X is associated with lower Y,” and needs prior evidence pointing that way. A non-directional (two-tailed) hypothesis predicts only that a relationship exists, without specifying its direction, and is the safer choice where evidence is mixed or absent.
How many hypotheses should a clinical psychology dissertation have?
Most taught dissertations manage one primary hypothesis and one or two secondary hypotheses well within their word count and sample size. More than that usually means the discussion chapter cannot address each one with enough depth to score well.
Do I need to state a null hypothesis in the report itself?
Check your own department’s template — some ask you to state H0 explicitly alongside H1, others only expect it implied by the statistical test you report. Either way, you should know what your null hypothesis is, because it is what your chosen test is actually evaluating.
Can my hypothesis be about attitudes rather than clinical symptoms?
Yes, and for an undergraduate dissertation this is often the more realistic route — attitudes towards mental illness, stigma, help-seeking intentions and public perceptions of specific conditions are all legitimate clinical psychology topics that do not require clinical-service access.
What ethics body reviews a UK clinical psychology dissertation?
For research that does not involve NHS patients, staff or premises, your own university’s departmental or faculty research ethics committee normally reviews the application, following BPS Code of Human Research Ethics principles. Any design that would need NHS involvement should be raised with your supervisor early, since HRA approval is not available at standalone undergraduate level and is a lengthy process even at master’s level.
Should I name the exact statistical test in my hypothesis?
Not in the hypothesis wording itself, but the test should already be decided before you finalise it — a hypothesis phrased around a correlation needs a correlational test, and one phrased around a group difference needs a test built for comparing groups. Deciding the test after writing the hypothesis in vague language is a common source of mismatched methodology.
What if my hypothesis is not supported by my results?
A non-significant or contrary result is not a failed dissertation — markers reward critical discussion of why a hypothesis was not supported (sample size, measurement limitations, population differences from the original studies) far more than they reward a hypothesis that happened to be confirmed.
Can I use a validated scale I found online without permission?
Not automatically — some validated psychology scales are free for student use, others require registration or a fee, and a few are restricted to qualified clinicians. Check the specific licence for any instrument before building your hypothesis around it.
How does a hypothesis connect to my theoretical framework?
The framework is what explains why you expect the relationship your hypothesis states — for example, a cognitive model of anxiety is what justifies predicting that a specific thinking pattern will relate to a specific anxiety outcome. Choosing the framework before finalising the hypothesis usually produces a more defensible, less arbitrary prediction.