Which Mental Health Scale Should You Use in a Psychology Thesis? (Canada, 2026)

For most Canadian psychology theses measuring depression and anxiety, the PHQ-9 paired with the GAD-7 is the strongest default: both are public domain, brief, validated in Canadian primary care and community samples, and free to reproduce in full in your appendix. If your study needs one instrument covering three constructs at once, the DASS-21 is the best-supported runner-up, also free and widely used in Canadian undergraduate and clinical samples. A nursing thesis measuring burnout rather than mood faces the same licence-versus-validity trade-off among a different instrument family, worked through in the comparison of which burnout scale to use in a nursing thesis.

What about the Rosenberg Self-Esteem Scale and the Perceived Stress Scale?

Reach for these two when your construct is self-esteem or perceived stress specifically, rather than a mood or anxiety disorder symptom cluster. The Rosenberg Self-Esteem Scale is a ten-item, public-domain measure of global self-worth first published in 1965 and still one of the most widely cited instruments in social and developmental psychology research; its brevity and free availability make it a common secondary or covariate measure alongside a primary outcome. The Perceived Stress Scale (PSS-10) measures the degree to which situations in a respondent’s life are appraised as stressful, which is conceptually distinct from a symptom-count measure like the DASS-21 stress subscale, and is the better choice when your research question is specifically about appraisal and coping rather than symptom severity.

Both instruments have French-language Canadian validation studies available, which matters directly if your sample is drawn from Quebec or another bilingual setting; confirm you are citing the validated French version rather than translating the English items yourself, since an unvalidated translation undermines the reliability evidence you are relying on.

Why does licence status matter as much as validity?

Because a validated instrument you cannot legally reproduce in your thesis appendix, or cannot afford to administer to your full sample, is not usable regardless of how strong its psychometric properties are. The BDI-II and the STAI are both extensively validated and still assigned in many Canadian courses, but both are commercially licensed: Pearson charges a per-use fee for the BDI-II, and Mind Garden licenses the STAI per administration, which becomes a real budget line for a sample of a hundred or more participants. Check with your library first; some Canadian university libraries hold institutional access to commercial instruments through their test collections, which can eliminate the per-participant cost if your sample size is within the license terms.

When should you use the PHQ-9 and GAD-7?

Choose this pairing when your study needs a brief, low-burden screen for depression and anxiety severity rather than a full diagnostic instrument, and when you want results directly comparable to the primary-care and population-health literature that also uses these two tools. Both were developed and validated against DSM criteria, both take under three minutes to complete, and both are scored by simple item summation, which keeps your data-cleaning section short. Their brevity is also their limitation: a nine- and seven-item screener will not capture the same nuance as a longer clinical interview, so state explicitly in your limitations section that you are measuring symptom severity, not making a diagnosis.

When does the DASS-21 fit better than two separate screeners?

Choose the DASS-21 when your research question genuinely needs the stress dimension alongside depression and anxiety, or when you want one instrument with three subscales rather than the added complexity of managing and reporting two separate tools. Developed by Lovibond and Lovibond and made freely available for research and clinical use, the DASS-21 has strong validation data in Canadian undergraduate and community samples and produces three subscale scores from twenty-one items, each scored 0 to 3 on a self-report severity scale over the past week.

Report each subscale’s Cronbach’s alpha separately in your results, since a combined-scale reliability figure obscures whether one of the three subscales performed weakly in your specific sample; the thresholds for what counts as an acceptable alpha, and what to do about a weak subscale, are covered in the explainer on what counts as an acceptable Cronbach’s alpha.

When should you use the K10 instead?

Choose the Kessler Psychological Distress Scale (K10) when your thesis compares your sample against Canadian population data, since it is the instrument Statistics Canada uses in the Canadian Community Health Survey to measure psychological distress nationally. Using the K10 lets you situate your own sample’s distress scores against published national and provincial benchmarks in your discussion chapter, which two-screener or clinical instruments generally cannot offer. It is ten items, publicly available, and quick to administer, though its single total-distress score is less differentiated than the DASS-21’s three subscales.

Are the BDI-II and STAI ever still the right choice?

Yes, most often when your thesis needs direct comparability to a specific body of prior published research that used the licensed instrument, or when your supervisor’s lab has an existing institutional license or normative dataset built on it. The STAI’s separate state and trait anxiety scores are genuinely useful when your design manipulates a short-term anxiety-inducing condition and you need to distinguish momentary anxiety from a participant’s baseline trait level, a distinction the GAD-7 does not offer. Budget the licensing cost into your ethics application and your research budget from the start, not after your Research Ethics Board has already approved a protocol that assumed a free instrument.

Does your Research Ethics Board care which instrument you choose?

Yes, indirectly: your REB reviews the burden and risk your instrument places on participants, and a longer, more clinically loaded instrument such as the BDI-II or STAI may draw closer scrutiny of your distress protocol, referral pathway and debriefing procedure than a brief screener does. The approval process itself is the same Tri-Council pathway regardless of which instrument you choose, covered in the guide to REB approval and TCPS 2 rules in Canada, but budget extra review time for a study using a clinically loaded instrument with a vulnerable population.

Do digital or app-based versions of these scales change anything?

The underlying psychometrics do not change simply because a screener is delivered through a survey platform such as Qualtrics or REDCap instead of on paper, but the mode of administration is itself something you must report and, where the original validation was on paper, justify. Several of these instruments now have separately published validation studies for computerized or online administration, and citing the mode-specific validation, rather than assuming paper-based psychometrics transfer automatically, is expected in a methods chapter that reports an online study. Where you built a custom digital form rather than using a licensed platform’s existing implementation, note any deviation from the original item order or response format, since even minor formatting changes have been shown to shift response patterns on self-report scales.

How does reproducing the instrument in your appendix work?

For public-domain instruments, PHQ-9, GAD-7, K10, Rosenberg and the DASS-21, you may typically reproduce the full instrument in your thesis appendix with a citation to the original validation source, though confirm your specific university’s requirement, since some schools of graduate studies still ask for a formal permission letter even for public-domain instruments. For copyrighted instruments such as the BDI-II or STAI, you generally cannot reproduce the items themselves; instead, cite the instrument, describe its structure and scoring in prose, and direct readers to the publisher for the actual item content. The general copyright logic is the same one that governs reproducing any figure or table from a published source in your thesis, covered in the guide to do you need permission to use a figure in your thesis.

How many participants do you need to power a study using one of these scales?

The scale itself does not set your sample size; your effect size, your design and your planned analysis do. If you are comparing group means on a DASS-21 subscale or correlating K10 scores with another variable, the sample-size logic is the same one used across quantitative psychology research generally, worked through in the guide to how many participants a thesis actually needs. What the instrument does change is your minimum: a scale with a documented floor or ceiling effect in your population may need a larger sample to detect a true difference reliably.

How do you write the instrument-selection paragraph in your methods?

Name the instrument, its author and validation, its licence status, and why it fits your construct better than the alternatives you considered. Here is a worked example you can adapt.

Depression and anxiety symptom severity were measured using the Patient Health Questionnaire-9 (Kroenke, Spitzer & Williams, 2001) and the Generalized Anxiety Disorder 7-item scale (Spitzer et al., 2006), both public-domain instruments validated in Canadian primary-care samples. These were selected over the Beck Depression Inventory-II and the State-Trait Anxiety Inventory on the basis of administration burden and licensing cost across a target sample of 150 participants, and because their brevity supports the study’s planned online administration format.

Can Tesify help you write your instrumentation section?

Choosing between a free screener and a licensed clinical instrument is a decision your budget and your research question settle together, but writing the instrument-selection justification, the psychometric reporting paragraph and the appendix framing in the precise register a psychology committee expects is exactly the drafting work Tesify is built to help with.

Draft your instrumentation section with Tesify

If you are also weighing which AI writing tool to draft the surrounding methods prose in, the comparison of the best AI tools for writing a thesis in 2026 covers the same licence-and-fit logic applied to software rather than a psychometric instrument.

Frequently asked questions

Can I modify the wording of a public-domain instrument for my study?

Generally no, not without documenting and justifying the change, since altering item wording invalidates the published psychometric properties you are relying on. If you must adapt items, for instance translating them, report this explicitly and note that your version’s psychometrics are not identical to the original.

Is the PHQ-9 diagnostic on its own?

No. It screens for and grades depression symptom severity; a formal diagnosis requires a clinical interview. State this distinction explicitly in your methods and discussion chapters to avoid overstating what your data show.

Do I need clinical training to administer these screeners?

Self-report screeners like the PHQ-9, GAD-7, DASS-21 and K10 do not require clinical training to administer, but your REB protocol should still specify a referral pathway for participants who screen positive for significant distress.

Which instrument is best for an online or remote study?

The shorter public-domain screeners, PHQ-9, GAD-7 and K10, translate most cleanly to an online survey format, since their brief item count reduces drop-off compared with the forty-item STAI or the licensing complications of administering a copyrighted instrument outside a controlled setting.

Can I compare my results directly to a published Canadian norm?

Only for instruments with published Canadian normative data at your population level; the K10 is the strongest option here because Statistics Canada publishes national and provincial distributions from the Canadian Community Health Survey.

Do I cite the original validation paper or the version I actually used?

Both where they differ: cite the original validation study for the instrument’s psychometric provenance, and separately cite any Canadian validation study if you are relying on Canadian-specific reliability or normative data.

What if my supervisor insists on the BDI-II despite the cost?

Ask whether your department or your supervisor’s lab already holds an institutional license or a bulk-purchased set of forms, which is common, before assuming you must pay the per-use fee yourself.

Should I report Cronbach’s alpha for a validated instrument even though it is already published?

Yes. Report the alpha calculated in your own sample alongside the originally published value, since reliability can shift meaningfully across populations, languages and administration formats.