Everyone says exercise helps depression. Fish oil is supposed to reduce anxiety. Cutting sugar should improve your mood. Vitamin D helps with seasonal blues.
Maybe. But does it work for you?
The only way to know is to measure.
Why guessing doesn't work
"I feel better" is subjective. Memory is unreliable. Placebo effects are real. And when you're trying multiple things at once, you can't tell what's actually helping.
Standardized assessments solve this. They give you numbers instead of impressions. Numbers you can compare week over week, before and after, with and without.
This is how clinical research works. You can do the same thing for yourself.
The basic method
1. Establish a baseline. Take an assessment before you change anything. Do this for 2-3 weeks to get a stable starting point.
2. Change one thing. Add exercise. Start a supplement. Cut alcohol. Whatever you're testing.
3. Keep taking the assessment. Weekly is usually enough. Same day, same time, same conditions.
4. Compare. After 4-8 weeks, look at your scores. Did they change?
5. Confirm (optional). Stop the intervention for a few weeks. Do scores return to baseline? If so, the effect was probably real.
This isn't perfect science (there's no control group) but it's far better than guessing.
What to measure
For most lifestyle experiments, these assessments work well:
PHQ-9: Depression symptoms. Captures low mood, energy, sleep, appetite, concentration. Good for testing interventions that target motivation and energy.
GAD-7: Anxiety symptoms. Captures worry, tension, restlessness. Good for testing interventions that target stress and calm.
DASS-21: Depression, anxiety, and stress in one assessment. Takes slightly longer but gives you three separate scores. Good when you're not sure which dimension will be affected.
Pick one and stick with it. Switching assessments mid-experiment makes comparison harder.
Exercise
The research here is strong. A 2024 meta-analysis found exercise significantly improves PHQ-9 and GAD-7 scores. Effect sizes are comparable to antidepressants for mild to moderate symptoms.
What to track:
- Type of exercise (cardio, strength, yoga, walking)
- Frequency (times per week)
- Duration (minutes per session)
- Intensity (light, moderate, vigorous)
Start simple. "30 minutes of walking, 4 times per week" is specific enough to test. Track your mental health scores weekly.
Most studies show effects within 4-8 weeks. If you don't see improvement by then, the type or dose of exercise might not be right for you, or exercise might not be your key lever.
Diet
Harder to isolate than exercise, but possible. Common experiments:
Cutting alcohol. Easy to test because it's binary. Stop for 4 weeks, track scores.
Reducing sugar/processed food. Define what you're cutting specifically. "No added sugar" or "no fast food" is testable.
Mediterranean diet. Research links this pattern to lower depression risk. High in fish, vegetables, olive oil, nuts. Low in processed food.
Elimination diets. Some people react to specific foods. If you suspect something (gluten, dairy, etc.), cut it for 4-6 weeks and watch your scores.
Keep a simple food log alongside your assessments. You don't need to track calories. Just note major changes. "Started eating fish twice a week" or "stopped drinking soda" is enough.
Supplements
The evidence varies wildly. Some supplements have decent research support. Others are mostly marketing.
Stronger evidence:
- Omega-3 fatty acids (fish oil): Multiple studies show modest effects on depression. Typical dose: 1,000-2,000mg combined EPA/DHA daily. Allow 8-12 weeks.
- Vitamin D: Relevant if you're deficient (common in winter or if you're indoors a lot). Get tested first if possible.
- Magnesium: A meta-analysis found significant effects on depression scores. Typical dose: 200-400mg daily.
Weaker or mixed evidence:
- St. John's Wort: Some evidence for mild depression, but interacts with many medications. Check with a doctor first.
- SAMe: Used in Europe for depression. Some positive trials, but expensive.
- Probiotics: The gut-brain connection is real, but specific recommendations are still unclear.
Important: Supplements aren't regulated like drugs. Quality varies. Buy from reputable brands. And tell your doctor what you're taking, because interactions are real.
How to run a clean experiment
The biggest mistake is changing too many things at once. If you start exercising, taking fish oil, and cutting alcohol simultaneously, you'll have no idea which one helped (or if they cancelled each other out).
Better approach:
1. Pick one intervention
2. Define it precisely (what, how much, how often)
3. Commit to 6-8 weeks minimum
4. Track weekly with the same assessment
5. Look at the trend
If you want to test multiple things, do them sequentially. Exercise for 8 weeks. Then add fish oil for 8 weeks. Then try cutting alcohol. Each phase gives you clean data.
What counts as a real effect?
On most assessments, a 5-point change is considered clinically meaningful. That's the threshold researchers use to say "this intervention made a real difference."
Smaller changes (2-3 points) might be real but could also be noise. Look for consistency across multiple weeks rather than putting weight on any single score.
Direction matters more than magnitude. A steady downward trend in depression scores over 8 weeks is meaningful even if individual weeks fluctuate.
Example: Testing exercise
Week 0-2: Baseline. Take PHQ-9 weekly. Average score: 8 (mild depression).
Week 3: Start running 3x/week, 20 minutes each.
Week 4: PHQ-9 = 7
Week 5: PHQ-9 = 6
Week 6: PHQ-9 = 8 (bad week at work)
Week 7: PHQ-9 = 5
Week 8: PHQ-9 = 4
Week 9: PHQ-9 = 4
Week 10: PHQ-9 = 3
Average weeks 8-10: 3.7 (minimal symptoms)
That's a clear signal. The score dropped from 8 to under 4. Running appears to help.
Caveats
This isn't medical treatment. If you're struggling significantly, work with a professional. Self-experimentation is useful for optimization, not for managing serious symptoms.
Placebo effects are real. Expecting something to work can make you feel better temporarily. That's why the optional "stop and see if it returns" step matters.
Correlation isn't causation. Maybe your scores improved because of the supplement. Or maybe it was the season changing, a stressor resolving, or random variation. Be humble about conclusions.
Some things take longer. Exercise often shows effects in 4-8 weeks. Some supplements need 8-12 weeks. Don't give up too early, but don't keep going indefinitely either.
The point
Your body and brain are unique. What works for the average person in a study might not work for you. What works for your friend might not work for you.
The only way to know is to test it yourself, systematically, with real data.
Mental health assessments give you that data. Use them.
Start your experiment today
Survey Doctor makes it easy to track the PHQ-9, GAD-7, or DASS-21 on a consistent schedule. Establish your baseline, start your intervention, and watch the data. Take your baseline assessment now.