Belief is the weakest part of this sentence. Plenty of people believe in gratitude's power and practice it twice a year.
The effect isn't mystical and it isn't guaranteed. What's reasonably established is narrow: people who regularly write down specific things they're thankful for report ruminating less and sleeping better. That's a modest claim, and modest claims are the ones that hold up.
So don't believe in it. Run it. Three specific lines, most nights, for a month, then judge from your own data.
A thing you've tested doesn't need believing.