Gratitude journalling became ubiquitous on the basis of a small number of studies, and the subsequent research is worth looking at because it is a good case study in how findings travel.
The original work
Early experiments asked participants to record things they were grateful for, weekly or daily, and compared them to groups recording hassles or neutral events.
The gratitude conditions reported better mood and, in some studies, other benefits including sleep and exercise.
The studies were small, the samples were mostly students, and the outcomes were self-reported by people who knew which condition they were in.
Which is not a criticism of the researchers — it is standard for early work — and it is a reason for caution about how far the findings were extended.
What subsequent research found
Meta-analyses have generally found small effects on wellbeing measures, smaller than the original studies suggested.
Effects on depression and anxiety symptoms specifically have been weaker still, and several reviews concluded the evidence does not support gratitude interventions as a treatment for either.
Comparisons against active control conditions — writing about anything positive, or about daily events — have frequently found little difference, which suggests the specific gratitude content may not be the active ingredient.
The frequency finding
One of the more interesting results.
Studies comparing weekly to daily practice have found weekly performing better in some cases.
The proposed explanation is adaptation — doing it daily makes it routine and the content repetitive, which reduces the effect.
Which runs directly against the standard advice to do it every day, and it is the sort of finding that does not travel because it complicates a simple message.
Who it works for
Some evidence suggests effects differ by baseline. People already high in dispositional gratitude gain less, which is unsurprising.
There is also evidence that for some people, particularly those experiencing genuine hardship, gratitude prompts produce guilt or a sense of invalidation rather than benefit.
Which is worth naming, because the framing of gratitude as universally positive leaves people who feel worse assuming they are doing it wrong.
The critique worth taking seriously
Gratitude practice directs attention toward what is good in a situation, which is useful when the situation is acceptable and something is being missed.
Applied to a situation that genuinely should change, it may reduce the discomfort that would otherwise motivate change.
This concern has been raised in organisational contexts particularly, where gratitude programmes have been introduced in workplaces with genuine problems.
The evidence on whether this actually occurs is thin, and the mechanism is plausible enough to warrant attention.
What survives
Writing expressively about experience has its own literature, separate from gratitude specifically, with reasonably consistent findings on wellbeing.
Which suggests the act of structured reflection may carry more of the benefit than the gratitude framing does.
Expressing gratitude to another person — as opposed to recording it privately — has produced somewhat stronger effects in several studies, and the social component is the plausible explanation.
Where I land
It is free, it takes minutes, the downside is minimal for most people, and the realistic expectation should be a small effect rather than a transformation.
Doing it weekly rather than daily, writing in some detail rather than listing, and occasionally telling somebody directly are the versions with the better evidence.
And if it makes you feel worse, that is a documented response rather than a failure, and stopping is a reasonable decision.
The positive psychology context
Worth knowing since gratitude research emerged from a specific movement with specific commitments.
Positive psychology was founded partly as a corrective to a field focused on pathology, which was a reasonable observation.
It also generated a body of interventions promoted with considerable confidence on the basis of early studies.
Several flagship findings from the field have since failed to replicate at their original effect sizes, which is not unique to it and has been particularly visible there because the claims travelled so far.
The replication context
Psychology generally went through a period of reckoning about replication, and interventions research was affected.
Small samples, flexible analysis choices and publication of positive results only all inflated reported effects.
Preregistration and larger samples have become more common in response, and newer studies in this area generally report smaller effects than older ones.
Which is the field working correctly rather than a scandal, and it means older confident claims should be discounted rather than dismissed.
The version I would actually suggest
Weekly, in some detail, about specific things rather than categories, and occasionally directed at a person rather than a page.
What to expect
A small improvement in mood, unreliable, more likely if the practice is varied rather than repetitive.
Which is worth the few minutes and is not going to change a difficult situation, and expecting otherwise is where the disappointment originates.