Reacción a "A week without social media can reduce mental health symptoms in young people, according to a study"

Is the study based on solid data and methods?

"The study has some methodological strengths, although the results do not support all the conclusions as expressed in the summary.

There are two main strengths. The first is the measurement of psychological variables using EMA (Ecological Momentary Assessment), which involves collecting data in real time throughout the day using momentary reminders and electronic diaries. As these measures do not rely on memory or retrospective collection, they are generally considered less vulnerable to bias or distortion.

The second strength is the collection of information on mobile device use through statistics generated by the device itself, rather than relying solely on self-reports, which are known to be unreliable, as demonstrated by the low correlation between subjective estimates and objective records".

How does it fit in with previous work? What new insights does it provide?

"The main novelty lies in the methodological advances in recording. As for the results, they are very similar to those already known.

First, there is little correlation between objective usage time and emotional or psychological well-being. Only subjective perceptions show some relationship, which is often interpreted as meaning that people who already consider their use of the device to be inappropriate tend to feel less satisfied with their overall psychological state. However, this relationship disappears when objective indicators are used.

The other result presented as relevant is the reduction in depressive and anxiety symptoms in those who voluntarily participated in the social media “detox” phase. This effect, as previously observed in earlier studies, is very modest, with effect sizes between d = -0.05 for loneliness and d = -0.44 for depression and anxiety. It should be remembered that, in absolute terms and using the usual standards, 0.3 is considered a small effect, 0.6 a medium effect, and 0.9 a large effect, so the results fall within the range of small or small to medium effects. Furthermore, the percentages that appear in the summary and tables, such as a 16.5% reduction in anxiety, are meaningless, as the depression and anxiety scales do not have an absolute zero.

These effects can be expressed in another way. For example, a d = -0.44, the maximum observed in this study, implies that 62% of people undergoing ‘detoxification’ would have a lower level of depression than a randomly selected person who did not participate in the intervention, which is only 12% above chance.

In relation to the above, the study also notes that those who started with higher levels of anxiety or depression benefited more from detoxification. However, the initial sample had very low levels, which limits the margin for improvement. In other words, those who already had few symptoms could not improve, while those who were worse off could. It is therefore inevitable that the improvement is concentrated among those who started out in a worse initial state. It is surprising that such an overinterpretation has passed peer review".

Are there any important limitations to consider?

"Definitely. Beyond the interpretive nuances about effect sizes, there are several significant limitations.

The first is that participation in the ‘detoxification’ phase was voluntary. Not only was there no control group, but the intervention group was self-selected. In intervention studies, self-selection is considered bad practice, as it introduces bias by allowing only the most motivated or those with the highest expectations to participate.

The second limitation is related to the first. The ‘detoxification’ could not be applied under double- or single-blind conditions. Participants knew they were in the intervention and could intuit its purpose. A basic principle in this type of study is to prevent participants from knowing or guessing the condition they are in or the purpose of the intervention. In this case, achieving this was virtually impossible. The authors themselves acknowledge this in the limitations section, stating that “participants' behavior may have been affected by reactivity to being monitored and by self-selection, which limits causal validity.” However, acknowledging this does not solve a problem that directly affects the conclusions, which are much more visible than the limitations in the body and summary of the article.

The third limitation is the absence of a control group. This means that the reduction in anxiety and depressive symptoms could be due to external factors, such as spontaneous remission, and not necessarily to the intervention. Consequently, it is not possible to establish causal inferences".

How relevant is this study in clinical practice?

"Unfortunately, in practical and clinical terms, its relevance is very modest, and there is a risk that it will join the long list of inconclusive and easily overinterpreted studies that abound in this field.

The main merit is methodological, thanks to the use of EMA and objective indicators of use. In terms of effects, it can be concluded that people who already view their use of mobile phones and social media negatively and who have high levels of dissatisfaction could benefit from reducing it. In the general population, however, an intervention of this type has, at best, an almost imperceptible effect. In all likelihood, expectation biases, self-selection, and the lack of a control group inflate the results, which are already small. This leads to the conclusion that the findings suggest the opposite of what is claimed: mass interventions to reduce the use of social media or mobile devices are very unlikely to be effective or cost-efficient".