Tromholt recruited (via Facebook, of course) 1,095 Danish participants, who were randomly assigned to one of two conditions. The ‘treatment group’ were instructed not to use Facebook for one week, and were recommended to uninstall the Facebook app from their phones if they had it. At the end of the study, 87% of the treatment group reported having succesfully avoided Facebook the whole week. Meanwhile, the ‘control group’ were told to continue using the site normally.
The results showed that the treatment group reported significantly higher ‘life satisfaction’ and more positive emotions vs. the control group (p < 0.001 in both cases). These effects were relatively small, however, for instance the group difference in life satisfaction was 0.37 on a scale that ranged from 1-10...
This is a nice little study, but in my mind it doesn’t prove all that much. The trial wasn’t blinded – i.e. the participants of necessity knew which group they were in – and the outcome measures were all purely subjective, self-report questionnaires.It's this last point that I want to pick up as well. I worry that much of what is observed in this study is a Hawthorne effect - that the participants who were asked to give up Facebook for a week anticipated that the study was evaluating whether it increased their happiness, and reported what the researcher expected to see. The outcome measures were all self-reported measures of life satisfaction and emotions, which are easy for the research participants to manipulate (whether consciously or not). The author tries to allay this concern in the paper:
The critical point here is whether the participants have formulated their own hypotheses about the effects of quitting Facebook and that these hypotheses, on average, are pointing in the same direction. On the one hand, the individual formulation of hypotheses may have been facilitated by the pretest and the selection bias of the sample. On the other hand, the participants’ hypotheses may not be pointing in the same direction due to the fact that the direction of the effects found in the present study is not self-evident. Hence, there may be limited experiment effects at stake. If experiment effects did affect the findings of the present study, they might even turn in the opposite direction because people, in general, perceive Facebook as a source to positive feelings.I'm not convinced. Especially since, as this Guardian article on the research notes, the study was performed by the Happiness Research Institute. I'm not sure how any of the research participants could miss the significance of that and not realise that the study expected an increase in happiness to result.
Having said that, there are supplementary results reported in the paper that might partially allay those concerns. The effects on happiness were greater for those who were heavier users of Facebook, which is what you would expect to see if the effect is real, and this is not something that could be easily spoofed by the actions of participants.
One final word: it could be possible to improve this study by moving away from, or supplementing, the subjective wellbeing (life satisfaction) questions, such as by measuring levels of the stress hormone cortisol in the research participants before and after giving up Facebook. Of course, this would increase the cost of the study, because it could no longer be run solely online. Something for future research perhaps?