Imagine: you have just received a request to assess the mental health needs of the survivors of a large-scale disaster. What are your major challenges: Logistics? Resources? Communication? Getting an overview of who is in the area, who has been assessed, and who still needs to be? And if needs are identified, getting the appropriate services in place?

Probably all of the above. In addition, it may be dangerous to enter the disaster area altogether, for example due to radiation risk.

Therefore, wouldn’t it help if you could conduct needs assessments via social networks like Facebook? If these give the same information as face-to-face interviews that you would conduct otherwise, it would be worthwhile as a low-cost-low-risk approach.

A research team from Israel set out to test whether a Facebook sample and a traditional face-to-face sample would show differences in mental health and disaster-related data after the 2011 Fukushima nuclear disaster.

They had 2 samples of Japanese survivors:

• a convenience sample of 133 people (66% women, mean distance to Fukushima 271 km)
• a Facebook-recruited sample of 40 people (38% women, mean distance to Fukushima 248km).

Participants in both groups completed self-report questionnaires on posttraumatic stress, disaster-related thoughts, life satisfaction, health, social support, and meaningfulness of life. Some of these were measured with a single question. Data collection with the convenience sample was at least partly via interview, but the extent of this is not exactly described (also, who conducted these interviews? Did the interviewers run a health risk after all?).

Main result: there were differences. The Facebook sample reported higher perceived social support, higher life satisfaction, higher self-rated health, and lower posttraumatic stress levels than the face-to-face sample. The size of the effects varied from small (for posttraumatic stress, health) to medium (for social support) to large (for life satisfaction).

My first thought was “The difference in stress levels is because there were many more females in the convenience sample, and they tend to have higher levels of PTSD.” However, the researchers have taken gender differences into account in their analyses.

The authors conclude that data collection via Facebook may lead to an underestimation of mental health symptoms. However, in my view (and the authors describe this partially as well), it is too early to come to conclusions. For example, the article doesn’t describe how the two groups were recruited. Maybe the Facebook sample was actively recruited while the convenience sample were people who somehow presented with a request for support? The two groups may have differed in various ways not accounted for.

It is possible that the Facebook sample truly had a lower level of stress responses (e.g. because of their higher levels of support). To test this, one could take a larger group of Facebook users and randomly assign them to face-to-face or via-Facebook data collection. If they do not differ in outcomes, then it may well be that we can approach Facebook users via Facebook and find other ways for approaching non-Facebook users.

Reference:

Ben-Ezra M, Palgi Y, Aviel O, Dubiner Y, Evelyn Baruch, Soffer Y, & Shrira A (2013). Face it: Collecting mental health and disaster related data using Facebook vs. personal interview: The case of the 2011 Fukushima nuclear disaster. Psychiatry research, 208, 91-3 PMID: 23200780