4 ms·

After reviewing the report and the associated studies, there are no indicators that this is a causal relationship - just correlations. My immediate reaction is

After reviewing the report and the associated studies, there are no indicators that this is a causal relationship - just correlations. My immediate reaction is "why are we ruling out the possibility that they are on Instagram / Fb / (insert network of choice) because they do not have healthy EQ (emotional intelligence as a proxy for "mental health" here?)" I don't have skin in this fight, but I am extremely skeptical of "correlation over causation" with specious conclusions that justify certain opinions that are popular here on HN.

learnstats2 9y ago
Statistics rarely gives causal indicators (and arguably it would always be impossible to distinguish cause and effect in this case). So, this is an unhelpful criticism. Is there a correlation? Yes. Is there a causation? We can't know for sure, but we also can't rule it out and it is certainly worth considering and reporting on in lay terms.
- jm_l 9y ago
  Couldn't you settle this by controlling for initial mental health in two groups, having one group start using Instagram, and tracking any changes? If your point is that it would take work to separate correlation and causation then yes, agreed, but I don't see why you would think we can't know for sure.
  - extra88 9y ago
    That would probably be considered unethical because the expected change would cause harm. Instead, perhaps you could have group A stop using Instagram and group B continue using Instagram then compare. It would be important to look deeper to see if there are differences within group A between those who and did not replace using Instagram with another social platform.
  - learnstats2 9y ago
    In addition to the other concern, you don't have a good control here because you intend to select from people who don't (or do) currently use Instagram - but this is already a self-selecting group.
jrs95 9y ago
IMO it's probably both. Probably harmful for most people, but maybe only a bit, and for some it can be a terrible negative feedback loop.
rossdavidh 9y ago
Especially as an indicator of one social media platform vs another, it seems unwise in its design. Essentially, it appears that they are trying to gauge each platform's impact on young people's mental state by...asking those young people to guess at the answer. Hmmm...problematic as it is to ask someone (young or not) to evaluate their own mental state, asking them to evaluate which social media platform that they use is responsible for that, seems even more problematic. This does not look like the world's best experimental design, even with the same budget you could have done better.

learnstats2 9y ago

Statistics rarely gives causal indicators (and arguably it would always be impossible to distinguish cause and effect in this case). So, this is an unhelpful criticism. Is there a correlation? Yes. Is there a causation? We can't know for sure, but we also can't rule it out and it is certainly worth considering and reporting on in lay terms.

jm_l 9y ago
Couldn't you settle this by controlling for initial mental health in two groups, having one group start using Instagram, and tracking any changes? If your point is that it would take work to separate correlation and causation then yes, agreed, but I don't see why you would think we can't know for sure.
- extra88 9y ago
  That would probably be considered unethical because the expected change would cause harm. Instead, perhaps you could have group A stop using Instagram and group B continue using Instagram then compare. It would be important to look deeper to see if there are differences within group A between those who and did not replace using Instagram with another social platform.
- learnstats2 9y ago
  In addition to the other concern, you don't have a good control here because you intend to select from people who don't (or do) currently use Instagram - but this is already a self-selecting group.

jrs95 9y ago

IMO it's probably both. Probably harmful for most people, but maybe only a bit, and for some it can be a terrible negative feedback loop.

rossdavidh 9y ago

Especially as an indicator of one social media platform vs another, it seems unwise in its design. Essentially, it appears that they are trying to gauge each platform's impact on young people's mental state by...asking those young people to guess at the answer. Hmmm...problematic as it is to ask someone (young or not) to evaluate their own mental state, asking them to evaluate which social media platform that they use is responsible for that, seems even more problematic. This does not look like the world's best experimental design, even with the same budget you could have done better.