Aim: The difference in effectiveness of online and offline interactions to overcome psychological problems is unexplored for groups like female, older, health worker and mental health treatment takers. Moreover, the link of psychological problems of the social interactors is not investigated which could be insightful for clinical intervention. The aim of this study is to fill these research gaps.Subject and Methods: We explored a large dataset having 11,333 Japanese peoples’ data amid the mild lockdown in 2020. To assess psychological problems, we used well-established and clinically validated scales. To compare between groups, we did the T-test and Mann-Whitney U Test and to explore links between variables, we built a Bayesian consensus network by bootstrapping data 10,000 times.Results: We find lower loneliness, depression, and psychological distress among the social interactors. However, interaction effectiveness appears to vary by psychological problem. Compared to only offline interactors, only online interactors have a lower loneliness, larger social network, and also higher COVID anxiety (p<.05). On the other hand, offline interactors have lower depression, psychological distress, and higher healthy sleep (p<.05). Surprisingly, despite doing both types of interaction, 54% of the mental health treatment takers are depressed which is higher than other groups. We also find varying links among psychological problems, lifestyle, and stressors depending on the interaction and people.Conclusions: Our findings suggest there is no single mode of interaction which can be effective universally for all people. Therefore, support to overcome psychological problems should be centered around people’s specific needs and problems.