This book is an easy read, it contains some good ideas and potential information, and it provides some food for thought with regard to its main topic, so I’m not about to advise anyone against checking it out just based on the quibbles that I may have had (some of which may be due to decisions by the publisher/editor to oversimplify some of the content and clearly aim for a pop science flavour to it).
Still, I am afraid that quite a number of points bothered me about this book. I found myself not quite buying some of the interpretations of experimental manipulations. I understand that the descriptions are simplified, so it may very well be that additional manipulation checks just weren’t mentioned, but the issue extended beyond experimental assumptions to interpretations of behaviour in other studies.
If my issues with the book were limited to interpretations, it wouldn’t be quite so bad, but I was also left absolutely unsure about whether the magnitude of the effects were really reported in a fair manner. (For individuals not familiar with statistical methods: the vast majority of psychological differences still result in huge overlap between groups, so simply observing a “difference” doesn’t mean that this difference also has practical relevance.)
I am less concerned about the author’s own studies, but his interpretations of other studies are unfortunately questionable. For example, he introduces the infamous Facebook study, in which emotional content within news feeds was manipulated. The author describes the findings this way: “The results were unambiguous; people’s moods moved towards the contents of their manipulated news feeds. Those who saw more sad posts subsequently posted more sad updates themselves. Those who saw more happy posts showed a similar mirroring.”
Having looked at the actual effects reported in the journal article, these statements are extremely misleading. Yes, the study found “significant” statistical differences. However, these differences were only statistically significant because they had several hundred thousand “participants” to magnify these differences. To draw a direct example from the article: “When positive posts were reduced in the News Feed, … the percentage of words that were negative increased by B = 0.04% (t = 2.71, p = 0.007, d = 0.001)”.
In other words 0.04% = 1 word for every 2,500 words typed across many, many people! That is such a small effect that it is surely hardly worth mentioning, much less presenting as “unambiguous”.
The non-critical description of the Facebook study wasn’t the only point in the book where I found myself questioning either the effect size or the statistical power of the study design of a cited study, which suggests that the author may not have paid attention to these details in at least some of the findings reported in the book (particularly when they weren’t based on studies from his own lab). On top of that, some of the conclusions drawn from fMRI studies seem vastly overstated (with some of them sounding like reverse inferencing).
Casual readers may not place much importance on the next point, but as someone with an interest in the research, I was disappointed in the lack of in-text citations. (I also found it strange that the text cites a particular study by Wilhelm Hofmann several times throughout the book — but at one point introduces Kathleen Vohs, lists her credentials, then describes the study conducted by Hofmann in Germany with words such as “after she studies people’s behavior in their natural environs”, even though it is quite probable that she only played a role in writing the manuscript.)
Overall, I think that the book has many important ideas, but I’m unsure about the strategy of writing in this supposed accessible manner. And for a book meant to summarise other studies, it didn’t do a good job of critically presenting these studies, thus leaving me with too much uncertainty regarding the ideas.
Emotional success: The motivational power of gratitude, compassion and pride (David DeSteno)