Peer-Review is Censorship not Quality Control

The establishment in psychological science prides itself on their publications in peer-reviewed journals. However, it has been known for a long time that peer-review, especially at fancy journals with high rejection rates, is not based on the quality of an empirical contribution. Peer-review is mainly based on totally subjective criteria of quality. Rather than waiting for three month, editors should just accept or reject papers and state clearly that the reason for their decision is their subjective preference.

I resigned from the editorial board of JPSP-PPID after I received one of these action letters from JPSP.

The key finding of our study based on 450 triads (student, mother, father) who reported on their personality and well-being in a round-robin design (several years of data collection, 100,000 dollar in research funding) was that positive illusions about one’s personality (self-enhancement) did not predict informant ratings of well-being. Surely, we can debate the implications of this finding, but it is rather interesting that positive illusions about the self do not seem to make individuals happier in ways that others cannot perceive. Or so I thought. Not interesting at all because apparently, self-ratings of well-being are perfectly valid indicators of well-being. So, if informants don’t see that fools are happier, they are still happier, just in a way that others do not see. At least, that was the opinion of the editor and as it had the power to decide what gets published in JPSP, it was rejected.

I am mainly posting the editorial letter here because I think the review process should be transparent and open. After all, these decisions influence what gets published when and where. If we pride ourselves on the quality of the review process, we shouldn’t have a problem to demonstrate this quality by making decision letters public. Here everybody can judge for themselves how good the quality of the peer-review process at JPSP is. That is called open science.

Manuscript No. PSP-P-2019-1535

An Integrated Social and Personality Psychological Model of Positive Illusions, Personality, and Wellbeing Journal of Personality and Social Psychology:  Personality Processes and Individual Differences

Dear Dr. Schimmack,

I have now received the reviewers’ comments on your manuscript. I appreciated the chance to read this paper. I read it myself prior to sending it out and again prior to reading the reviews. As you will see below, the reviewers and I found the topic and dataset to be interesting. However, based on their analysis and my own independent evaluation, I am sorry to say that I cannot accept your manuscript for publication in Journal of Personality and Social Psychology:  Personality Processes and Individual Differences.

The bottom line is that the strongest theoretical contribution the model would appear to produce is not currently justified in the paper and the empirical evidence presented regarding that contribution does not rise to the occasion to support publication in JPSP.

I’ll start by mentioning that both Reviewers commented on the lack of clarity of the presentation, and unfortunately, I agree. Reviewer 1 commented overall, “There were passages with a lack of clarity.” Just focusing on the Introduction, I read it multiple times to try to understand what you see as the central theoretical contribution. I found that entire literatures were overlooked (well-being) or mischaracterized (social psychological perspectives on well-being), and illogical arguments were advanced(p. 4). Terms were introduced without definition (e.g., hedonic balance) but later included in the statistical model, and – when later comparing the terms in the model to the literature reviewed – I found a lack of discussion of or justification for the paths that were actually tested and that seem to be at the heart of what you see as the main contribution to the literature (as indicated by the first paragraphs of the Discussion section). Echoing Reviewer 1’s more general point, Reviewer 2 commented specifically on the model, stating: “The authors should describe the statistical model in much more detail to make the statistical analyses easier to follow, more transparent, and replicable.” Again, I agree.

I bring that up at the outset of the letter because you will see me refer to lack of clarity here and there, below. But now I’d like to set aside writing to focus on the theoretical and empirical contribution, which are the heart of the matter.

Through a careful reading of the terms in the model, analyses, and the first paragraphs of the Discussion section, here is what I understand are the claims about the central theoretical and empirical contributions (by “central” I mean the contributions that would make this work cross the threshold for publication in JPSP:PPID): you believe (subjective) well-being has a “truth” to it in the way personality traits might, and a bias to it. You think the “truth” (public view) estimate of well-being is a more important outcome than the “bias” estimate. As such, you draw the conclusion that you have overturned a seminal paper and the field of social psychology’s perspective on well-being because the measure you care about, the “truth” estimate of well-being, does not correlate with self-ratings of positive illusions. (This conclusion appears to be drawn despite the fact that positive illusions about the self and self-reported well-being are indeed correlated, which replicates prior positive illusions literature.)

Given that the argument turns on this interesting idea about truth and bias estimates of well-being, I’ll focus on well-being. There is a huge literature on well-being. Since Schwarz and Strack (1999), to take that arbitrary year as a starting point, there have been more than 11,000 empirical articles with “wellbeing” (or well-being or well being) in the title, according to PsychInfo. The vast majority of them, I submit, take the subjective evaluation of one’s own life as a perfectly valid and perhaps the best way to assess one’s own evaluation of one’s life. So if you are staking the conclusion of your paper on the claim that in fact others’ agreement with a person about whether that person’s life is good is the best representation of one’s well-being, and researchers in the field should dismiss the part about the evaluation that is unique to the evaluator, then that needs to be heartily justified in the paper’s Introduction. The onus is on the authors to do that and I do not believe it is there.

Instead, from what I can tell you appear to be relying on an assumption that, because well-being is consistent with statistical properties of personality in that “wellbeing judgments show agreement between self-ratings and informant ratings (Schneider & Schimmack, 2009; Zou, Schimmack, & Gere, 2013) and are much more stable than the heuristics-and-bias perspective suggests (Schimmack & Oishi, 2005)” (p. 7), therefore the conceptual problem is the same as for measures of personality. It is not. It is of course well-established on theoretical grounds why personality traits are useful to assess from multiple perspectives. But for the question of well-being, this is literally about my subjective feeling about my life; on what grounds do others’ perspectives take a higher priority than the self’s? I agree that it is an interesting question to know if others can see my well-being the way that I do, but this so-called “truth” estimate speaks to quite a different research question than what most of the well-being research field would consider to be an important question. If you think it is important or even more important than the way it has been traditionally done (which I surmise you might, based on what appears to be the dismissal of 30 years of research on positive illusions and well-being in the Discussion), it is up to you to (a) define and measure well-being as it relates to the contemporary psychological literature, (b) explain why this subjective assessment should not be taken at face value but instead needs multi-rater reports to make accurate or meaningful inferences, then (c) explain why each of your predictors would map on to each of these two estimates (i.e., truth and bias) and (d) why those paths matter for the broader literature.

I do see where you talked about positive illusions and “positive beliefs” (which I think you equate with wellbeing but it was unclear) side by side in the introduction (e.g., p. 4), but not where you (1) recognized (a) positive illusions about personality and (b) wellbeing estimates as distinct constructs and (2) justified why one would be associated with the other.

If you make those arguments – situated in the contemporary literature on well-being, and reviewers for a future submission agree with the logic and potential theoretical contribution – the next hurdle of course is the empirical contribution. Assuming the models are correct (see both reviewers’ comments on this), this paper would make empirical contributions in its conceptual replications of prior findings and a few other interesting observations. But the biggest theoretical contribution you appear to want to claim is that “Overall, these results challenge Taylor and Brown’s seminal claim that mental health and wellbeing are rooted in positive illusions.” Yet, (a) you do present evidence that the link between positive illusions about the self and well-being as assessed by the self are correlated, as has been done previously in that literature, and (b) this conclusion appears to be drawn based on null effects using a measure that is not established (i.e., “truth”). (And please see Reviewer 1’s concerns about the cross-sectional nature of the findings as well as the fact that measures use few items.)

Overall, this dataset is rich and the idea of considering convergence and bias in well-being estimates is interesting. To produce a paper that will have a strong impact, I suggest you take a close look at your modeling approach (Reviewer 2), take a close look at your conceptual model itself (not the results) and map it on to the points in the literature that it most closely addresses (e.g., novel questions about separating well-being into truth and bias), and consider what additional evidence might bolster that theoretical or methodological contribution.

Additionally, Reviewer 1 commented on the framing of the paper, on antagonistic language, and on editorializing, and I agree on all fronts. The frame is much too broad, sets up a false dichotomy between social and personality psychology, and the evidence does not rise to the occasion to either (a) take down the paper the Introduction sets up as the foil (i.e., Taylor and Brown, 1988) or (b) allow personality psychologists to “win” the false competition between social and personality psychology about whether positive illusions contribute to well-being.

Other comments:

–              Please justify the use of these two sets of life evaluations but not hedonic balance as indicators of well-being, based on contemporary literature and evidence on well-being and how these should relate to one another. (I note that, incidentally, Schwarz and Strack include happiness judgments in their review of well-being.)

–              In the Method section, what was the timescale of the “hedonic balance” assessment. Was it “right now”? The past 24 hours? Two weeks?

–              Both reviewers were experts in SEM methods and personality; please do take a close look at their methodological comments, which were quite thoughtful and helpful as I considered my decision.

–              I had similar questions as Reviewer 1 regarding the fact that student gender was lumped together relative to mother and father reports, where gender is naturally separated. I agree that there is low statistical power to address this empirically but just wanted to let you know that this thought independently came up for two of us.

In closing, I would like to thank the reviewers for their constructive comments, and I look forward to reading more about this research in the future.

For your guidance, I have appended the reviewers’ comments, and hope they will be useful to you as you prepare this work for another outlet.

Thank you for giving us the opportunity to consider your submission.


Sara Algoe, Ph.D.

Associate Editor

Journal of Personality and Social Psychology:  Personality Processes and Individual Differences

1 thought on “Peer-Review is Censorship not Quality Control

Leave a Reply