"The Cultural Narratives of Francophone and Anglophone Quebecers: Using an Historical Perspective to Explore the Relationship between Relative Deprivation and Collective Well-Being"
Evelyne Bougie; Esther Usborne; Roxane de la Sablonnière; Donald M Taylor
British Journal of Social Psychology
Dear Dr. Usborne
Thank you for submitting your manuscript entitled 'Cultural narratives of Francophone and Anglophone Quebecers ...' to the British Journal of Social Psychology. We have now received two reviews of the manuscript, and I have had a chance to read it carefully. Unfortunately, on the basis of the reviews, I am unable to accept your paper for publication in our journal at this stage. However, both the reviewers and myself believe that a substantively revised manuscript might be publishable, and this is an option you may wish to consider. I am going keep this editorial letter brief because the reviewers have written extremely detailed and well-informed responses to your article (see below). I would invite you to consider their comments and recommendations carefully should you choose to submit a revised draft for further consideration.
In brief, we can all see the potential in this manuscript. The attempt to situate the study of identity, group threat and relative deprivation in their historical context is bold and interesting. In addition, although I am not entirely convinced that your methodology is able to sustain the kinds of claims that you want to make, it does represent an innovative attempt to include an historical dimension that social psychologists often talk about but rarely study empirically.
As you will see, the reviewers raise several substantive issues, which concern the clarity of your key constructs and measures, the clarity of your theoretical explication of the relationship between RD, identity and well-being, and the meaning and operationalisation of the concept of 'historical context'. Reviewer 1 implies that study 1 might be unnecessary to report: at the very least, I would expect you to address some of his or her concerns in this respect. Both reviewers make numerous, helpful comments about other aspects of design, analysis and theoretical grounding. Usually, with such an extensive set of revisions, I would go with a straight 'reject' decision: when you start the revision you may find yourself wishing I had! On this occasion, however, I am also mindful of the novelty and potential value of the underlying research project. Moreover, I appreciate that the study of historical processes presents us with all kinds of methodological challenges (which is
probably why it is so rarely attempted). Accordingly, I feel that a more flexible and sympathetic approach is warranted and offer you the opportunity to revise and resubmit.
To conclude: let me wish you the best of luck with preparing your revised manuscript (should you opt for this route). I would also like to thank the two reviewers for delivering such detailed, intellectually engaged and comprehensive responses. This not only made my job easier, but the reviews were interesting to read in their own right.
Reviewer #1: MS# BJSP1141
This is a very interesting and novel line of research. I share the authors' enthusiasm for the idea that the collective identity of large social groups (e.g., cultural, national, etc groups) is in part the product of a shared understanding of the group's history. I also find the concept of identity clarity to be intriguing. It seems a potentially important addition to our understanding of the psychology of collective identity. In addition, the data in Study 2 seems interesting and supports the potential importance of historical events/periods that are currently understood by the group to be times of relative deprivation.
However, that said, I have a number of concerns with the current manuscript, not the least of which is that I am not entirely sure what the authors have actually demonstrated. I will detail my concerns below, but it is my opinion that this paper is currently not ready for publication in BJSP although I think a future version of the paper could be.
PRIMARY CONCERNS & QUESTIONS.
1. Study 1:
1a. The real "action" in the paper is all in Study 2. It appears to me that Study 1 is more of a "pretest" to establish the relevant history to be considered in the real Study (Study 2). Although the methodology in Study 1 is interesting (the cultural narrative), it really does not measure any of the critical concept outlined in the introduction (relative deprivation, collective well-being, or collective identity clarity), and the study does not test any of the theoretical or empirical questions set up in the introduction.
1b. Further, I do not understand the "second goal" (p.11) of Study 1. What exactly is the value in determining which narrative chapters were perceived to be intergroup threats? As described in the paper, "intergroup threat" can emerge from a variety of events or situations other than collective relative deprivation. This seems especially true in the case of the "nadir" experience which is described as "characterized by extremely negative emotions such as despair, disillusionment, terror or guilt." Most of these are not the emotions usually thought to emerge from RD. Guilt is not at all consistent with RD and emerges from feeling illegitimately overcompensated relative to the outgroup or from recognition that the ingroup has unjustly harmed the outgroup. Despair might result from the death of a visionary leader which does not seem the same as group-based RD. Terror might arise from a natural disaster; again not the same as collective RD.
So, one point is that intergroup threat is not necessarily synonymous with RD; thus how is the goal of determining which periods represent intergroup threats relevant to the primary question set out in the introduction -- the relationship between historical RD and current well-being.
However, an additional point is that I am not sure why measuring group threat in Study 1 was necessary at all. It appears that all of the chapters that were discovered in Study 1 were used in Study 2, independent of whether they were perceived as threats or not in Study 1. Also, the degree of deprivation felt when thinking about each chapter/periods was directly measured in Study 2, and it was these measures that were used to test the hypotheses in Study 2. It is not clear why it was necessary/valuable to know which chapters were seen as threats in Study 1.
1c. Also, it is not clear to me how these "intergroup threat" data were organized/analyzed. On page 13, we are told that a specific request was made for participants to report a nadir experience, but later we are told that the authors analyzed the extent to which participants "SPONTANEOUSLY described events that represented times of intergroup threat" I am not sure how something that was specifically requested can be considered spontaneously described.
1d. Finally, the results of Study 1 are described in an odd mix of broad summary statements, a smattering of numbers and a single t-test. Also, these different types of information often seem to contradict each other. For example, we are told that Francophones averaged 5.6 chapters and Anglophones 3.8 chapters. Later, we are told this difference is highly significant. Yet, on page 15, the summary conclusion appears to be "A large number of participants, regardless of their group, reported events associated with the five following historical periods:." What systematic process was used to get us from a highly significant difference to this agreed upon a set of 5 chapters (which are then used in Study 2). Similarly, according to the descriptions on pg. 14, it appears that Anglophones by in large ignored everything from the conflict between French and English colonizers and the rise of Francophone nationalism 300 years later. Francophones, on the other hand, appear to
describe a number of different events and periods in these 200 years. How do we reconcile this with the claim on page 15 that, "both Anglophones and Francophones pointed to similar periods as making up their cultural narrative"?
The first sentence of the section on "Perceptions of Threat" on page 15 provides another example of why I am confused. The sentence indicated that "similar events emerged in the narratives of both (groups)" but that the "attention given to the reported historical events differed between the (groups)." We are not explicitly told how the "emergence" of an event was determined and how the amount of "attention given" to that event was then determined. What are the criteria for an event to "emerge" and how was the level of "attention given" measured? We are told that Francophones "spent more time on the early conflict" but in the previous pages we are told that 95% of Francophones described this chapter, while only 81% of Anglophones described his period. It would appear to me that this would indicate a 14% difference in the "emergence" of these events - contradicting the earlier statement about the similarity in emergence. Are we to understand that Francophones also
talked for more time, and/or provided more details about this period? Similarly, we are told that "Anglophones accorded relatively greater attention to more contemporary historical events, such as the Quiet Revolution era." However, on the previous pages we are told that 80% of Francophones included this period as at least one chapter in their narrative, while only 78% of Anglophones included a chapter for this period.
The point of all this is that I am not at all sure how the "rich" data from the narratives was used/analyzed/organized to produce evidence for the claims made on pages 13 through 16.
To summarize..First, despite its very interesting methodology, I am not sure that Study 1 tells us much that is relevant to the stated purpose of the research. It seems primarily to provide an interesting method of pretesting to establish the relevant "chapters" to be described in Study 2. Secondly, it is not at all clear to me how the data from these cultural narratives were distilled down to the 5 Chapters that produced the historical narrative used in Study 2. There appears to be just as much evidence for inconsistencies in the narratives of the two groups as for the consistency that the authors claim produced these chapters.
2. Clarifying the key concepts?
2a. There appears to be a lot of "slippage" in the terms/concepts. In a number of cases, two or more concepts appear to be equated in ways that I am not comfortable with. For example, it appears that the terms 'intergroup threat' and 'collective relative deprivation' are being used interchangeably. As I have already described, it seems that intergroup threat is a much broader term than collective relative deprivation and I believe the story presented here would be much more convincing if the authors limited their discussion to the more specific concept of RD.
At times, the authors also seem to be equating 'collective well-being' and 'collective esteem.' While self-esteem is often considered to be one element of well-being, more complete measures of well-being usually include other aspects (most commonly something related to a sense of satisfaction, but there are a number of models of subjective well-being that include much more). Also, the definition of collective well-being on page 2 seems identical to a definition of collective self-esteem. If these are, in fact, meant to be the same, I am not sure what is gained by having two labels for the same concept.
There appears to be two solutions here. The first would be to restrict the terminology being used to the more specific concepts that are being considered. The second would be to provide a clearer distinction between these concepts and describe more clearly the authors' understanding of how they are related. For example, in the abstract the authors state "In a period of intergroup threat, relative deprivation might serve to clarify collective identity." In this statement it appears that intergroup threat and relative deprivation are different concepts with intergroup threat perhaps leading to or creating feelings of relative deprivation. However, this possible distinction and relationship between the two concepts is not clearly developed. Rather, the two terms seem to be used interchangeably elsewhere.
I would also suggest that the author's be consistent in labeling these concepts as collective (e.g., "collective relative deprivation" not "relative deprivation"). I also think that it is wise to use the term "collective" rather than "social" (p. 3) to refer to group-based processes. While Social Identity Theory did initially use the term "social" to refer to group-based processes, many researchers suggest that this might imply that interpersonal and intrapersonal processes are somehow not "social" in nature.
2b. The concept of "collective identity clarity" remains a bit unclear to me. Given that this is the most novel of the paper's primary concepts, this one needs particular attention in terms of definition and elaboration. When the concept is introduced (bottom p.3 - top of p. 4), it is described only as ".clarify one's collective identity." and in the subsequent example it is contrasted with "defining" identity (".might actually sever to clarity or define the individual's Jewish identity."). It is not at all clear how a clear (or clarified) collective identity might differ from an unclear (or unclarified) collective identity.
When the concept is presented in more detail later we are told it is ".the extent to which one has clear or confident beliefs about one's collective identity." It is later described as "a clear understanding of this identity." However, I am not sure what would constitute clear beliefs or a clear understanding. Does this mean that one can list a larger set of attributes (ingroup stereotypes), or that these attributes are more clearly integrated so that they appear to describe a consistent and organized characterization of the ingroup or a typical member? Or perhaps it means that the person has "clarified" the content of the collective identity by more formal enacting and testing of the group stereotypes. Also, what exactly is the relationship between clarity and confidence? Exactly what is the individual confidence about? Is it that stronger feelings of confidence when describing the content of the group identity (i.e., its attributes, experiences, relationships with
other groups) is 'evidence' of greater identity clarity, or is it that confidence itself 'is' clarity?
It might also be useful to distinguish identity clarity from the strength of identification with the group. If clarity, in part, determines the degree to which one feels the group is valued and respected (collective esteem), shouldn't it also result in greater identification with the ingroup. Alternatively, the stronger one identifies with the ingroup the more one might spend time investigating and elaborating the content of the identity. Thus, stronger ingroup identification may lead to greater identity clarity.
2c. Finally, the confusion around terms also led to some concerns about the measures used in Study 2. For example, the concept of Collective Well-Being/Collective Esteem is defined as "the degree to which one perceives one's collective identity as having worth, respect and value." Yet, the measure of this concept is the Private and Membership subscales of the Collective Self-Esteem Scale. While the Private subscale seems consistent with the definition, the Membership subscale does not. As the authors recognize, this subscale deals with judgments of the individuals own worth as a member of the group (it really seems to have little to do with the worth, respect and value of the group itself). More importantly, I couldn't find anything in the introduction to explain why thinking that the ingroup was more relative deprived at a past period in history should lead someone to feel like they are now a more valuable member of the ingroup. In addition, wouldn't the Public
subscale, which measures the perceived value and respect given by others to the ingroup be more consistent with the definition the authors provide for Collective Well-Being? This subscale seems to be an "evaluative dimension of collective self-esteem" (the reason given for including the other two subscales). Finally, the two subscales are combined without much explanation for why. Although the alpha for the combined scale is good, the authors might want to provide other evidence that these two subscales are best represented as a single scale.
I am also not entirely clear about how the Collective Identity Clarity was measured. We are told that participants rated "how confident or sure they felt about their overall ratings of what happened to their group during each chapter." What are these "overall ratings of what happened."? Are these the measures of RD or were there other questions about the details of the events of the chapter? If these were in fact simply measures a one item measure of how confidence the participant is with their ratings of the amount of RD, then I am not at all sure that they should be described this as measuring a separate concept of "Collective Identity Clarity." Further, with clarity measured as 'the confidence that one has about the degree of deprivation that group was suffering at that particular time,' I am not at all sure that these data can support claims that this research "points to the potential importance of threatening periods in history for defining that group's identity
and thus creating higher levels of collective well-being." I don't believe that confidence about the level of deprivation is evidence that this deprivation defines the group identity.
3. What is meant by "historically contextualized"?
The wording in a number of places (including the abstract) leads me to think the researchers were actually measuring RD and identity clarity at different periods in a group's history. This, of course, is an extremely interesting possibility, especially for real social groups like Anglophone or Francophone Quebecers. The only ways I could imagine this being done would be to acquire a "time machine" or design an ingenious manipulation that could effectively convince participants that the current period in history is consistent with another particular period in the past. In fact, and obviously perhaps, this is not what the authors have done. Instead, they have measured the perceptions of participants who are situated in the current political/social/historical context but who are asked to consider/think about the degree of deprivation or privilege their group enjoyed at different periods in history. Thus, the research is not set in or contextualized by different historical
contexts, as all participants are clearly immersed only in one historical context (the present). Rather the research takes the unusual (and perhaps interesting) approach of having participants think about their group's deprivation relative to a consistent comparison outgroup at a variety of different times in the past. While this is an interesting idea and I am intrigued by the possibilities that such an approach might offer, I am pretty sure that doing this does not "examine perceptions of relative deprivation and collective identity clarity at different key periods in a group's history." (p. 17).
Similarly, and perhaps more generally, I am also a bit unsure about exactly what role history is thought to play here. I am absolutely convinced of the general point that the current collective identity of a group (especially a large societal group) is, in part, a result of members' shared representation of the group's history (especially their history of intergroup relations with relevant comparison groups). However, I am not sure that we need always measure collective identity while reminding people about particular points in their group history. So, it is not clear to me what the authors mean when they say "research that sets out to empirically explore predictors of collective identity and well-being most often measure these variables in a decontextualized manner, divorced from any historical context." If the shared understanding of the group history actually serves to create the current identity (it produces the content and psychology of the current identity)
then measuring identity in the present is not divorced of history, but includes it. Specifically reminding people about particular points in history might well alter current understandings of the collective identity, but I am not sure that this is any more "historically contextualized" than measures that don't involve specifically raising the salience of a particular point in time. I think the authors key point that some points in history are particularly important in generating the current identity is evidence of this. Some points in history are used in the construction of the current identity and are "there" in the current identity whether we explicitly remind people about them or not. Other points in history are not "there" in the current identity and reminding people of these might alter the current identity in ways that produce a different conception of the current ingroup identity. The point is, reminders of history can be used to influence current perceptions of
the group identity, but I am not sure that these remainders make the content of the described identity "more contextualized" per se.
4. Questions about the proposed model?
I agree with the statement "the historical context in which feelings of relative deprivation are measured must be taken into account" (p.3) if this means that we need to consider that feelings of collective relative deprivation at any point in time are influenced by the historical trends in feelings of relative deprivation for that group. When I began the paper, I thought that the point effort was going to involve an effort to combing RD based on temporal comparisons with the ingroups situation (an interesting area of RD research) with intergroup comparisons. That is, I thought the authors were going to combine these two ideas about RD to create a model that included both temporal and intergroup RD. For example, if my group is currently gaining status relative to the comparison outgroup compared to the relative status it has enjoyed in the past, I am likely to feel less relatively deprived than if my group held the same current relative position, but were loosing
ground compared to the past. That is, my current feelings of relative deprivation are not only influenced by current deprivation relative to the outgroup but also by how the current level of RD compares to previous levels of RD compared to that outgroup. The key idea is that current feelings of RD influence current feelings of collective esteem, but that current feelings of RD are in part determined by comparisons with past RD relative the same comparison outgroup.
This model provides a reasonable explanation for why those who perceive higher degrees of collective RD at a salient point in the past should indicate higher collective esteem in the present. This is because the high RD in the past makes current RD seem lower, thus leading to higher current collective esteem. Similarly, this explanation could also provide an explanation for the finding described on p. 31, that if participants are reminded of a period in which the group was relatively advantaged "a positive event" compared to the outgroup this would make current comparisons more negative (e.g., stronger feelings of current RD) and thus lower current collective esteem.
However, the authors appear to be proposing that current levels of RD are not involved, that the level of RD that one associates with a particular point in the past directly effect current collective esteem, and this relationship is partially mediated by the degree to which the individual has a clear representation of what the collective identity was at that point in the past. The temporal placement of all this seems somewhat strange to me. It seems reasonable that current collective identity clarity (assuming my understanding of this term is correct) should influence current collective esteem (and perhaps the even broader personal and collective well-being). However, I do not understand why a clear understanding of "who we use to be" should have this effect. In fact, if I have a very clear understanding of who we use to be and a very unclear understanding of who we currently are, being reminded of this past clarity of identity would make the current lack of
clarity particularly painful.
In fact, the authors seem to change the temporal story in the final pages of the paper. On page 28, it appears that the authors are describing current feelings of threat for Francophones versus Anglophones by referencing the small % of Francophones in Canada and North America. However, their own data appears to show that if anything Francophones feel less RD than Anglophones when thinking about the current chapter of the history narrative. Similarly, it appears that the discussion of identity clarity and "defining collective identity" are described as current understandings of the current (not past) group identity. For example, the example of the Jewish identity on p. 32 seems to be clearly talking about "defining the identity of a modern Jewish person." The past event creates identity clarity and definition in the current collective identity. Yet, in the early part of the paper and in the way the concepts were measured, identity clarity is described as clarity of
the past collective identity (what we use to be). My broad point here is that after reading this paper, I am still not clear how past RD and past Identity Clarity should influence current Collective Esteem without reference to current RD or current Identity Clarity.
MINOR & MORE SPECIFIC QUESTIONS & SUGGESTIONS.
Generally, I found parts of the introduction sections quite repetitive. I believe that pages 2-9 could be tightened somewhat.
P. 3, line 4 & 5: The authors claim that "numerous others have found." but then sight only three studies with no use of "e.g.,". Why not say 3 studies, or are there others (in which case an "e.g.," is necessary)?
P. 3, line 14: I don't think that research can "imply" something. Research "provides evidence for" something and that evidence might "suggest" something. Authors might imply that the research suggests something, but the research doesn't "imply."
P.3, line 16-19: I am confused by this sentence. How can research "demonstrate" something but never have "systematically tested" it? Do you mean that the research findings "suggest"?
P. 12, Participants: There is no information on how and where the participants were recruited.
P. 18, 1st line of Measures section: I am not sure what ".devised in a more structured manner." means. Perhaps the problem is with the work "devised."
P. 27, middle paragraph: In the middle of the paragraph, the authors describe a drop in beta from .27 to .19 as "completely disappeared", because .19 is not significant. This equation of lack of significance with no effect at all is indeed surprising when later in the paragraph we are told that a Sobel test that failed to reach statistical significance was important and worthy of interpretation. In my opinion, you can't have it both ways. Either you completely ignore non-significant results (they are the same as zero) or you don't.
P.31, first paragraph: If previous research has shown that thinking about historically positive events produces a negative relationship between RD and collective esteem, why did the current research not show this relationship during periods of apparent relative privilege for each of the two groups?
P.31, first line of Methodological Contributions. Both studies don't employ the Cultural Narrative method. It is used only in Study 1.
In summary, despite all my criticisms, I found this work to be intriguing and there may very well be an important contribution here. The narrative approach used in Study 1 is interesting and may have a number of important benefits, but we need a much clearer understanding of how this data is used/analyzed/organized in order to accept the claims made about the findings from this approach. But most importantly the authors need to clarify and make clearer the temporal aspects of the model and be consistent and clear in the definition and subsequent measurement of the key constructs.
Reviewer #2: Comments for authors of BJSP1141: The cultural narratives of Francophone and Anglophone Quebecers: Using an historical perspective to explore the relationship between relative deprivation and collective well-being
There was much to like about the manuscript. I really like the main thrust of the authors' argument that understanding the meaning of and relationships between social psychological variables requires a grasp of historical context (though would add that Gergen's arguments, as I understand them, go somewhat further than the need to place just our variables and measures in historical context). I also very much like the combination of methods across studies, which was innovative and puts into practice the sort of methodological plurality that many other papers only pay lip service to.
Nevertheless, I do have some non-trivial reservations about recommending the paper for publication in its present form. These mostly centre on theoretical issues and on the operationalisation of key constructs, and are likely to take some substantive work to address.
At several points, I felt that I was just on the verge of fully understanding the theoretical case for expecting the relationship between relative deprivation (RD), collective identity clarity (CIC) and well-being to differ depending on historical frame, but was then left short-changed. I understand (and agree with) the arguments relating to intergroup threat providing a basis for collective identity at specific points in time, but I struggled to see a clear rationale for why this should change the nature of the relationship between RD, CIC and well-being, rather than simply increasing mean levels of each. As it stands, the argument reads that under conditions of heightened threat, threat starts to positively predict well-being via CIC, so a score of 4 on the RD scale will tend to be associated with similarly high scores on the other variables - but it's not clear why under conditions of low threat at a mean level, the same score on RD would no longer be associated with high
scores on the other variables. Subjectively, the level of perceived threat (as measured on the RD scale) is the same, so why is it doing different things at different times? The answer must be something along the lines that particular historical timepoints affect the nature and meaning of 'threat' as measured on the RD scale, as well as its mean levels. As it stands, I don't get the sense of why - or indeed, how - this might occur. In short, the authors need to elaborate why greater mean levels of RD also changed the nature of RD and its relationship with other variables.
This leads on to another gripe I had about the RD concept. For all the build-up peppered with references to historical meaning and contextualisation, it was a bit of a disappointment that the rich, varied and multi-faceted nature of threat was reduced to a six-point scale labelled relative deprivation. To my mind, threat - especially as it pertains to long-standing conflicts involving struggles over language, culture, identity, political voice and economic inequalities - is not one thing, and is certainly not fully encompassed by the relative deprivation concept. RD does map quite closely onto the notion of status or esteem threat, but this can be quite different from threats that relate to identity itself which have been shown to play a crucial role in struggles involving ethnolinguistic groups, as a sometimes parallel but distinct issue to social inequality. Indeed, the lead-in to the study made me half expect a discussion of how notions of deprivation and threat are
constituted and remade over the course of a group's history, and the RD scale seemed somewhat anaemic as a result. In view of the previous points relating to the need to address meaning as well as means, the paper would therefore benefit from discussion of what constitutes 'threat' or 'relative deprivation', and how this itself is historically contingent and variable in meaning. As it stands, the RD framework sits awkwardly with the rhetoric regarding historical contextualisation.
A related issue is that these judgements of RD and CIC are all made with the benefit of historical hindsight - all participants know what happened next, as it were, and specifically know that 'we' are still around to tell the tale. How might this play out if RD and CIC were measured contemporaneously during a 'nadir'? Does the positive relationship between RD, CIC and well-being hold up when you don't know how things worked out subsequently, or when you aren't asked to consider their relationship in explicit comparison to other points in history? To use a concrete example that recurred in the manuscript, would this relationship hold for Jews who were questioned in the midst of the Holocaust, rather than 60 years later?
I thought that the CIC measure itself was also quite problematic, not least of all because it wasn't entirely clear how it was measured. Was it a single item in each case? More fundamentally, it didn't actually seem to refer to the clarity of identity as such; rather, it related to how sure or certain participants were about what happened to the group. The idea of CIC brought to mind things like ingroup homogeneity, self-stereotyping, or entitativity, or some other subjective indicator of ingroup integrity. I didn't get a sense that this was really what was being addressed, so much as certainty about what happened to the group at that time - clarity of historical knowledge, rather than of identity. The paper would benefit from more detail on the nature of these measures, how they map on to the notion of identity clarity, and why the authors chose to operationalise it in this manner, when the above, well-established constructs seem intuitively (to me at least) to map more
closely on to the notion of collective identity clarity.
One indirect reason for doing so is that some of the authors' arguments put me in mind of other research that looks at the connection between forms of threat and well-being. In particular, the rejection-identification model (e.g., Branscombe, Schmitt, & Harvey, 1999) suggests that minority group members' experiences of discrimination can have negative direct consequences for well-being, but positive indirect consequences for well-being through identification with the minority group. Although they focus on discrimination rather than relative deprivation, and on identification rather than identity clarity, there is clear resonance with the approach suggested here. As such, the authors may want to do some work to elaborate and differentiate their model from the rejection-identification model.
One other substantive issue is that, theoretical considerations aside, the evidence for a mediated relationship between RD and well-being via CIC is actually pretty weak. For Francophones, the mediation is only partial. What is accounting for the residual positive association between RD and well-being? Does this mean that CIC is useful but ultimately insufficient to account for this association? For Anglophones, the indirect effect isn't even reliable. There may be some statistical reasons for this, and the authors may wish to try bootstrapping procedures to test indirect effects, rather than the Sobel test (which can be less stable and/or sensitive for small sample sizes: Shrout & Bolger, 2002; see also Mackinnon et al., 2002), and might provide more compelling results
Otherwise, the authors might consider other ways of testing their hypotheses. For example, one implication of the authors' reasoning is that the correlations between RD, CIC and well-being should be stronger at some historical periods than at others. However, they don't directly test this. Showing that this was the case would at least provide some evidence for this 'weaker' version of the historical contingency hypothesis.
I found the paper to be a bit repetitious in places, especially in Introduction and in parts of the Results and Discussion sections. A reduction in length focused on cutting out redundancy would benefit the paper.
The authors could explain the rationale and benefit of the developmental trajectory analysis in Study 2. For one thing, what does this show that couldn't be gleaned from a glance at the means in the table, and a 2 (group) X 5 (time-point) mixed ANOVA to look at how the linear and curvilinear trends fit over time-point, and differ between (i.e., are moderated by) group? Second, I was puzzled by term 'trajectory'. As I understand it, this sort of terminology is typically reserved for measures taken at multiple points in time (i.e., longitudinal designs), as opposed to perceptions of different points in history measured at one time. In what sense do the present data display a trajectory, any more than any other within-subjects variable in which all measures are taken at once? It might be fairer to clearly and consistently say that these are perceived historical trajectories of the group, or something along those lines.
Branscombe, N. R., Schmitt, M. T., & Harvey, R. D. (1999). Perceiving pervasive discrimination among African Americans: Implication for group identification and well-being. Journal of Personality and Social Psychology, 77, 135-149.
MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A comparison of methods to test the significance of the mediated effect. Psychological Methods, 7, 83-104.
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New procedures and recommendations. Psychological Methods, 7, 422-445.
* Please submit your revised paper via the Journal's online peer review system, Editorial Manager: http://bjsp.edmgr.com/ * Include a separate document listing the changes made in the new version of the paper. This document should not reveal your identity.
* Avoid sending PDF files if possible as these can create problems in the final stages of the peer review process (PDF figures and tables are acceptable, however). Most word processing formats are supported.
Visit the BPS Journals website (www.bpsjournals.co.uk) if you would like more detailed information on the Journal's submission guidelines for new and revised papers.