Do social dominance orientation and right wing authoritarianism similarly predict both explicit and implicit attitudes?

Dear Authors,

I finally had the chance to read the revised manuscript entitled “Do Right-Wing Authoritarianism and Social Dominance Orientation Predict Implicit Attitudes?”.

I am very pleased with your revision of the manuscript. You followed very carefully the suggestions the other reviewers and I gave you. Particularly, I find that the introduction makes it clear now what your objectives and hypotheses are, as well as the relevance of your work. The importance of investigating implicit associations has also become evident.

The method section also underwent significant and positive changes. The description of the methodological framework of the original study is much more comprehensive and easy to understand, and the analytical plan is well-developed.

As a result, I would only like to congratulate you on the great work you have done on this manuscript. I wish you success with this project!

Following are the last - minor - things I want to report.

Table 1 row 5. Consider indicating in the notes that “status quo” does not have a counterpart. It is explained only later in the manuscript.

Notes 2 on page 15. “A two-stage approach (Yuan & Lu, 2008) is preferable in for the Ideology 2.0…” Please, remove “in.”

Notes 3 on page 19. “the “SSpower()” function provided by the “semTools” package in R (CITE).” Insert proper citation.

The power analysis described on page 19 is not a sensitivity (what is the minimum detectable effect size given N, alpha, and power) but a post hoc analysis (what is the power to detect a specific effect size, with a given alpha and N). However, I suggest you perform a sensitivity analysis, as it could be more informative (Perugini et al., 2018 https://doi.org/10.5334/irsp.181) 

Page 19, last row. “Miminum” is misspelled.

The authors have addressed my comments in a convincing manner. I now find it much easier to follow the rationale and planned analyses. I also appreciate that the authors included a strategy for type 1 error control and a novel approach for evaluating model fit. I am looking forward to reading the Stage 2 manuscript.

To the Authors,

In the registered report entitled "Implicit Ideologies: Do Right-Wing Authoritarianism and Social Dominance Orientation Predict Implicit Attitudes?”, you describe a study that aims to investigate whether and how Right-Wing Authoritarianism and Social Dominance Orientation affect explicit and, crucially, implicit preferences for authority, tradition, and hierarchies.

Overall, I believe you aim to address an interesting and relevant research question. However, I also think there are improvement areas in both the introduction and methodology sections.

Theoretical Introduction

I like that you discussed implicit preferences by considering their limitations and critical issues. This is of extreme importance. However, I was left with the impression that it does not emerge why to care about implicit associations. In the paragraph “Implicit vs. Explicit Attitudes” (pp. 4-5), it would also be important to note why it is relevant to investigate implicit attitudes in addition to explicit ones. E.g., implicit measures are less susceptible to demand characteristics and can help measure responses that may be introspectively inaccessible. Further, even though implicit and explicit measures appear to assess the same underlying attitude in some research, they can still tap into more spontaneous evaluative reactions and more controlled responses. In other words, although the attitude might be the same, its expressions might differ.

In the next section (pp. 5-6), you stated, "Given the findings relating to the IAT and political orientation, we might expect RWA and SDO to predict implicit attitudes in a similar way to that of explicit attitudes. However, research employing RWA and SDO to predict implicit attitudes is relatively scarce and offers mixed findings.” The mixed results might be due to the heterogeneity of the attitudes examined. Thus, I do not feel it is a compelling reason to explore implicit attitudes. Perhaps it might be more effective to stress instead that much research has shown that RWA and SDO are significantly linked to various attitudes. If SDO and RWA are predictive of those attitudes, you can expect both to be related to the implicit attitudes of interest, which are closer and more relevant to the considered ideologies (which you noted in the abstract but not elsewhere in the RR).

Another major issue I see in the RR’s introduction is that it lacks a clear description of the research question and hypotheses. Indeed, from the “Present Study” section, it is hard to understand what the dependent variables are and how RWA and SDO are expected to relate to each of them (both in terms of direction and magnitude of the effect). For example, as you provided examples of IAT pairings that should be more relevant to RWA than SDO and vice versa, I find it puzzling that you did not speculate on differences in effects direction and size for each of them.

Related to the dependent variables, I learned later in the RR that you plan to analyze 13 concepts (implicitly and explicitly). However, it is unclear to me what criteria you followed to select them, as for some pairings, it might not be evident why they are relevant (e.g., 1950/2050 IAT pairings). Further, what pairings are relevant to either SDO or RWA and which are relevant to both?

Methodological Section

Although data collection and measurement details are described elsewhere, I believe this section should stand on its feet. For example, it should be clear what procedure was used to collect data (e.g., order of the measures, counterbalancing), data exclusion criteria (e.g., how to treat outliers), how variables will be computed (e.g., I guess explicit attitude will be the difference between the preference for x and y), and the sampling plan (which is briefly described in the Table under Study Design section) should be presented in more details. Further, it might also be helpful to understand the minimum effect size detectable given the available sample size (sensitivity power analysis).

Further, I believe the analysis plan should be enriched with relevant information, such as how the variables will be treated in the SEM model and the fit indices you will use to evaluate the goodness of fit. Further, do you plan to check whether the missing mechanism is as expected by design (in planned missingness design, missing data points are MCAR by definition; thus, you might use Little’s MCAR test to check this on your data)?

As I learned from the OSF documentation, data were collected from different countries (e.g., WEIRD and non-WEIRD countries). Thus, how do you plan to control the hierarchical structure of the data and measurement invariance, especially on SDO and RWA measures (there might be inconsistencies in the measurement models across cultures)? 

In conclusion, I know you are addressing a relevant research question, and I see how important the results would be for understanding how ideologies relate to different attitudes measured at implicit and explicit levels. It is disappointing that all this potential has yet to emerge from the RR. I encourage you to provide a more detailed description of your objectives, to stress further the relevance of using implicit measures in addition to explicit measures, and to add the methodological details that would permit everyone to replicate your study. Good luck with this research!

Minor points

p. 5, row 3: Please, add a reference to IAT.

p. 5, row 7: Remove “toward” from the text.

p. 5, row 5-7: From the description, it should emerge that the implicit association measured with the IAT is relative in nature.

p. 6, row 14: “Emphasize” needs the s.

Sometimes words are spelled in British, and some others in American.

1. While the Author idea is a great one. the author did not define or state clearly in the abstract what Right-Wing Authoritarianism looks like and how they are formed. This should be tested as to how people pick up ideology and how they influence the society.
2. Social Dominance Orientation: How do they behave in the current age, and historically should be added to the study design.
3. sample size is also small to correctly perform this study. I believe a sample size of over 1000 would be much more efficient than what the author had chosen.
4. What predictive model will be developed from this study? this will help other researchers reuse the model developed by this study.

If these questions are address it will benefit and strengthen this study greatly.