Explaining Disparities in Paper Impact across Similar-Level Journals: An Analysis of Sociology Publications

Scientific impact has long been a focal point in the field of science studies,^[1] generating extensive research on its measurement and determinants.^[2–4] In the field of science studies, impact frequently operationalized through citations, represents the extent to which a paper influences human intellect or cognition^[3] and signifies the dissemination and expansion of knowledge.^[5] Given its close relationship with awards, honors, and career progression for scientists,^[6,7] journal rankings and reputation for institutions,^[7–9] resource allocation for policymakers,^[10] and national scientific prestige,^[11] the influence of impact transcends individual papers.^[12]

Despite the abundance of literature exploring citation-based impact, an essential yet inadequately addressed problem remains: why do papers under similar prestige (e.g., published within similar-level journals) often exhibit significant disparities in their citation counts? Prior research has predominantly examined external institutional factors such as journal reputation, ranking, disciplinary paradigm development etc., are primarily tied to the visibility and social recognition of a paper, highlighting a consensus mechanism that enhances visibility and impact through established reputations.^[13–16] However, these external factors alone are insufficient for explaining citation variations among papers under equivalent reputation and ranking. Furthermore, these external indicators typically become accessible only after a paper’s publication, limiting their usefulness in predicting paper impact beforehand.

Previous studies indeed have suggested internal indicators, such as paper type, length, reference structure, novelty, immediacy, recency etc., relate to the content characteristics of the paper itself, as influential predictors of scientific impact.^[17–21] However, a common limitation across these studies is their overly broad comparative scope, often confounding paper quality and social reputation mechanisms. It remains unclear whether internal characteristics genuinely predict impact or simply reflect structural advantages rooted in journal reputation, disciplinary visibility, or author status. In other words, when external “consensus-based” mechanisms dominate, the explanatory power of internal features may be overstated.

This paper argues that the variations in impact among papers of similar quality cannot be attributed solely to randomness, in contrast to the social interaction mechanisms proposed in classic sociological studies, such as those examined in music lab experiments, nor to the cumulative advantage of scientific impact driven by prestige-based or consensus-based mechanisms. Instead, it aims to provide an endogenous explanation by analyzing impact variations among papers of similar quality based on their internal characteristics. Using journal articles from the sociology field with comparable ranking levels as a case study, the research extracts intrinsic knowledge features of these papers and specifically examines the effect on citations. This study adopts a cognitive perspective, aiming to explain citation disparities within journals of equal prestige by focusing on how authors make micro-level strategic decisions in assembling and structuring knowledge. Accordingly, we select three internal mechanisms-context novelty, content novelty, and knowledge focus-which capture how authors cognitively structure their work: how disparate knowledge is connected, how thematic content is recombined, and how narrowly the intellectual content is focused. The findings reveal that while novelty does not effectively differentiate citation counts among such papers, the degree of knowledge focus better accounts for the observed impact variations.

Which papers are more likely to be cited? A common assumption is that high-quality papers receive more citations.^[22] However, there is no formulaic approach to definitively assess the quality of scientific research.^[23] Due to its relative objectivity, citation count has become a widely used metric for evaluating paper quality.^[3,24,25] Highly cited papers are often associated with greater creativity and significance,^[26–28] and citation advantages tend to accumulate over time, a phenomenon widely known as the Matthew Effect in scientometrics.^[5,29,25] This effect is largely driven by reputation-based mechanisms, where heightened visibility enhances the probability of further citations.^[15]

Existing studies has predominantly examined external institutional factors, highlighting the role of reputation in shaping citation impact. Academic evaluation systems such as impact factors, the H-index, and CiteScore institutionalize reputation-based assessments. In reputation-based system, journal reputation, in particular, has been found to be a strong predictor of citation counts.^[30] Papers published in high-impact journals are more likely to attract citations due to increased visibility,^[17–19] and mainstream journals generally exhibit higher citation rates than their less prestigious counterparts.^[13] Author reputation also plays a crucial role, as papers by well-established scholars are more frequently cited.^{[13,17,18,31,32]} Disciplinary differences further influence citation patterns-fields with more established paradigms tend to attract higher citations, as seen in the greater visibility of natural sciences compared to social sciences.^[15] Research fields with highly developed paradigms not only receive more resource support but also benefit from faster peer review and shorter publication cycles.^[16,33]

Despite the predictive power of reputation-based metrics, these systems primarily serve as post-publication evaluation tools, meaning that citation potential can only be assessed retrospectively, offering little predictive value before publication. Furthermore, citation counts are not perfect indicators of quality but rather reflections of perceived importance, constrained by institutional and social mechanisms.^[24] Given these limitations, the challenge remains: how can the impact of a paper be predicted before publication, independent of external reputation signals?

Some scholars argue that citation potential is embedded in a paper’s intrinsic attributes rather than external reputation.^[17] Factors such as paper length, document type, and reference patterns have been shown to influence citation counts. For instance, longer papers generally attract more citations,^[17–19] and review articles tend to be cited more frequently than empirical studies.^[18] Structural components, including Figures and Tables, also contribute to a paper’s visibility and impact.^[20,21] Additionally, citation patterns play a crucial role-papers that incorporate more recent and authoritative references often achieve higher citation counts.^{[17–19,34,49]}

Beyond the formal features above, the role of innovation in citation impact has drawn increasing attention. Creativity is often linked to paper quality, and highly novel papers are more likely to achieve substantial impact.^[35] Two dominant paradigms of innovation-combinatorial innovation and disruptive innovation-have been proposed to explain variation in scientific influence. Disruptive innovation is typically assessed based on a paper’s long-term effect on existing knowledge structures.^[36] Specifically, the disruptive index captures whether subsequent literature continues to cite a focal paper while disregarding its predecessors, making it a fundamentally retrospective metric. As such, it relies on post-publication citation dynamics and is less suitable for studies aiming to predict scientific impact based on pre-publication internal characteristics.

In contrast, combinatorial innovation focuses on the production process itself, arguing that certain knowledge combinations are more innovative than others.^[37] Uzzi introduced combinatorial innovation to scientometrics, developing the widely adopted concept of combinatorial novelty by analyzing journal citation patterns.^[38] Their findings suggest that impactful scientific contributions often emerge from combinations of highly traditional knowledge with novel, unexpected elements. This combinatorial approach has been found to increase the probability of a paper becoming a high-impact publication,^[39] a conclusion supported by subsequent research.^[37]

However, novelty alone may not fully capture the internal cognitive structure of a paper. While novelty emphasizes the distance or atypicality of knowledge combinations, it does not account for the depth or coherence with which ideas are developed and integrated. This aspect of cognitive structure-depth-is closely related to the notion of specialization, which reflects the extent to which research is grounded in a focused and consistent knowledge domain. A growing body of research has examined specialization as a key dimension of scholarly careers, primarily at the disciplinary or author level. These studies explore how scholar’s alignment with specific fields or subfields shapes their professional development. Greater specialization is often associated with higher levels of expertise, clearer thematic positioning, and more stable career trajectories.^[40–45] Furthermore, focus-as a strategy of specialization-suggests a higher likelihood of success. For example, more focused knowledge is more likely to stand out in the field,^[46] promote research productivity,^[40] and even lead to groundbreaking or disruptive outcomes.^[47] These findings suggest that focus is not merely a stylistic feature of scholarship, but potentially a cognitive mechanism shaping scientific recognition. However, these studies aimed at the level of individual researchers, few have operationalized at paper-level knowledge focus as a measurable construct or tested its relationship with scholarly impact. Yet it still remains unclear how such cognitive strategies manifest at the level of individual publications, especially among papers that are published under similar institutional and disciplinary conditions.

In summary, prior research has identified a wide range of factors influencing scientific impact, which can be broadly categorized into external characteristics and internal characteristics. While external factors are well-established predictors of citation performance, they primarily operate through reputation-based or consensus-driven mechanisms, and as such, can only be observed after a paper is published. Studies of pre-publication internal factors although works, however, they often confounding paper quality and social reputation mechanisms due to their overly broad comparative scope. As a result, existing research has not adequately distinguished how specific internal characteristics contribute independently to variations in citation impact within closely matched journal contexts, highlighting the need to explore those factors that can effectively explain impact differences within micro-contexts.

To address this gap, the study focuses on articles published in three leading journals of comparable reputation in a single field of sociology-American Journal of Sociology, American Sociological Review, and Social Forces. Anchored in a cognitive framework, the study examines three internal characteristics: context novelty, content novelty, and knowledge focus, which capture how authors cognitively structure their work, and investigates whether and how these cognitive-level features can predict differential citation outcomes under conditions of structural equivalence. Our findings indicate that neither form of novelty effectively differentiates citation impact among papers published in journals of comparable ranking. In contrast, knowledge focus proves to be an effective predictor of citation variation, suggesting that even when quality differences are difficult to discern, focus can meaningfully shape scholarly recognition.

As identified in the literature review, the empirical question guiding this study is: Among papers published in journals of similar disciplinary standing, to what extent do internal characteristics-specifically content novelty, context novelty, and knowledge focus-predict differences in citation impact? To address this question, we detail below the data sources and procedures used in this study, including data collection and preprocessing, variable measurement, and the analytical strategy.

As previously discussed, this study aims to explain variations in impact among papers of comparable quality. However, existing research based on reputation-driven or consensus-based external characteristics-which are only observable through post-publication evaluations-as well as studies relying on pre-publication external indicators-which often conflate paper quality with social reputation mechanisms-are limited in their ability to address this question. To overcome these limitations, we deliberately control for disciplinary context and journal prestige in our data selection. Specifically, we focus on three journals within a single discipline-sociology-namely, American Journal of Sociology, American Sociological Review, and Social Forces-that are of similar ranking or prestige level, as a representative example. This design allows us to examine differences in scholarly impact and the underlying mechanisms under conditions where social structural factors are held constant.

The data used in this study were primarily drawn from the Web of Science (WoS) and the Microsoft Academic Graph (MAG), two widely recognized databases of scholarly publications. We retrieved all research articles information published between 1999 and 2022 ¹by conducting journal-specific searches in WoS. We extracted metadata such as year, titles, abstracts, keywords, and DOIs from WoS. We then manually search and download the content texts of these papers to facilitate text analysis using open-access sources and institutional subscriptions where available based on information of titles and DOIs. In total, we obtained 3,190 papers: 752 from American Journal of Sociology, 931 from American Sociological Review, and 1,507 from Social Forces (see Table 1). In addition, we also collected the historical impact factor data of the three journals from the WoS platform. Citation data were obtained by matching each paper’s DOI with the Microsoft Academic Graph (MAG), which contains comprehensive citation records across disciplines.

Before analysis, the content texts of each paper were batch-converted to editable .txt files, from which we parsed structural components such as titles, abstracts, main body texts, and footnotes. For the purpose of capturing content-level features, we focused our analysis on the main body of each article, excluding appendices and references. Figure 1 presents an overview of the data collection and preprocessing procedure employed in this study.

Figure 1:
Data collection and preprocessing workflow.

This study aims to investigate how internal characteristics of scholarly papers contribute to variations in citation impact-especially under conditions where external structural factors such as journal prestige or disciplinary scope are relatively comparable. By focusing on papers published in journals of similar rank and field, we can examine whether internal features can still meaningfully account for differences in scholarly visibility once broader institutional or reputational influences are held constant. Prior studies have identified a range of internal features that tend to correlate positively with citation counts. However, much of this evidence comes from heterogeneous samples across fields and journal tiers, where internal features are often confounded with paper quality or reputation-based mechanisms. As a result, it remains unclear whether such internal attributes retain their predictive value in more structurally uniform contexts.

To address this question, we test the extent to which three specific internal characteristics can distinguish impact among papers published in journals of comparable standing. While this selection inevitably excludes other potentially relevant features, we believe that focusing on context novelty, content novelty, and knowledge focus is beneficial, as these three factors capture deeper cognitive dimensions of scholarly work. That is, they speak directly to how authors organize, integrate, and position knowledge within the intellectual space: how disparate ideas are connected, how thematic content is recombined, and how narrowly the argument is focused. Specifically, context novelty measures the uniqueness of a paper’s knowledge base, based on how typical its cited journal combinations are; content novelty captures the rarity of thematic combinations, derived from topic modeling; and focus assesses the degree of concentration in a paper’s topic distribution. The measurement details are outlined below.

Context novelty captures the degree to which a paper draws on atypical combinations of prior sources-specifically, the rarity of journal pairings in its reference list. This measure reflects how distinctive or unconventional a paper’s knowledge base is in relation to prevailing citation patterns. Specifically, we use commonness as a proxy variable for context novelty. Following previous study,^[37] we define the commonness of each journal pair (journals i and j) in year t as follows:

We identify all journal pairs cited by a given paper and calculate their commonness based on how frequently each pair co-occurs in other papers published in the same year. We define the reference universe as all papers in the sample published in year t, and calculate the expected co-occurrence of each journal pair (i, j) based on their marginal frequencies. The observed-over-expected ratio indicates how typical or atypical the journal pairing is. For each paper, we assign commonness scores to all cited journal pairs, then take the 10^th percentile of these values as a conservative estimate of contextual distinctiveness. This value is log-transformed and sign-reversed so that higher scores reflect greater context novelty. Intuitively, a paper citing highly uncommon journal combinations is considered to possess a more original knowledge foundation.

Content novelty captures the originality of a paper’s thematic structure by assessing the rarity of its core topic combinations relative to recent literature, which distinguishes content-level novelty from citation-based indicators by focusing on the internal composition of ideas. The underlying premise is that papers recombining commonly separated topics may introduce novel conceptual linkage or problem framing, thereby increasing their visibility and impact. Content novelty is defined as the rarity or uniqueness of the combination of two representative topics:

Specifically, using paper text from the three target journals, we trained an LDA topic model and identified the two most prominent topics for each paper based on topic loadings, we then computed the co-occurrence frequency of selected topic pairs within a three-year window preceding each paper’s publication. The rarity of a topic combination was calculated by comparing its observed frequency to the expected frequency under an assumption of independence. Lower observed-over-expected ratios indicate more novel thematic pairings.

Knowledge focus captures the degree to which a paper concentrates its intellectual content within a narrow thematic scope, as opposed to distributing its content broadly across multiple thematic areas. The underlying idea is that a more focused paper may offer greater conceptual clarity, stronger domain specificity, and thus more easily identifiable scholarly contributions.^[40,42]

We operationalize focus using the Herfindahl-Hirschman Index (HHI), a widely used measure of concentration originally developed in economics, as:

In our context, it reflects how unevenly a paper distributes its content across different topics. For each paper, we take the posterior topic distribution produced by the LDA model and compute the sum of the squared topic loadings. This yields a score ranging from 0 to 1, where values closer to 1 indicate a higher concentration of content in fewer topics (i.e., greater knowledge focus), and values closer to 0 indicate a more evenly distributed topic profile (i.e., less focus).

To examine the relationship between internal characteristics and scholarly impact, we employ an Ordinary Least Squares (OLS) regression model, with the number of citations a paper receives within five years of publication as the dependent variable. This five-year citation window is commonly used in bibliometric research^[2] to capture a paper’s medium-term influence, and we also tried other window (like 10 years). The three focal independent variables-content novelty, context novelty, and knowledge focus-represent distinct dimensions of cognitive structuring. To account for temporal variation in citation practices and exposure time, we include publication year fixed effects as control variables. We also include team size as a control variable to partially account for author-level effects such as visibility or collaboration-based advantages, which may be correlated with reputation or institutional standing. In addition, we also conducted some robustness checks.

As detailed in the Data Collection and Preprocessing section, journal impact factor data for the three selected journals were retrieved from the Web of Science database. Figure 2 illustrates the changes in impact factors for the three journals from 1999 to 2022. While all three journals exhibit an overall upward trajectory, notable differences emerge over time. The American Sociological Review (ASR) and Social Forces show more substantial growth-particularly after 2015-whereas the American Journal of Sociology (AJS) displays a slower and more gradual increase. Before 2013, AJS and ASR had comparable impact factors, both consistently higher than Social Forces. However, in subsequent years, ASR pulled ahead, while Social Forces gradually closed the gap with AJS, even surpassing it in recent years.

Figure 2:
Changes in impact factors.

Although all three journals are considered top-tier outlets in sociology, this divergence raises questions about whether internal publication characteristics-beyond journal reputation-may account for differential citation dynamics. In the following sections, we explore this possibility by examining the internal cognitive features of individual papers across these journals.

Before analyzing the topical composition of the journals, we evaluated the coherence scores of LDA models with varying numbers of topics (ranging from 5 to 40). As shown in Figure 3, coherence peaks at 25 topics (0.629), but remains relatively stable. While 25 topics offered the highest quantitative score, we also considered semantic interpretability through manual inspection of topic-word distributions. A 20-topic solution was selected as it provided a balance between thematic clarity and model coherence, facilitating both analytical tractability and conceptual relevance.

Figure 3:
Number of topics and topic coherence score.

Table 2 presents the 20 extracted topics along with their most representative vocabulary terms (only displaying the topics with clear significance). These topics reflect major thematic domains within sociology, including law and crime, education, social movements, religion, gender, and more.

Figure 4 illustrates the temporal distribution of topics across the three journals. ASR and Social Forces exhibit relatively balanced topic coverage, with no sustained dominance of particular themes. In contrast, AJS demonstrates a stronger and more persistent emphasis on certain topics-especially social movements and, during 2011-2017, violence and terrorism. This thematic concentration may partially contribute to the journal’s slower growth in impact factor over the years. However, this observation remains tentative, as the relationship between topical focus and journal-level citation outcomes is complex and likely influenced by multiple factors. It is also important to distinguish journal-level topic patterns from paper-level knowledge focus, which is the main object of analysis in subsequent sections.

Figure 4:
Topic distribution over years.

Importantly, this journal-level pattern does not contradict our subsequent findings at the individual-paper level, where we observe that greater knowledge focus-the degree to which a paper concentrates its thematic content-can enhance citation impact among papers of comparable journal ranking. In other words, while excessive topical repetition may limit a journal’s reach, cognitive focus at the article level may help a paper stand out in a crowded intellectual space.

Figure 5 shows the variations in content novelty and context novelty for the three journals. Regarding content novelty, there are significant differences between the three journals. The content novelty of AJS is lower than the other two journals for most of the time, especially lower than ASR. The content novelty of AJS and Social Forces fluctuates more widely, while ASR’s content novelty changes relatively steadily. In addition, the content novelty of AJS and Social Forces shows considerable fluctuation, while ASR remains relatively stable with a slight increase throughout the observed period. For context novelty, AJS is higher than the other two journals in most cases, particularly higher than Social Forces. At the same time, AJS’s context novelty fluctuates significantly more than the other two journals. AJS and ASR show significant peaks in context novelty in some years, while Social Forces remains relatively stable.

Figure 5:
Changes of novelty.

Furthermore, consistent with previous research,^[48] the correlation between the two types of novelty we measured is weak (coefficients<0.03), supporting the idea that they capture distinct dimensions of innovation. These indicators will be analyzed further in the regression models to assess their predictive value for citation impact.

Figure 6 displays the temporal patterns of knowledge focus across the three journals. ASR maintains the most stable and consistently high level of focus, suggesting that its published papers are more thematically concentrated. In contrast, AJS exhibits a gradual decline in focus-especially between 2007 and 2013-and remains lower overall than the other two journals. Social Forces shows the greatest volatility, with substantial year-to-year fluctuations across the observation period. Whether such journal-level patterns translate into citation differences at the paper level is addressed in the regression analysis that follows.

Figure 6:
Changes of knowledge focus.

To examine how internal characteristics of papers shape their scholarly impact, we employ an Ordinary Least Squares (OLS) regression model. Following previous work,^[18,36,49] we model five-year citation counts as a function of the three main internal characteristics: context novelty, content novelty, and focus.

Before conducting the regression analysis, we performed multicollinearity diagnostics for the main explanatory variables. we conducted collinearity diagnostics using two methods. First, a Pearson correlation matrix (see Table 3) shows that all pairwise correlations among the three main predictors are below 0.03, indicating extremely weak linear relationships. Second, we computed Variance Inflation Factors (VIF), all of which are approximately 1.0-well below conventional thresholds for multicollinearity concerns (see Table 4). These results confirm that the three predictors capture conceptually distinct dimensions of internal structure.

Table 5 presents the descriptive statistics for both the independent and dependent variables. To construct the analytical sample, we first matched the records from the three sociology journals with citation and reference data from the Microsoft Academic Graph (MAG), yielding an initial sample of 1,117 papers. Since the dependent variable is citation count within five years of publication, we excluded papers published after 2015 to ensure sufficient time for citation accumulation. In addition, content novelty is calculated based on the rarity of a paper’s top two topic combinations compared to those in the previous three years. Given that our dataset starts in 1999, we could only compute content novelty for papers published in 2002 or later. As a result, papers published before 2002 were excluded from the final model. After applying these restrictions, the final sample size used for regression analysis is 967 observations.

Table 6 presents the OLS model results for the individual characteristics of the papers and their citations. In this model, as mentioned earlier, we use the citation count within five years after publication as the dependent variable (we also tried 10-year window), with content novelty, context novelty, focus as the main explanatory variables, and control for year. We also include team size as a control variable to partially account for author-level effects, which may be correlated with reputation or institutional standing. The detailed results of the main model are shown in Table 5.

From Table 6, we observe that neither content novelty nor context novelty has a statistically significant effect on five-year and ten-year citation counts after publication. This contrasts with findings from previous studies that have associated higher novelty with greater impact and visibility.^[38,48,50] One possible reason for this divergence is that, when controlling for structural comparability across similarly ranked journals, the marginal advantage of novelty may be diminished.

In contrast, knowledge focus shows a strong and statistically significant positive association with citation impact (p<0.01). Papers that exhibit higher thematic concentration tend to receive more citations, suggesting that cognitive coherence may enhance visibility even in settings where external prestige signals are held constant. While this result underscores the potential value of focus as an internal quality indicator, further research is needed to explore how it interacts with other mechanisms such as author reputation or network position.

To assess the robustness of our findings, we re-estimated the models using alternative topic model specifications (15 and 30 topics). As shown in Table 7, knowledge focus continues to exhibit a statistically significant positive relationship with citation counts, regardless of topic number. In contrast, the effects of content and context novelty remain largely non-significant, even with only a marginal negative coefficient for content novelty. As a further sensitivity analysis, we also extended the citation window from five to ten years. This longer-term measure helps assess whether the observed patterns persist beyond initial post-publication dynamics. The results remain largely consistent: knowledge focus retains a significant positive effect, while novelty measures again show limited predictive value. Overall, these analyses reinforce the robustness of our main findings. The positive association between thematic concentration and citation impact holds across topic models and time windows, suggesting that focus operates as a stable internal characteristic may associated with greater scholarly visibility in similarly ranked journals.

In summary, our results suggest that the distinctiveness of a paper’s topical focus is positively associated with citation impact-i.e., papers with more concentrated knowledge tend to receive more citations. In contrast, neither context novelty nor content novelty shows a statistically significant effect in any of the model specifications, despite their theoretical prominence in prior literature. The summarized results of the models are provided in Table 8. These findings do not negate the broader value of novelty-based approaches, but indicate that within the controlled context of journals of similar ranking, knowledge focus emerges as a more robust internal predictor of citation differences. Further research could explore whether the effects of novelty operate differently across fields or in less structurally comparable publication settings.

While our analysis concentrates primarily on the internal cognitive features of academic papers-such as their conceptual focus and thematic novelty-we fully acknowledge that citation impact is also significantly influenced by a range of external social and institutional factors. These may include, but are not limited to, prevailing disciplinary citation norms, the professional reputation of the authors, the prestige and visibility of their affiliated institutions, and patterns of self-citation. Such factors often operate in subtle and overlapping ways, shaping how scholarly work is received and recognized within academic communities.

To partially control for structural variation across different publication venues, we deliberately limited our sample to articles published in three highly prestigious sociology journals. This sampling strategy serves to mitigate some sources of heterogeneity, especially those related to journal prestige or disciplinary subfields. However, this focus also introduces limitations with respect to the broader applicability of our findings. Specifically, it constrains the generalizability of our conclusions beyond the specific field of sociology, and may not reflect dynamics present in other disciplines or less prominent journals.

Additionally, we rely on topic modeling using Latent Dirichlet allocation (LDA), a widely used but inherently limited computational technique. LDA’s outcomes are sensitive to the selection of parameters (such as the number of topics), and the interpretive process involved in labeling topics introduces a degree of subjective judgment. As a result, the model may not fully capture the nuanced and layered meanings embedded in complex scholarly texts.

Prior research on scientific impact has emphasized both external institutional mechanisms-such as journal prestige, author reputation-and internal characteristics of scholarly work, including length, citation density, and novelty. While these factors have offered valuable insights, a persistent question remains: among papers published in similarly prestigious venues, what explains the variation in scholarly impact? In such cases, where structural advantages are relatively controlled, internal cognitive characteristics may play a more consequential role. This study contributes to that line of inquiry by examining how papers’ novelty and knowledge focus relate to citation outcomes within three similar-level sociology journals. More broadly, our study underscores the importance of looking beyond traditional prestige-based or network-driven explanations of scientific recognition. Especially in contexts where external status signals are already saturated, internal cognitive features-such as how ideas are assembled, framed, and integrated-may matter more than previously assumed. This perspective also invites renewed attention to micro-level variation: small differences in how papers articulate knowledge may produce outsized effects in how they are received and cited, even among works of ostensibly similar quality.

Our analysis shows that, under conditions of structural comparability, knowledge focus-as measured by thematic concentration-is a stable and consistent predictor of citation impact, whereas both content and context novelty lose their statistical significance. This finding suggests that, even when author reputation and journal visibility are held relatively constant, how narrowly and coherently a paper organizes its intellectual contribution may enhance its chances of being cited. Focus may improve a paper’s conceptual legibility and recognizability within its field, facilitating easier uptake by other researchers. In contrast, the benefits of novelty-particularly when not accompanied by sufficient cognitive anchoring-may be more limited or delayed.

These findings partially align with, but also diverge from, prior research. For example, previous studies have emphasized the positive association between novelty-particularly atypical combinations-and citation impact.^[38,39] However, their analyses typically draw from broader, cross-disciplinary datasets with high variation in journal prestige and author visibility. In contrast, our study holds these structural factors relatively constant, suggesting that the predictive value of novelty may be contingent on variation in social prestige or disciplinary breadth. Previous studies also have shown that the relationship between novelty and impact may follow an inverted-U pattern: moderately novel papers are more likely to be cited, while extremely novel ones may be too disconnected to be immediately recognized.^[51–53] Furthermore, highly novel work often carries reputational and interpretive risks-it may be overlooked, misunderstood, or underappreciated, especially in its early stages. While our study does not directly test these mediating mechanisms, our findings are consistent with the idea that in structurally saturated environments, the payoff of novelty may diminish or become more unpredictable. That said, our results do not challenge the theoretical importance of novelty, but instead point to its conditional effects.

On the other hand, our finding that knowledge focus predicts impact resonates with emerging work on thematic coherence and research specialization. For instance, some studies demonstrate that scholars who pursue more focused intellectual trajectories tend to receive greater recognition.^[41,42] While most of this literature has focused on the author level, our findings extend this insight to the paper level, showing that coherence within a single publication may matter as much as career-long focus.

Our findings also contribute to understanding the relative changes in journal impact factors. We observed that among the three journals examined, the American Journal of Sociology (AJS) has experienced a slower increase in its impact factor over the past two decades. In parallel, our topic modeling shows that AJS has consistently concentrated on a narrower set of themes-particularly a sustained emphasis on topics related to social movements-compared to ASR and Social Forces. While this overlap in trends is noteworthy, we do not claim a causal relationship between thematic concentration and journal-level impact. These observations may raise a potential hypothesis: that reduced topical diversity may limit a journal’s reach across subfields or audiences. However, this proposition remains speculative and would require dedicated empirical testing in future research.

This study set out to investigate the extent to which internal cognitive characteristics of scientific papers-namely, context novelty, content novelty, and knowledge focus-affect their subsequent citation impact when published in sociology journals of comparable ranking. While the role of novelty has long been emphasized in the literature as a central driver of scientific influence and recognition, our empirical findings challenge this assumption by showing that its predictive strength weakens in settings where structural factors, such as journal prestige and editorial reputation, are held relatively constant. In other words, novelty alone does not necessarily translate into higher impact when external evaluation cues are uniform across publication venues.

In contrast, knowledge focus-operationalized as the degree of thematic concentration within a paper-consistently and significantly predicts citation performance. This suggests that when the structural playing field is level, the internal organization of ideas, particularly the coherence and specificity with which a research contribution is articulated, becomes a more decisive factor in gaining scholarly attention. These findings align with and extend a growing body of scholarship that underscores the importance of cognitive structure in shaping how scientific work is received and valued by the academic community. They also contribute to an emerging understanding that scholarly visibility may depend as much on how ideas are framed and presented as on the novelty of the ideas themselves.

Several limitations of this study warrant careful consideration. First, while we aimed to control for journal-level prestige, we could not account for other potentially influential factors such as institutional affiliation, network embeddedness, or prior visibility. Second, although we focused on three high-ranking journals in sociology, the findings may not generalize to fields with different citation practices or epistemic cultures, and thus restrict the broader applicability of our conclusions. Third, our analysis emphasized novelty and focus as internal features, but other cognitive dimensions-such as clarity, methodological rigor, or rhetorical structure-may also shape impact. Future research should expand the scope of internal characteristics and explore how they interact with social mechanisms to influence scholarly visibility. However, ultimately, this study advances our understanding of how impact disparities emerge among papers of comparable quality, and underscores the importance of internal coherence-beyond novelty alone-as a mechanism of scholarly visibility.