Lexical density in English newspapers – a cross-analysis of the New York Times and Arab News Open Access

This study explores three research questions:

This research is the first attempt to compare the lexical density and the overall readability of newspapers representing two different cultures, The New York Times and Arab News. Therefore, it reveals patterns in cross-cultural journalism that can guide organizations in adjusting their writing patterns for increased engagement of audiences in the global segment. Moreover, the results can help news entities in the development of effective strategies to enhance the readability and comprehensibility of articles without sacrificing their information-carrying capacity.

Lexical density: The proportion of content words relative to the total word count in a text or corpus (Ure, 1971).

Content words: Nouns, verbs, adjectives, and adverbs that carry meaning (Ure, 1971).

Readability: The ease with which an individual can read and understand a material (Kembaren and Aswani, 2022).

Comprehensibility: The degree to which a text can be understood by its intended audience (Kembaren and Aswani, 2022).

Lexical items: Individual words that have semantic content and contribute to the overall lexical density in a text (Khorina and Handani, 2022).

Lexical density indicates the information-carrying capacity in written and oral discourse. Lexical density is measured as the proportion of content words, which primarily encompass adverbs, verbs, nouns, and adjectives, in the corpus or overall text (Ure, 1971). In other approaches for computing lexical density, lexemes are expressed as a ratio of a ranking clause (words that convey a complete idea) (Halliday, 1985). Most studies rely on Ure’s (1971) method in which the functional perspective of language is de-emphasized, and multiple words are not integrated into a single unit (Aziz and Riaz, 2024). Ure’s (1971) method is more straightforward and applies primarily to linguistic appraisal.

Several studies have examined lexical density in spoken and print media outlets. Examination of TNYT, Business of Fashion, and BBC News using Halliday’s (1985) and Ure’s (1971) approaches led to lexical density scores above 50% (Kembaren and Aswani, 2022; Khorina and Handani, 2022; Rahayu and Syaifullah, 2022). Khorina and Handani (2022) used a sample of 10 pieces of written and spoken news from BBC News. The findings indicate that written news had slightly higher average lexical density (55.05%) compared to spoken samples (52.15%). Nevertheless, research has indicated that articles in magazines vary significantly depending on the topic and target audience. In their research, Sari and Ekawati (2021) found that the lexical density for articles in Reader’s Digest Magazine covering various topics was 64.5% for travel, 54.68% for parenting, 58.84% for language, 61.56% for animals, and 61.34% for health. The findings indicate that the majority of the articles could be classified as intricate since their lexical density values were above 60%.

A paucity of research has established a relationship between lexical density, readability, and comprehension in different contexts. Bakuuro (2024) applied both Flesch’s (1948) and Gunning’s (1952) frameworks to examine readability of high school English texts. The authors inferred a negative correlation between lexical density and readability. Both Ure’s (1971) and Halliday’s (1985) approaches provided closer values of linguistic complexity. Liu and Dou (2023) indicate that higher ratios of lexemes are associated with a greater cognitive load during the processing of the information. In another study (Rahayu and Syaifullah, 2022), TNYT articles had a lexical density of 58.1% congruent with a level 12 of grammatical intricacy, compared to the Business of Fashion's lexical index of 54.2% and corresponding grammatical intricacy level of 8. Nevertheless, the above study did not examine the relationship between readability and lexical density. While widely used in the literature, readability formulas have been criticized for focusing on a limited number of textual features to inform reader comprehension (Crossley, 2024). Studies have established additional factors that influence reader comprehension. For example, Ji et al. (2023) demonstrated that information distribution over long documents influence comprehension. In addition, Hackemann et al. (2022) indicated that technical terms and longer sentences with multiple clauses can substantially reduce the comprehensibility of a corpus.

Studies have examined English-language newspapers in ESL countries, revealing important cultural variations (Ghani et al., 2022; Liu and Dou, 2023; Salihoglu and Karatepe, 2023). Using Halliday’s (1985) model of systemic functional linguistics, Ghani et al. (2022) found that the ratios of lexemes in Pakistani and UK news articles were 49.89 and 53.09%, respectively. The findings suggest that cultural adaptations in non-native English-speaking countries often lead to simplified syntactic structures. Liu and Dou (2023) found that translation processes in cross-cultural contexts impact lexical complexity. They found that direct expression of content in English was linked with more complex vocabulary and content words compared to the interpretation of texts from Chinese or Russian to English.

Some studies have examined lexical density indices in Western newspapers such as BBC News and TNYT (Kembaren and Aswani, 2022; Khorina and Handani, 2022; Rahayu and Syaifullah, 2022; Sari and Ekawati, 2021). However, limited amount of research that examined Western and Middle-Eastern English-language newspapers did not consider the relationship between information density and corresponding readability levels (Ghani et al., 2022; Salihoglu and Karatepe, 2023). Therefore, this study fills this gap by comparing the lexical density and readability of articles from different cultural contexts, Western and Middle-Eastern publications. To address these gaps, this study is structured into three objectives:

• (1)

  To examine the difference between lexical density of articles published in TNYT and AN.

• (2)

  To investigate the impact of lexical density on readability and comprehensibility.

• (3)

  To identify factors beyond lexical density that affect readability.

The examination of these objectives will address the gap in the role of lexical and structural choices in cross-cultural news. In addition, the objectives provide a framework for understanding how English-language newspapers can adjust and present their messages in a manner that aligns with the diverse international audiences. The objectives support a study approach that combines quantitative and qualitative approaches, which contributes to a better understanding of textual complexity, as well as other cultural-linguistic aspects influencing the reader's comprehension. The findings are relevant to news agencies aiming to disseminate detailed informational content while ensuring that their articles demonstrate reasonable comprehensibility and readability for both the local and the international audiences.

This study applied a mixed-method approach to examine the lexical density and readability of TNYT and AN newspapers. The quantitative method involves analyzing numerical data to elucidate facts about the phenomenon associated with the population. In addition, it reduces bias associated with subjective interpretation of results inherent to qualitative designs. A quantitative approach is relevant given that the current study involved comparing linguistic measures for the two newspapers. A qualitative content analysis allowed for in-depth analysis of contextual factors affecting text comprehensibility.

The sample in this study involved 100 newspaper articles published between 2022 and 2024 and covered topics related to climate change and environment. Fifty articles were obtained from TNYT and the remaining ones from AN. A purposive sampling was applied to help the researcher select articles that are relevant to the study objectives. The sampling approach can enhance the internal validity of findings by focusing on participants who truly represent the phenomenon under investigation. The method was useful for the collection of specific and meaningful data related to lexical density in media discourse.

Articles were retrieved from the official websites of TNYT and AN. The selection criteria involved English-language articles and a minimum of 300 words. In addition, articles from wire services were excluded from the study.

Ure’s (1971) formula was applied to evaluate the lexical density of words within the selected articles. The content-to-total word ratio indicates the percentage of nouns, verbs, adjectives, and (lexical items) in the text. Prepositions are usually excluded from the computation due to a lack of semantic content. The formula is given below (Khorina and Handani, 2022):

In addition, the Flesch-Kincaid Grade Level, developed by Kincaid et al. (1975), was used to calculate the readability aspect. Readability is defined as the proficiency level required for individuals to comprehend the text; therefore, it was synonymous with comprehensibility (Kembaren and Aswani, 2022). Flesch-Kincaid framework accounts for the sentence length and syllable count to establish the grade level required to comprehend each text. As shown in Table 1, the complexity of the text increases as the score moves from a low value to higher Flesch-Kincaid Grade Level scores (Kincaid et al., 1975). The Automatic Readability Checker, a text analysis software, which analyzes discourses based on Flesch-Kincaid Grade Level, was applied for data entry and processing. Manual verification and adjustment were performed. Data were stored in an Excel spreadsheet for further statistical analysis. Flesch-Kincaid formula for appraisal of readability is as follows (Kincaid et al., 1975):

Descriptive statistics were completed to determine the means and standard deviations for lexical density and Flesch-Kincaid grade-level scores. The impact of lexical density on readability scores was examined through the computation of Pearson's correlation coefficient. In addition, independent t-tests were conducted to examine if the mean lexical density of the two newspapers varied significantly.

The qualitative content analysis involved selecting five articles from each of the two newspapers. A thorough examination of the articles is essential for familiarization and understanding assumptions and patterns in the content (Nicmanis, 2024). Sentences and phrases from the articles were analyzed to determine cultural-linguistic factors that can affect the comprehensibility for general audience. The researcher formed categories and themes that allowed the identification of the similarities and differences between AN's and TNYT's climate change reporting styles.

Several measures were taken to ensure that the findings reflected the linguistic characteristics of each newspaper. Reliability focuses on the consistency of measures, while validity indicates the accuracy of measures. Ure’s (1971) lexical density formula and the Flesch-Kincaid Grade Level (Kincaid et al., 1975) has been proven reliable in previous studies. In addition, two researchers independently measured the variables for a subset of articles. The results indicated a high level of inter-rater reliability. The utilization of linguistic measures from a respected scholar, Ure (1971), and unambiguous operationalization of readability and lexical density enhanced the content and construct validity of the study. Nevertheless, external validity might be limited to newspapers with audiences and styles similar to those employed in this research.

The sample for this study comprised 100 articles from TNYT and AN. Table 2 shows the descriptive statistics of the lexical density for the two newspapers. The summary of the individual lexical density elements for each newspaper is shown in Table 3. The average readability values in the form of Flesch-Kincaid Grade Level for TNYT and AN are depicted in Table 4. In addition, this study involved testing the statistical significance of the difference between TNYT and AN lexical density, which is shown in Table 5. A correlational analysis of the lexical density and Flesch-Kincaid Grade Level scores of all the articles from two newspapers is shown in Table 6.

Arab News extensively relied on local context examples, especially from the Middle East and North Africa (MENA) region. The publication’s staff referenced detailed country-specific carbon capture and renewable initiatives. Articles focused on Saudi Arabia’s projects, MENA as a food deficit economy, and the region’s potential water scarcity. The analysis revealed more global references with less contextual details in TNYT.

Arab News and TNYT demonstrated different strategies for presenting technical content regarding climate change issues. Arab News, for instance, provided explicit definitions or analogies for terms such as voluntary carbon market and circular carbon initiative. On the other hand, TNYT writers mostly assumed that their readers had prior knowledge of environmental discourse and international policy frameworks. The New York Times used unexpanded acronyms and did not include explanatory phrasing for scientific terms such as thermal batteries, carbon capture, and clean hydrogen. However, similar to TNYT, one of the AN articles assumes that the audience has background knowledge about global climate initiatives due to the use of terms such as The Conference of Parties (COP27) in the form of acronyms and without contextual elaboration.

Structural analysis focused on sentence complexity, paragraph organization, and explanatory transitions. The New York Times demonstrated a higher frequency of complex sentences than AN (62 vs 54%). On the other hand, AN articles had shorter paragraphs (average of 3–4 sentences) than TNYT (6–7 sentences). However, both AN and TNYT showed comparable use of explanatory transitions.

Table 2 shows that TNYT has a slightly higher mean lexical density (57.96%) compared to the AN sample (56.54%). The above findings align with Ure’s (1971) study which established that lexical density for written news exceeded 40%. The results shown in Table 3 indicate that the lexical items that accounted for the greatest proportion of the lexical density in both sets of newspapers were nouns, with TNYT having a slightly higher percentage (35.32%) than AN (32.43%). Independent t-test results (p > 0.05) indicate that the difference between lexical density in TNYT and AN is not statistically significant, as shown in Table 5. Therefore, TNYT and AN are comparable in terms of information-carrying capacity.

Prior studies have concentrated on lexical density in Western media outlets such as TNYT and BBC News without comparing the values with Middle-Eastern media (Khorina and Handani, 2022; Rahayu and Syaifullah, 2022). However, consistent research has established that the typical lexical density for news articles and other publications that use formal writing is 50–60% (Maufiroh et al., 2024; Ramadhani et al., 2023). However, some AN and TNYT articles had lexical density scores that exceeded 60%, as depicted in Table 2. According to Maufiroh et al. (2024), values ranging between 61 and 70% are interpreted as high and indicate a text that is difficult to read. Additionally, these above-average lexical density patterns are common in specialized documents assuming that the audience has an in-depth understanding of the subject matter.

A correlational analysis was performed to determine the impact of lexical density on readability. As shown in Table 6, lexical density has a positive correlation (0.6332) with Flesch-Kincaid Grade-Level scores (used to measure the readability and comprehensibility of the text in this research). Higher Flesch-Kincaid Grade-Level scores usually indicate that the text is more difficult to read (Kembaren and Aswani, 2022). Therefore, the correlation analysis indicates that reading comprehension for TNYT and AN articles decreases as lexical density increases. Information-dense texts require a higher cognitive load for readers (Liu and Dou, 2023). Akin to the findings in this study, Bakuuro (2024) found that lexical density had a negative relationship with readability. Nevertheless, the researcher focused on academic context rather than the media. Rahayu and Syaifullah (2022) found that The New York Times articles had higher grammatical intricacy than Business of Fashion due to higher lexical density.

Table 4 shows that the average readability levels of the two media are TNYT (13.45 Grade Level) and AN (12.90 Grade Level). The scores indicate that the articles can be read and understood by individuals with at a least college education. In addition, AN and TNYT articles can be categorized as difficult to read (see Table 1). Given the moderate to strong positive correlation established in Table 6, newspapers should consider lowering lexical density as one of the measures to increase readability and make their content suitable for individuals with an average education level.

The qualitative analysis offers insight into additional factors other than lexical density that can affect the comprehensibility of AN and TNYT articles. Although the two newspapers have similar lexical density, AN’s regional adaptation, technical content explanation, and lower sentence complexity compared to TNYT can enhance reader comprehension for non-native English speakers. Arab News focuses on MENA’s environmental concerns and uses locally relevant examples such as Saudi Arabia’s renewable energy initiatives to elaborate on impacts and practical implementation strategies. The New York Times reports climate change issues from a global perspective and demonstrates higher technical content with minimal contextual explanation, reflecting its assumption of reader familiarity with environmental discourse. Prior research has established that complex sentence structures and technical vocabulary reduce the comprehensibility of texts (Crossley, 2024; Hackemann et al., 2022). Therefore, in can be inferred that TNYT’s linguistic features can hinder reader comprehension more than AN.

This study has established that higher lexical density is associated with lower readability and comprehensibility. Therefore, it is prudent for news organizations to include lexical density in their writing policies. Recommended optimal density values that can allow news articles to balance information-carrying capacity and comprehensibility for the general audience should range from 50% to 60%, as discussed by Maufiroh et al. (2024). The print media should use a multi-tiered approach in which a significant proportion of articles are written with low lexical density for the general public and additional in-depth articles are disseminated for individuals proficient in climate change and environmental policy. Also, based on the qualitative results, journalists should increasingly use contextual explanations for technical terms and integrate simple sentence structures and regional adaptation into their articles to reach individuals with average education levels.

One of the limitations of this research involves the exploration of two English-language newspapers from Western and Middle-Eastern world regions. This means that the findings cannot accurately reflect the audiences in these two regions. Additionally, this study did not involve the comparison of tolerance thresholds for lexical density among different groups, such as native and non-native English language speakers. In the future, researchers should examine the same topic by considering other Western and Middle-Eastern newspapers that were not covered in this study. In addition, interviews with experts and tests should be administered to compare the impacts of lexical density on comprehensibility between native and non-native English language speakers.

This study aimed to compare the lexical density of TNYT and AN articles published between 2022 and 2024 and their impact on the readability of the text. Ure’s (1971) content-to-total word ratio method and the Flesch-Kincaid Grade Level were used to measure lexical density and readability, respectively. An independent t-test, Pearson’s correlation coefficient, and qualitative content analysis were performed to answer research questions. The findings indicated a moderate positive correlation (0.6332) between lexical density and the Flesch-Kincaid Grade Level. Therefore, this means that the readability of the text reduces as lexical density increases. Arab News and TNYT have comparable lexical density averages, which are greater than 50%. The findings underscore the need for news organizations to establish optimal density ranges to disseminate informative content and simultaneously enhance reader comprehension. In addition, local context integration, explanatory phrasing of technical terms, and less complex structures should be considered as additional measures for increasing comprehensibility for audiences. Future research should examine lexical density and comprehensibility of newspapers from multiple cultural contexts.