The interaction effect of online review language style and product type on consumers’ purchase intentions

Online reviews, a form of electronic word-of-mouth, influence consumers' evaluation of products, purchasing decisions, and behaviors. Surveys indicate that over 70% of consumers refer to and trust online comments and about 85% report that reviews influence their purchase behaviors. While prior work has largely examined online review valence, quantity, quality, and similarity, less attention has been paid to the narrative style of reviews, despite text content having high reference value. Online shopping is relatively impersonal and lacks interpersonal warmth and sociability. From the perspective of social presence—the extent to which media allow users to perceive others as psychologically real—this study asks: Do language styles of online reviews influence consumers’ purchase decisions; are there differences across product types; and how/why do review language styles affect purchase decisions?

Online review research spans sentiment/emotion, perceived usefulness, and online WOM effects, but few studies focus on language style. Language style, distinct from language system, reflects how individuals use language to persuade in context. Prior classifications include emotional vs cognitive words (Pennebaker et al., 1997) and review characteristics such as expertise claims, valence, and argument density (Willemsen et al., 2011). Drawing on Kronrod and Danziger (2013), user-generated content is classified into literal vs figurative language styles. Literal language uses objective, plain words to describe characteristics and functions without added connotation; figurative language uses indirect meaning, metaphor, and exaggeration to convey connotation beyond lexical meanings, often correlating with higher emotional intensity. Product type is conceptualized as search vs experience (Nelson, 1970). Search products (e.g., cameras, washing machines) are evaluated by objective attributes accessible prior to purchase; experience products (e.g., movies, music, wine) are evaluated via subjective experience, with information difficult to obtain pre-purchase. Telepresence is an individual sensory experience of the media environment and comprises physical presence ("being there") and social presence ("being with others"). Social presence is the degree to which a person is perceived as real and connected to others via mediated communication, historically conveyed through nonverbal cues but also induced by text. In e-commerce, social presence affects trust, utilitarian/hedonic value, interaction, and shopping pleasure. It can be induced by product information presentation and socially rich text content. From social presence theory, language is a conduit of social presence; imaginative texts can evoke fantasies that reinforce preference and purchase probability. Hypotheses: H1: Literal (vs figurative) reviews increase purchase intention for search products. H2: Figurative (vs literal) reviews increase purchase intention for experience products. Mechanism: Social presence mediates these effects—H3: In search products, social presence mediates the effect of language style on purchase intention; H4: In experience products, social presence mediates the effect of language style on purchase intention. A conceptual model posits an interaction between review language style (literal vs figurative) and product type (search vs experience) on purchase intention via social presence.

Two experimental studies using 2 (language style: literal vs figurative) × 2 (product type: search vs experience) between-subject designs were conducted in China. Study 1: Participants and design—138 undergraduates from Guangzhou were randomly assigned to one of the four cells; after removing 12 invalid responses, 126 valid questionnaires remained (effective rate 91.3%). Stimuli—Search product condition asked participants to imagine purchasing a digital camera; experience product condition asked them to imagine buying a coat (stimuli adapted from Hassanein and Head, 2007; Mudambi and Schuff, 2010). The questionnaire presented product parameters and authentic consumer reviews sourced from taobao.com. Example literal review: "this dress will keep me warm in winter"; figurative review: "wearing this dress can become a fairy!" Measures—Manipulation checks: After reading provided definitions, participants rated review language style (1 = literalness, 7 = figurativeness) and product type (1 = search, 7 = experience). Purchase intention (7-point Likert) was measured with three items (Kronrod & Danziger): "The information in the online comments made me be willing to buy this product", "The information in the online comments made me decide to buy this product", and "I will also consider coming to this online store to buy this product". Reliability: Cronbach’s alpha for purchase intention = 0.953. Study 2: Design matched Study 1 with additional measurement of social presence. Stimuli reflected a search product (mobile phone) vs an experience product (hotel). Measures—Manipulation checks for language style (1 = literalness, 7 = figurativeness) and product type (1 = search, 7 = experience). Purchase intention measured as in Study 1. Social presence measured with items from Hassanein and Head (2007): "There is a sense of human contact on this website", "There is a sense of sociability on this website", "There is a sense of human warmth on this website" (7-point Likert). Reliability: Cronbach’s alpha = 0.918 (purchase intention) and 0.920 (social presence). Analytical approach—Two-way ANOVA tested main and interaction effects on purchase intention; moderated mediation was tested using Hayes PROCESS Model 8 with 5000 bootstrap samples.

Study 1 manipulation checks showed successful manipulations: Language style—figurative descriptions rated more figurative (Mfigurative = 5.71) than literal (Mliteral = 2.19), F(1,124) = 247.58, p < 0.01. Product type—coat perceived more as experience (Mcoat = 5.63) than camera as search (Mcamera = 2.33), F(1,124) = 190.96, p < 0.01. ANOVA: No significant main effects of language style, F(1,122) = 0.009, p = 0.926, or product type, F(1,122) = 1.766, p = 0.186, but a significant interaction, F(1,122) = 10.915, p = 0.001, consistent with H1–H2. Study 2 manipulation checks again confirmed successful manipulations: Language style—Mfigurative = 6.04 vs Mliteral = 1.85, F(1,123) = 254.41, p < 0.001. Product type—Mhotel = 6.08 vs Mphone = 1.90, F(1,123) = 507.73, p < 0.001. Reliability: Purchase intention α = 0.918; social presence α = 0.920. ANOVA: No main effects of language style, F(1,121) = 0.866, p = 0.354, or product type, F(1,121) = 0.438, p = 0.509; significant interaction, F(1,121) = 15.272, p < 0.001. Simple effects: For search products, literal reviews > figurative reviews on purchase intention (Mliteral = 5.60 vs Mfigurative = 4.25), F(1,121) = 11.614, p = 0.001, supporting H1. For experience products, figurative reviews > literal reviews (Mfigurative = 5.16 vs Mliteral = 4.32), F(1,121) = 4.467, p = 0.037, supporting H2. Moderated mediation (Hayes Model 8, 5000 bootstraps): Overall moderated mediating effect 95% CI [1.5772, 3.6496] (excluding 0). Conditional indirect effects via social presence: Search products—mean indirect effect = -1.5056, 95% CI [-2.1618, -0.9249] (significant), supporting H3; Experience products—mean indirect effect = 1.0575, 95% CI [0.5273, 1.6639] (significant), supporting H4.

The findings address the research question by demonstrating that review language style interacts with product type to shape purchase intentions via social presence. For search products, literal reviews align with consumers’ objective evaluation criteria, induce stronger social presence (perceiving reviewers as real, helpful others focused on the reader’s needs), and increase purchase intention. For experience products, figurative reviews more vividly convey subjective experiences, heighten social presence through emotional immersion and imagined usage, and thereby elevate purchase intentions. The results emphasize social presence as a key psychological mechanism in online shopping, showing that textual cues alone can compensate for the impersonality of e-commerce by fostering feelings of human warmth and contact. This advances understanding of how matching communication style to product type can optimize persuasive impact in online reviews and contributes to theory by positioning language style as an antecedent of social presence.

This research demonstrates that the effectiveness of online review language style depends on product type and operates through social presence. Literal language boosts purchase intentions for search products, while figurative language does so for experience products, with social presence mediating both effects. The study contributes theoretically by enriching literature on review language style and identifying language style as an antecedent of social presence in e-commerce contexts. Practically, platforms and marketers should curate and highlight review styles that match product type—objective, literal content for search products; vivid, figurative narratives for experience products—and leverage reviews to enhance social presence through warmer, sociable interactions. It also suggests building data-driven systems to tailor review displays by consumer preferences to increase conversion.

The study focuses only on favorable reviews; future work should include neutral and negative reviews. It examines social presence as the mediating mechanism; other mechanisms may be explored. Temporal factors (e.g., time interval) may moderate review effects and warrant investigation. Individual differences (e.g., reader involvement/expertise, reviewer reputation) could moderate interpretations and deserve study. The experimental method controls many confounds but may limit external validity; field studies and surveys could complement findings. The sample consists of participants from China, which may limit generalizability to other cultural contexts; broader, cross-cultural research is recommended.