Keywords

Online Food and Grocery Shopping, Purchase Frequency, Usability and Convenience, Hedonic Experience, Information Accessibility, Hygiene and Delivery Safety.

Introduction

The adoption of e-commerce in India has evolved from a novelty to a fundamental infrastructure, with online buying becoming increasingly integrated into daily household practices across several categories that exhibit distinct risk profiles and evaluation methods. Food and grocery provide a unique scenario due to their combination of habitual restocking, quality ambiguity, sensitivity to freshness, sanitation issues in transportation, and frequent purchasing behaviours that can swiftly transform an initial unsatisfactory experience into customer attrition or adverse word-of-mouth. In contrast to other durable or uncommon categories, grocery purchasing consistently subjects consumers to operational reliability, including substitutes, condition upon arrival, delivery timeliness, complaint resolution, and safe handling (Rogers, 2003). Behaviourally, this positions online grocery shopping as a high-frequency environment where minor frictions accumulate, and where trust and perceived danger can overshadow mere "usefulness" (Bauer, 1960; Slovic, 1987). In the context of adoption, online grocery can be understood as a perpetual acceptance framework where perceived utility and effort expectancy are significant, although are continually reassessed through service interactions and fulfilment efficacy (Davis, 1989; Venkatesh et al., 2003; Parasuraman, Zeithaml & Berry, 1988). In relational contexts, trust transcends mere disposition; it is empirically evaluated through secure transactions, reliable information, authentic reviews, and effective resolutions in the event of issues (Morgan & Hunt, 1994; Gefen, Karahanna & Straub, 2003).

A dataset-centric methodology is particularly advantageous in Indian e-commerce research, as numerous studies excessively depend on generic "trust" and "usefulness" metrics, neglecting category-specific operational risks (such as freshness, hygiene, and quantity compromise) and the social context of household endorsement in collectivist or family-oriented consumption environments (Triandis, 1995; Hofstede, 2001). The submitted measure specifically incorporates familial recognition and peer encouragement, facilitating an empirically based examination of social influence. The text encompasses perceptions of hygiene and delivery safety, which are especially prominent in the post-pandemic risk landscape and in categories involving perishable goods and frequent delivery interactions, where perceived contamination and safety measures can influence the willingness to persist (Slovic, 1987; Parasuraman, Zeithaml & Berry, 1988).

Review of Literature

The adoption and sustained utilisation in digital commerce are frequently analysed through the Technology Acceptance Model, wherein perceived usefulness and perceived simplicity of use influence attitude and intention (Davis, 1989). Subsequent integrative models identify performance expectancy, effort expectancy, social influence, and facilitating factors as fundamental determinants of behavioural intention and usage, with experience and voluntariness influencing these effects (Venkatesh et al., 2003). In online shopping environments, research consistently indicates that convenience, time efficiency, information availability, and user interface functionality are essential for initial adoption and subsequent purchases (Fishbein & Ajzen, 1977). However, trust and risk perceptions become more critical as transaction stakes, uncertainty, or perceived vulnerability escalate (Gefen, Karahanna & Straub, 2003; Pavlou, 2003). Online food and grocery categories also introduce special amounts of uncertainty about the condition of the products, evaluations of freshness, and delivery cleanliness all of which contribute to greater projected risk than categories that have more homogenous needs (Ajzen, 1991).

The perceived risk hypothesis emphasizes the fact that the consumers evaluate the perceived benefits and the uncertainty and the potential loss that vary between the individuals and the environment (Bauer, 1960). In the psychometric model, dread, unknown risk and institutional trust affect the risk perception which leads to behavioural consequences when the objective risk is minimal (Slovic, 1987). The dangers of online grocery shopping are not limited to the theft of payments but also include the problem of the integrity of fulfilment, broken goods, quantity inconsistency, wrong product, unsatisfactory substitutes, late delivery, and ineffective response to complaints. These risks are functional in nature, and therefore the service system becomes part of the product, which is well aligned with service quality paradigm, which underlines the importance of reliability, responsiveness, assurance, and compassion to the customer in terms of customer satisfaction and loyalty (Parasuraman, Zeithaml & Berry, 1988).

Trust is always noted to be important in e-commerce, as people conduct business without physically looking at them and necessarily relying on platform claims, payment systems and third party logistics (Pavlou, 2003). According to the relationship marketing theory, the prerequisite of sustaining valuable partnerships is the existence of trust and commitment, which implies that the perceived risk is lower and the willingness to continue increases (Morgan and Hunt, 1994). Digital platform trust is multi-dimensional and includes competence, integrity, compassion and predictability of the platform, confidence in the security of payments, and confidence in the quality of information, including belief in the review authenticity (McKnight, Choudhury & Kacmar, 2002; Gefen, Karahanna & Straub, 2003). Some of the trust-related factors in the uploaded data include: the security of payment, the reputation of the site, the credibility of the information and the authenticity of the reviews. This allows the study to operationalise a composite of trust/security in terms of measured indicators as opposed to assumed qualities.

The body of value and pricing literature suggests that people respond to not only the absolute prices but also to perceived fairness, bargain appeal, reference prices as well as promotion framing (Monroe, 1990; Kahneman and Tversky, 2013). Discounts and special offers in digital commerce serve as an economic stimulus and psychological stimulus that helps to reduce purchasing anxiety; however, they can also create distrust in the quality or hidden costs of some categories. Groceries shoppers who use the Internet often consider price as a value in opposition to reliability and freshness. The data include price is significant, reasonable price, attractive discounts, and bulk offers, which allow seeing the promotion and value perceptions in the subtlety of understanding that price is not the only variable.

The social impact is highly relevant to the situation of purchasing through the Internet in the Indian household setting since family members can support or discredit online purchasing decisions, and peer networks can spread the platform use standards (Venkatesh et al., 2003; Triandis, 1995). The scale includes certain questions on family appreciation and peer support. These will help in analysing a social media that many strictly individualistic models of internet shopping poorly characterise, especially in the areas like groceries that affect common household expenditure.

Methodology

The research utilises a dataset of 421 responses and 55 variables. The initial nine variables encompass demographic and background data represented as coded categories: gender, age group, marital status, highest qualification, financial independence, monthly family income, profession, number of dependents, and residential location. All respondents in the dataset confirm that they purchase food and grocery products online, indicating that the dataset comprises active users rather than a mixed sample of adopters and non-adopters. This is analytically beneficial for comprehending continuation and advocacy outcomes among users, although it constrains inferences about hurdles for total non-users.

The perceptual measurement block comprises 35 Likert-type items rated from 1 to 5, addressing platform usability, time efficiency, information accessibility, payment security, platform familiarity, information reliability, review authenticity, personalisation, risk of transit damage, risk of quantity compromise, value for money, availability of transaction modes, billing, delivery punctuality, packaging hygiene, delivery safety protocols, adequacy of replacements, responsiveness to complaints, overall pleasantness, lifestyle compatibility, visual appeal, effort required for quality selection, perceptions of expense, range of options, variety across platforms, freshness, quality equivalence with offline offerings, price prominence, reasonable pricing, attractiveness of discounts, bulk offer incentives, convenience of delivery time settings, order tracking, family endorsement, and encouragement from friends.

The dataset encompasses outcome and behavioural proxy variables such as purchase frequency (coded categories), intention to continue online shopping in the future (three categories), the influence of prior experience on repurchase intention (Likert scale 1–5), monthly spending category, maximum acceptable delivery fee category, preferred platform category, the primary purchase criterion category, the most significant drawback category, and likelihood of recommendation (Likert scale 1–5). This work clearly displays coded distributions of many behavioural variables that are categorised without labels in the file, rather than fabricating meanings for the categories. When the meaning of a code is highly inferable (e.g., continuing intention where code 1 predominantly signifies “Yes”), it is presented as an inference rather than a definitive designation.

Construct formation is exclusively driven by datasets. Multi-item composites are created by averaging items that conceptually group together inside the instrument. Reliability is evaluated by Cronbach’s alpha, acknowledging that alpha is influenced by the number of items and tau-equivalence assumptions, so it is regarded as a diagnostic tool rather than an absolute measure (Cronbach, 1951). The subsequent composites are derived from assessed elements: usability/convenience (user-friendliness, time efficiency, readily accessible information), trust/security (payment protection, reputable websites, reliable information, authentic reviews), perceived risk/effort (risk of damage, compromise in quantity, additional effort to ensure quality, perception of cost), service/logistics (billing accuracy, delivery punctuality, replacement policies, responsiveness to complaints, capability to schedule delivery, tracking systems), hygiene/safety (adherence to packaging hygiene standards, delivery safety measures), hedonic/aesthetics (enjoyable experience, visual appeal), variety (diversity of options, assortment across platforms), freshness/quality (freshness, quality equivalence with offline offerings), price/value (cost-effectiveness, price prominence, reasonable pricing, appealing discounts, bulk purchasing incentives), and social influence (family endorsement, peer encouragement). Single-item measured concepts that retain significance are maintained as individual items (personalisation and lifestyle fit), although they are not regarded as reliability-tested latent constructs.

The analytical strategy unfolds in three phases. Initially, descriptive distributions are presented for demographic codes and essential behavioural codes, accompanied with item-level means to discern strengths and shortcomings in the consumer experience (Hair et al., 2019). Secondly, reliability and inter-construct correlations are evaluated to validate composites and to elucidate any overlaps. Third, outcome models are estimated: an OLS model forecasting recommendation likelihood (1–5) based on the composites; an OLS model forecasting the repurchase-influence variable (1–5) from the composites; and a logistic model predicting continuance intention coded as “Yes” (a binary indicator where the value equals 1 on the continuance item) versus alternative responses, utilising the composites as predictors. These models are interpretative and explanatory rather than causative; the dataset is cross-sectional and self-reported, thereby prioritising connections over causality (Podsakoff et al., 2003).

Analysis

The demographic profile, presented as coded percentages, indicates a generally balanced sample with a little male predominance. Gender codes indicate 55.1% for code 1 and 44.9% for code 2. The age codes are predominantly found in codes 2 and 3, with code 2 comprising 54.9% and code 3 accounting for 27.3%. In contrast, code 4 represents 11.4%, code 1 constitutes 2.1%, code 5 makes up 4.0%, and code 6 is at 0.2%. The marital status distribution is 69.1% for code 1 and 30.9% for code 2. Education is predominantly represented in code 4 (47.0%) and code 5 (39.0%), with lesser proportions in other codes. The financial independence is 52.0 percent code 1 and 48.0 percent code 2. These distributions are more of a contextual control because there are no labels of category means in the dataset; however, they show that the sample is most likely to have a high proportion of young-to-mid adult population, which in India is the typical characteristic of app-based shopping adoption.

The frequency of purchases is also concentrated in one coded segment: 44.9% of the respondents have the purchase frequency of code 4, 17.3% have the purchase frequency of code 2, 12.1% have the purchase frequency of code 3, 10.0% have the purchase frequency of code 1, 8.6% have the purchase frequency of code 6 and 7.1. In the absence of code labels, interpretation should be cautious; however, the concentration indicates a prevailing frequency pattern (for instance, numerous grocery consumers purchase weekly or monthly), aligning with the replenishment aspect of grocery shopping and with continuity models where habit serves as a crucial stabiliser of repetitive behaviour (Triandis, 1995; Venkatesh, Thong & Xu, 2012).

Item-level refers to the assessment of the online grocery experience, highlighting its strengths and identifying areas of friction. The most highly evaluated assertions encompass the availability of several online transaction methods (mean ≈ 4.45), adequate billing facilities (mean ≈ 4.40), time efficiency (mean ≈ 4.30), payment security (mean = 4.25), and simplicity of order tracking (mean ≈ 4.24). The elevated ratings align with TAM/UTAUT predictions that performance expectancies (time efficiency) and facilitating conditions (transaction methods, billing) are fundamental to satisfaction and intention, especially for habitual tasks (Davis, 1989; Venkatesh et al., 2003). They also demonstrate that platforms have effectively established baseline transaction assurance, a crucial precursor to confidence in online buying (Gefen, Karahanna & Straub, 2003; Pavlou, 2003).

The lowest rated items give the source of perceived risk and quality ambiguity. The statement about the quantity compromise has the least mean (approximately 3.03) and the greatest variation therefore indicating that a large percentage is concerned about the inadequacy in the quantity, weight differences, or under-delivery. The average score of the statement online grocery items are normally expensive is quite low (3.29), meaning that there will be mixed perceptions, some of the respondents will consider online shopping to be costly because of the delivery cost, service cost, or high list price, and others may be neutral or disagree because the cost can be reduced through offers. The review authenticity (mean 3.47) is a significant weakness as it shows that trust in digital commerce is not purely created by familiarity with the platform or payment security, review authenticity is a different belief that creates risk perception and belief in making decisions (McKnight, Choudhury & Kacmar, 2002; Gefen, Karahanna & Straub, 2003). The average scores of value-for-money (around 3.50) and the quality being the same as with the offline offerings (around 3.55) are below the strong agreement level, indicating that online grocery shopping faces a persuasive problem. While convenience is recognised, consumers exhibit ambivalence regarding the reliability of online offerings in terms of value and quality compared to offline alternatives (Grewal, Monroe & Krishnan, 1998). This reflects a contradiction between reliability-assurance and outcome-quality in service quality: the system appears user-friendly and secure, however the actual product experience remains ambiguous (Parasuraman, Zeithaml & Berry, 1988).

Composite construct descriptors offer a more lucid strategic perspective. The average usability/convenience score is elevated (mean ≈ 4.14), suggesting that respondents predominantly see user-friendliness, time efficiency, and information accessibility as robust. The trust/security rating is elevated (mean = 3.92), indicating that payment security and platform trustworthiness are well-established. The price/value ratio is similarly elevated (mean ≈ 3.91), which may appear incongruous with the comparatively lower mean for "value for money"; however, the price/value composite incorporates discounts and promotions, which receive high ratings, thereby enhancing the overall perception of economic appeal despite the mixed fairness of the base price (Kahneman & Tversky, 2013; Monroe, 1990). Service/logistics averages are robust (mean in the upper 3s to low 4s range), although risk/effort is moderate, reflecting ongoing apprehensions regarding delivery damage, quantity compromise, and the additional effort required to ensure quality selection. Social influence is considerably elevated, indicating that home and peer environments typically endorse online grocery usage, albeit inconsistently.

Reliability diagnostics confirm that these composites are internally consistent. Cronbach’s alpha is roughly 0.74 for usability/convenience, 0.81 for trust/security, 0.78 for service/logistics, 0.77 for hygiene/safety, 0.77 for variety, and approximately 0.71 for price/value. These align with prevalent thresholds employed in behavioural research, acknowledging that alpha is influenced by item quantity and that constructs comprising two items frequently exhibit lower values despite being conceptually cohesive (Cronbach, 1951). Risk/effort (≈ 0.63), hedonic/aesthetics (≈ 0.64), freshness/quality (≈ 0.61), and social influence (≈ 0.65) are deemed appropriate for the development of exploratory, instrument-anchored composites within a category-specific context. The dataset facilitates significant multi-item measurement across various domains, enabling research to continue without the need to create absent latent structures.

There were 286 responders to code 1, 18 to code 2 and 117 to code 3 as indicated by the continuation intention question (Kim, Ferrin & Rao, 2008). One may infer that the number 1 refers to the Yes, the number 2 and 3 refer to the No and the Maybe/Not sure, respectively; however, this is a wise guess on the part of the study. The overall advocacy mean is likely to be 3.62 out of 1-5, with a median of 4, which means that the advocacy is generally positive, yet not invariably enthusiastic. Such tendency is typical of the service systems in which customers value convenience but feel the existence of a risk of execution which prevents strong recommendations (Parasuraman, Zeithaml & Berry, 1988; Reichheld, 2003).

Outcome modelling offers a clearer elucidation of how experiential impressions translate into advocacy and retention. The OLS model forecasting recommendation likelihood (1–5) based on composite constructs accounts for approximately 25.7% of the variation (R² = 0.257), which is significant for cross-sectional behavioural intention outcomes with perceptual predictors (Forsythe & Shi, 2003). Standardised coefficient patterns indicate that perceived risk/effort exhibits a significant negative correlation, implying that increased perceived risk/effort corresponds to diminished readiness to recommend. Social impact is significantly positive, indicating that when family and friends endorse online grocery purchasing, respondents are more inclined to promote it, in accordance with the principles of diffusion and normative influence (Venkatesh et al., 2003; Triandis, 1995). Usability and convenience, together with hedonic and aesthetic factors, demonstrate positive influences, suggesting that seamless and enjoyable encounters are more likely to be shared and recommended, consistent with the notion that experiencing satisfaction enhances advocacy (Cheung & Thadani, 2012). The multivariate model indicates a negative correlation between freshness and quality, which may arise when the two-item composite reflects a complex belief: respondents might recommend despite persistent doubts regarding freshness or offline equivalence, or multicollinearity with service/logistics and trust/security could induce suppression effects (Dinev & Hart, 2006). The accurate interpretation is not that "freshness diminishes recommendation" as a causal assertion, but rather that within the aggregated predictor set, the distinct variance in freshness/quality beliefs operates unexpectedly, indicating that quality beliefs may be intricately linked with risk perceptions and service experiences in complex manners (Podsakoff et al., 2003).

In the OLS model assessing the item "previous experience affects repurchase intention" (1–5), the explained variance is relatively low (R² ≈ 0.082), suggesting that this outcome may be less accurately measured, more susceptible to unmeasured habits and situational constraints, or encoded in a manner that diminishes linear predictability (Zeithaml, Berry & Parasuraman, 1996). Numerous coefficients have counterintuitive indications, indicating a probable measurement-coding problem: without clarification on whether “1” signifies “strongly agree” or “strongly disagree,” the directional interpretation remains uncertain. This paper regards the repurchase-influence model as exploratory and prioritises more interpretable results (recommendation likelihood and binary continuance), where elevated values distinctly indicate more advocacy and where the "Yes" intention group is recognisable by its predominance. This exemplifies why dataset-anchored work must refrain from over-interpreting outcomes with unknown scale anchors in the file (Jarvenpaa, Tractinsky & Vitale, 2000).

The logistic model forecasting continuing intention "Yes" (coded as 1 on the intention item) exhibits a pseudo-R² of around 0.169, signifying adequate explanatory power for a binary intention outcome. The highest odds ratio pertains to usability/convenience: a one-unit improvement in usability/convenience correlates with approximately 2.03 times greater likelihood of the intention to continue. societal influence is significant, with probabilities of 1.61, and service/logistics at about 1.49, indicating that continuity is closely linked to both the efficacy of personal experience and the prevailing societal endorsement (Kottler & Keller, 2009). These findings correspond with UTAUT's premise that effort/performance expectancies and social influence collectively influence intention, particularly in consumer technology settings where familial norms and peer recommendations are significant (Venkatesh et al., 2003). The model indicates that risk/effort and freshness/quality possess odds ratios below 1 (negative coefficients), aligning with the notion that perceived operational risk diminishes the inclination to maintain online grocery consumption (Bauer, 1960).

The critical interpretative synthesis is the baseline capability/credibility gap framework. The transactional infrastructures are rated highly by the participants, and most of the payment features, precise billing, transactions, and tracking features are rated highly in terms of the average rating of the items. The features are indicative of platform expertise and trustworthiness- crucial conditions of interaction (Parasuraman, Zeithaml & Berry, 1988; Gefen, Karahanna & Straub, 2003). However, the low scores of review authenticity, responsiveness of complaints, value-for-money assurance, and quantity compromise problems suggest that the system is not as reliable and able to recover as much as the ratings suggest (Liu & Shrum, 2002). This in service research implies that failures of moment of truth (wrong amount, poor condition, unresolved complaints) may become a major distraction to confidence and loyalty in comparison with the benefits gained due to convenience (Parasuraman, Zeithaml & Berry, 1988). The supermarket purchases made frequently in the context of risk study; therefore, occasional failures may be more pronounced with time, making the perceived risk higher and lessening the advocacy (Slovic, 1987) Table 1a & Table 2a.

Table 1a Sample Coded Demographics
Variable	Code	n	%
1. Gender	1	232	55.1
1. Gender	2	189	44.9
2. Age	1	9	2.1
2. Age	2	231	54.9
2. Age	3	115	27.3
2. Age	4	48	11.4
2. Age	5	17	4
2. Age	6	1	0.2
3. Marital Status	1	341	81
3. Marital Status	2	72	17.1
3. Marital Status	3	3	0.7
3. Marital Status	4	2	0.5
3. Marital Status	5	3	0.7
4. Highest Qualification	2	1	0.2
4. Highest Qualification	3	110	26.1
4. Highest Qualification	4	174	41.3
4. Highest Qualification	5	124	29.5
4. Highest Qualification	6	10	2.4
4. Highest Qualification	7	2	0.5
5. Are you financially independent?	1	215	51.1
5. Are you financially independent?	2	206	48.9
6. Family Income per month	1	21	5
6. Family Income per month	2	76	18.1
6. Family Income per month	3	51	12.1
6. Family Income per month	4	70	16.6
6. Family Income per month	5	71	16.9
6. Family Income per month	6	132	31.4
7. Your Profession	1	223	53.1
7. Your Profession	2	10	2.4
7. Your Profession	3	21	5
7. Your Profession	4	146	34.8
7. Your Profession	5	20	4.8
8. Number of dependents you have	1	218	51.8
8. Number of dependents you have	2	65	15.4
8. Number of dependents you have	3	81	19.2
8. Number of dependents you have	4	35	8.3
8. Number of dependents you have	5	15	3.6
8. Number of dependents you have	6	7	1.7
9. Residential Area	1	11	2.6
9. Residential Area	2	59	14
9. Residential Area	3	170	40.4
9. Residential Area	4	181	43

Table 2a Purchase Frequency
Code	n	%
1	42	10
2	73	17.3
3	51	12.1
4	189	44.9
5	30	7.1
6	36	8.6

The frequency of online food and grocery purchases is predominantly concentrated in frequency code 4 (44.9%), succeeded by code 2 (17.3%) and code 3 (12.1%), while the other codes constitute smaller clusters. Although there were no code names, this trend clearly shows that a large proportion of about half of the sample had a dominant frequency category that was a main routine. Concentration is important in that it implies that a number of the respondents are not just occasional users but likely have many years of experience with the platform in terms of ordering, delivery, and complaint-solving systems. This is noteworthy since contexts of high usage occasionally make reliability of operations and recovery a more vital criterion in satisfaction and recommendation than the original eagerness of adoption Table 2b.

Table 2b Continuance Intention
Code	n	%
1	286	67.9
2	18	4.3
3	117	27.8

Code 1 (67.9%) and 3 (27.8) have the highest intention to continue, with a small percentage in code 2 (4.3%). The dataset does not explicitly point to which code is the one that signifies the Yes/No/Maybe, however, the fact that the code 1 is the most common one in the majority of survey codings is rather strong evidence that it is the affirmative answer in most of the codings. This implies that the respondent sample is mostly made up of people who would continue to online shop groceries, and this data would be a better fit when examining continuation, loyalty, and the quality of experience instead of examining the causes of non-adoption Table 2c.

Table 2c Recommendation Likelihood Summary
Metric	Value
n	421
Mean	3.62
Std. dev.	0.85
Median	4
% rating 4 or 5	56.1

Likelihood of recommendation (15) was a mean of 3.62, median of 4 and 56.1 percent of those surveyed responded 4 or 5. It means that the advocacy position is mostly favourable, but not universal. The study shows that online grocery services are broadly recommendable to most users; however, a large number of them are ambivalent or skeptical to support it. The gap between utilisation and strong advocacy frequently reflects that the consumers are aware of the specific benefits of convenience but also of risks or lack of consistency that makes them unwilling to suggest the service to other people (Bhattacherjee, 2001) Table 3a.

Table 3a Highest-Rated Experience Items
Item	n	Mean	Std. dev.
23. Different modes of online transactions are available, for online Food and Grocery items	421	4.45	0.8
24. Proper billing facilities are available for online Food and Grocery items	421	4.4	0.85
13. Online Food and Grocery shopping helps to save time	421	4.29	0.91
15. Payment aspects of online Food and Grocery shopping are secured	421	4.25	0.89
44. Orders can be easily tracked, while buying Food and Grocery items online	421	4.24	0.87
16. Online sites providing Food and Grocery items are well-known	420	4.16	0.9
35. Huge range of choices are available, while buying Food and Grocery items online	421	4.14	0.98
36. More variety can be sought, while purchasing from different online sites	421	4.14	0.92

The best-rated claims are centred on the elements of transaction and convenience. The highest mean scores are on the availability of various transactions options (mean ≈ 4.45) and sufficient billing facilities (≈ 4.40), with time efficiency (≈ 4.29), payment security (≈ 4.25) and convenient order tracking (≈ 4.24) coming next. The high averages and relatively mid-range standard deviations indicate that the basic mechanics of the online grocery procedure which are payment, billing, tracking, and security perception are widely accepted and form a stable basis of consumer trust. The study shows that platforms have been most successful in the system trust level, where perceptions about the process of making a transaction are felt as safe and effective, as a condition that leads to the emergence of more advanced levels of loyalty Table 3b.

Table 3b Lowest-Rated Experience Items
Item	n	Mean	Std. dev.
21. Quantity of Food and Grocery items are compromised, while purchased through online sources	421	3.03	1.29
34. Food and Grocery items purchased from online platforms are usually expensive	421	3.29	1.16
18. The reviews on the online sites are authentic	420	3.47	1.01
22. Real value for money can be obtained, while buying Food and Grocery items online	421	3.5	1.08
38. There is no significant difference between quality of Food and Grocery items, available on online and offline platforms	421	3.55	1.13
37. Food and Grocery items purchased online are generally fresh	421	3.62	0.93
29. Complaints are given timely attention, for online Food and Grocery items	421	3.65	1.02
28. Replacements are provided, if the items delivered are not of adequate standard	421	3.68	1.15

The worst rated items imply areas of uncertainty. The minimum mean is related to quantity compromise concerns (mean ≈ 3.03), which also has the largest dispersion (SD ≈ 1.29), indicating a great lack of agreement between the respondents and suggesting that the issue of quantity compromise is a controversial one- some users are trustworthy, others are definitely not. The perception that internet foods are generally expensive is also relatively low (mean = 3.29), which shows that there is an ambivalent value perception despite the favourable rating of discounts and promotions in other situations. The validity of reviews is quite low (mean = 3.47), which can be interpreted as indicators of the fact that people do not trust information as much as they trust payments; they might believe that payments are secure but still do not trust the authenticity of online reviews. The lower score of value of money (approximately 3.50) and a quality match with offline shopping (approximately 3.55) indicates that there is an execution gap: the convenience is high, yet the belief in the value provided and its resemblance to offline shopping is not always high. The less solid components include complaints and replacement, which is noteworthy since service recovery is often the determining factor in turn consumers into potent promoters as opposed to continuing to do so because it is convenient Table 4.

Table 4 Composite Constructs: Reliability and Descriptive Statistics
Construct	Items	Cronbach's α	Composite mean	Composite SD
Usability & convenience	3	0.74	4.14	0.74
Trust & security	4	0.81	3.92	0.75
Personalization	1		3.82	0.94
Perceived risk & effort	4	0.63	3.52	0.79
Service & logistics	6	0.78	4	0.68
Hygiene & safety	2	0.77	3.99	0.83
Hedonic & aesthetics	2	0.64	4.05	0.75
Lifestyle fit	1		4.05	0.91
Assortment variety	2	0.77	4.14	0.85
Freshness & quality parity	2	0.61	3.59	0.88
Price value & promotions	5	0.71	3.91	0.66
Social influence	2	0.65	3.77	0.93

The data enables the development of reliable multi-item scales. The internal consistency of usability and convenience (α = 0.74, mean = 4.14) and trust and security (α = 0.81, mean = 3.92) are good and the average agreement is also high, indicating that the underlying adoption framework is cohesive and has a positive perception. Strong are service and logistics (α=0.78, mean=4.00) and hygiene and safety (α =0.77, mean=3.99), which suggests that the delivery process, tracking and scheduling, and hygiene and safety are considered to be mostly proficient. On the other hand, Perceived risk and effort (α = 0.63, mean = 3.52) and Freshness and quality parity (α = 0.61, mean = 3.59) have lower means and moderate reliability, which is consistent with the idea that perceptions of risk and quality are more context-sensitive and less consistent between users and orders. Price value and promotions (α = 0.71, mean = 3.91) show that the promotions and price signals are widely valued, but this should be combined with the less significant reaction to value of money, which suggests that promotions can increase the perceived attractiveness without properly addressing inherent fairness and value concerns. Social influence (α = 0.65, mean = 3.77) implies that online grocery shopping is largely supported by the family and friends, which, in the context of the Indian household setting, is likely to have a great impact on the recommendation behaviour Table 5a.

Table 5a OLS Model Fit: Recommendation Likelihood
Statistic	Value
N	420
R2	0.257
Adjusted R2	0.235
F-statistic	11.714
Prob > F	0.000

The OLS version of recommendation likelihood explains 25.7 per cent of the variance with an adjusted R² of 23.5 showing a strong level of explanatory power in a cross-sectional behavioural model which uses the perceptual characteristics. The model has a statistical significance in general as indicated by the F-statistic and the p-value of less than 0.001. The total predictor model is a significant explanation of the difference in willingness among the respondents to recommend online food and grocery shopping. The recommendation behaviour on this dataset is not by chance but it is systematically affected by the judgements experienced, especially on risk, social validation, as well as the general quality of the purchasing experience Table 5b.

Table 5b OLS Coefficients: Recommendation Likelihood
Predictor	B	Std. Error	Beta (Std.)	t	p	Sig.
Usability_Convenience	0.137	0.085	0.119	1.603	0.1097
Trust_Security	0.081	0.076	0.071	1.063	0.2882
Personalization	0.068	0.049	0.075	1.382	0.1677
Risk_Effort	-0.199	0.049	-0.183	-4.076	0.0001	***
Service_Logistics	-0.032	0.097	-0.025	-0.328	0.7434
Hygiene_Safety	0.050	0.062	0.049	0.809	0.4190
Hedonic_Aesthetics	0.152	0.081	0.134	1.880	0.0608
Lifestyle_Fit	-0.017	0.053	-0.019	-0.327	0.7435
Assortment_Variety	0.058	0.056	0.058	1.043	0.2978
Freshness_Quality	-0.108	0.054	-0.111	-1.990	0.0472	*
Price_Value_Promotions	0.094	0.088	0.073	1.073	0.2840
Social_Influence	0.154	0.055	0.167	2.813	0.0051	**

The strongest and unambiguous negative predictor of the probability of recommendation is Risk Effort, whose standardised beta is -0.183 and which has a very significant p-value. This implies that the more the respondents believe that they are at risk of being damaged, compromising quantity or making an extra effort in choosing quality items, the less will they recommend online shopping of groceries to their friends. It is an important finding of the paradigm since it confirms that advocacy is suppressed by operational and perceptual friction.

Social influence is another important positive predictor with a standardised beta of 0.167. This shows that among members of the family that support online grocery shopping and friends influence the respondents to support it, respondents are likely to recommend it to others. This finding is especially important in the Indian context where the approval of the house and peer validation often determine the consumption behaviours. Recommendation is not simply a personal judgment of platform effectiveness but it is a socially accepted action as well (Eisenbeiss et al., 2014).

Freshness Quality has a negative coefficient that is statistically significant. This is not to be taken in a naive way as indications that increased freshness reduces recommendations. It implies that of the overall set of multivariate predictors, that the specific variance of the specific composite is suppressive or paradoxical (Hennig-Thurau et al., 2004). This can be because at the point of freshness and quality judgements and trust, logistics, and perceived risk, there is multicollinearity or suppression of those. The attitudes towards freshness and offline equivalency are included in a more complex perceptual system and cannot work in a simple linear manner.

Hedonic Aesthetics has a positive relationship and is on the verge of significance indicating that even pleasant and attractive events can affect recommendations, even though the effect is not above the conventional 0.05 level in this model. However, Usability Convenience, Trust Security, Service Logistics, Price Value Promotions, and other categories do not show any special importance when all the predictors are taken together, even though some of them have relevance on the descriptive or bivariate levels. This means that the largest net effects on recommendations are due to a reduction in perceived risk and improvement in social endorsement Table 6a.

Table 6a OLS Model Fit: Previous Experience Affects Repurchase Intention
Statistic	Value
N	420
R2	0.082
Adjusted R2	0.055
F-statistic	3.040
Prob > F	0.0004

The OLS equation that evaluates the effect of previous experience on repurchase intention explains only 8.2 percent of the variance with an adjusted R 2 equal to 5.5. Although the model is significant in general, its ability to explain is significantly lower than that of the recommendation model. This implies that the repurchase-influence item is either less predictably consistent with the composite predictors, more dependent on other unmeasured factors like habit or situational need, or more difficult to interpret because of the ambiguity in the coding direction of the response scale Table 6b.

Table 6b OLS Coefficients: Previous Experience Affects Repurchase Intention
Predictor	B	Std. Error	Beta (Std.)	t	p	Sig.
Usability_Convenience	-0.254	0.120	-0.173	-2.107	0.0357	*
Trust_Security	0.149	0.108	0.103	1.382	0.1679
Personalization	-0.058	0.069	-0.050	-0.837	0.4033
Risk_Effort	0.075	0.069	0.055	1.095	0.2743
Service_Logistics	0.072	0.136	0.045	0.528	0.5974
Hygiene_Safety	-0.011	0.087	-0.008	-0.124	0.9016
Hedonic_Aesthetics	-0.102	0.114	-0.071	-0.899	0.3693
Lifestyle_Fit	0.124	0.075	0.104	1.648	0.1001
Assortment_Variety	-0.053	0.079	-0.042	-0.668	0.5046
Freshness_Quality	0.061	0.076	0.050	0.801	0.4236
Price_Value_Promotions	-0.305	0.124	-0.186	-2.465	0.0141	*
Social_Influence	-0.023	0.077	-0.019	-0.295	0.7682

The results of this model are surprising especially in Usability Convenience and Price value Promotions which are not only positive but also statistically significant. These constructs would normally have a positive impact on repurchase-based outcomes in a sound directional scale. Their negative indicators strongly suggest that the given dependent variable can have some ambiguity in coding, or that the respondents interpreted the phrase in a way that does not fit directly into a simple higher is better explanation. The exact setting of the scale anchor of this item is not fully clear according to the file; therefore, the directional meaning of these coefficients must be approached with a grain of salt.

The low R² and the unusual angles of coefficients both imply that it is an exploratory model but not a confirmatory one. It is still useful because it shows that the predictor set has much less stable explanatory power of this outcome than either recommendation likelihood or continuing intention. This finding suggests that the views of respondents regarding the presence of previous experience as a predictive of repurchase might be more complicated, less direct, or more connected to personal habits and purchasing requirements than can be explained by the current predictors Table 7a.

Table 7a Logistic Model Fit: Continuance Intention (YES =1)
Statistic	Value
N	420
Pseudo R2 (McFadden)	0.169
LL-Null	-263.735
LL-Model	-219.141
LR chi2	89.188
Prob > LR	0.000

The logistic regression model that predicts continuance intention has a McFadden pseudo-R2 of 0.169, which indicates a good level of explanatory power of a binary outcome. The likelihood-ratio chi-square test is significant, and this means that the model performs significantly better compared to a null model that lacks predictors. The set of experiential composites is helpful in explaining why some respondents report having continuing intention and others do not Table 7b.

Table 7b Logistic Coefficients: Continuance Intention (YES = 1)
Predictor	Logit B	Std. Error	Wald z	p	Odds Ratio	Sig.
Usability_Convenience	0.710	0.271	2.618	0.0089	2.034	**
Trust_Security	0.035	0.241	0.147	0.8831	1.036
Personalization	0.079	0.157	0.505	0.6139	1.083
Risk_Effort	-0.285	0.174	-1.640	0.1010	0.752
Service_Logistics	0.397	0.307	1.293	0.1960	1.488
Hygiene_Safety	0.007	0.194	0.034	0.9732	1.007
Hedonic_Aesthetics	0.243	0.256	0.952	0.3411	1.275
Lifestyle_Fit	0.071	0.170	0.417	0.6764	1.074
Assortment_Variety	-0.022	0.183	-0.119	0.9052	0.978
Freshness_Quality	-0.378	0.183	-2.061	0.0393	0.685	*
Price_Value_Promotions	0.047	0.283	0.168	0.8667	1.049
Social_Influence	0.476	0.173	2.747	0.0060	1.609	**

Usability Convenience is the largest indicator of continuing intention. The odds ratio of 2.034 suggests that one unit changes in this composite is associated with a probability of about doubling the probability of planning to continue online shopping of food and groceries. This influence is substantial, and it proves that the ease of use, time-saving, and the ability to obtain information are important to customer retention. It means that even with a somewhat defined user base, a smooth functionality determines user intentions to stay to a considerable extent.

Social influence is a significant positive predictor with odds ratio of 1.609. This means that the more support a person has, especially family and friends, the higher the chances of continued consumption. This supports the previous recommendation model and proves that the social environment has an effect on advocacy and it directly promotes retention.

The variable Freshness Quality has statistically significant negative coefficient with odds ratio of 0.685 indicating reduced chances of persisting as the unique variation of the construct in the model as the unique variation of the construct increases. Like the recommendation model, this does not justify an inept assertion that novelty is a pure underminer of continuance. A more subtle perspective implies that the beliefs in freshness and parity of offline and online qualities are correlated with such dimensions as trust, logistics, and perceived risk, and their independent statistical significance is less when these other factors are held constant. This implies overlap, suppression or a more complex belief system instead of a direct negative effect on quality judgments.

The Risk Effort is found to be negative which is expected by theory and the odds ratio is less than 1, but, in this case, does not reach conventional statistical significance. Service Logistics is also positive and of moderate size with odds ratio of 1.488, though not statistically significant in the overall model. This means that the quality of logistics is definitely notable but its total effect can overlap with usability and other related service perceptions. The other predictors are not significant when they are mutually adjusted, which implies that continuance is specifically concentrated on smooth usage and social approval whereas quality-related perceptions have a more complex and even overlapping meaning.

Taken together, these tables suggest that the less perceived risk, improved usability, and greater social reinforcement are the major determinants of online grocery retention and advocacy in this dataset. The results reveal that the perceptions of freshness and quality cannot be perceived separately, and they are embedded into a bigger context of trust, logistics, and risk-evaluations.

Findings

The dataset has a number of meaningful, practical implications that can be assessed with regard to the measured variables. The online grocery customers in this sample are Indian, making it evident that these customers are already aware of the convenience of online shopping as the levels of ratings on efficiency, convenience, transparency of bills, and payment methods are high. This is an exemplary utility and labor-saving characteristic in line with technology adoption models (Davis, 1989; Venkatesh et al., 2003). Secondly, the impressions of trust and safety are mostly positive, especially in terms of payment safety and knowledge of the platform. This shows that the main barrier of can I safely transact is properly managed in this group of users in line with e-commerce trust studies which show that, institutional and technical guarantees can form foundation confidence (McKnight, Choudhury & Kacmar, 2002; Gefen, Karahanna & Straub, 2003).

The perceived fulfilment risk and effort are the main barriers to advocacy and retention, which are not related to interface or payment matters. The issue of quantity compromise and the necessity to be more vigilant in the choice of quality is also significantly evident. The higher the perception of risk or effort, the lower the willingness to recommend in the recommendation model, which means that consumers might still be using the platforms due to convenient reasons but unwilling to recommend them because of the fear of potential accountability on negative experiences. This tendency of using privately and carefully advocating usually happens in spheres where outcome is heterogeneous and socially visible (Bauer, 1960; Slovic, 1987).

Fourth, social influence is not a secondary one; it belongs to the category of the most important factors of endorsement and persistence. Once the family members promote online grocery shopping and their friends promote it, consumers will be more interested in recommending the service and expressing their intentions to continue using the service (Zeithaml, 1988). This indicates that online grocery shopping adoption and continuance in India is not solely due to an individual utility but rather it is part of domestic culture that is influenced by social and cultural values (Triandis, 1995; Hofstede, 2001; Venkatesh et al., 2003).

Service and logistics effectiveness are of utmost importance: timeliness of delivery, sufficiency of substitutions, responsiveness to complaints, schedule of deliveries and tracking of them all have significant positive correlations with retention intention. This is in favour of a service-dominant perspective: online grocery is a fulfilment service with a digital interface, and therefore the necessity of operation stability and recovery measures to ensure that customers.

Sixth, the promotional and pricing markers are also largely positive, specifically discounts and offers, but the value-for-money perception is not as high as the appeal of discounts, which means that the offers can be valued, but it does not entirely eliminate the worry about fairness and the total cost. This difference is critical to platform strategy: massive discounts can be used to increase short-term satisfaction, but enduring loyalty demands invariable value propositions and clear charges (Monroe, 1990; Kahneman and Tversky, 2013).

Managerial Implications

This analytic dataset is very informative to the management of online food and grocery platforms in India. Amongst the greatest implications is that the platforms have been successful to build strong reliability in transactions as indicated by the high perceptions of payment security, billing precision, and tracking orders. However, these talents are now becoming norms as opposed to competitive differentiation. This means that managers have to shift their strategic focus of transactional efficiency to experience assurance, particularly in reducing perceived risks of the compromised quantity, damaged products, and inconsistency of the quality offered. The significant negative effect of perceived risk and effort on referral behaviour shows that the slightest failures in operations can significantly reduce advocacy, which plays a crucial role in digitally mediated markets where the word-of-mouth is dominant.

An important conclusion is made in the favourable impact of social factors which is robust. The results show that the support of family members and peers significantly enhances both the probability to get recommendations and the desire. This implies that virtual grocery services should not limit their target audience to individual levels, but actively create a campaign that will increase social acceptance. Adoption and retention can be improved with the help of referral programs, family-oriented promotions, and community-based engagement strategies that involve the online grocery shopping as part of the social setup of families. In a collectivist consumption environment, close social groups approval is an effective trust device to supplement the platform-based warranties.

The importance of usability and ease in increasing continuing intention is another argument to the need to continue improving interface design and user experience optimisation. Even though the consumers are in relatively high level of adoption, they are still concerned with the ease of navigation, quick delivery of order, and the availability of information. The managers are hence encouraged to invest in less cognitive load in product selection, more personalisation algorithms, and quicker reorder processes. Uninterrupted experiences increase retention and indirectly create advocacy through making the service more shareable and recommendable.

The multifaceted and even the contradictory nature of the subject of freshness and quality perceptions tells us that the problem of quality control in online grocery is not merely a product issue, but a perception management issue. Seeing that the perception of freshness is associated with the trust, logistics, and prior experiences, the platforms should possess a detailed quality assurance policy that includes the availability of sourcing information, real-time product status updates, and the reliability of replacement policies. It is important to improve the ways to resolve complaints, as the data shows that discontent with the after sales service may diminish the positive effects of convenience and promotions.

Pricing and promotion findings show that as much as the pricing and promotions are effective to attract the consumers, there is a possibility that they might not fully translate in to a perceived value of money. Promotional strategies should not be relied upon by managers who need to focus on producing the long-term value perception. Open pricing, quality control and quality service delivery are very instrumental in keeping consumer confidence and curbing price-induced loss.

Scope for Further Research

The present research provides useful information on online shopping of groceries, but there are still many avenues of further research. A major limitation is that the dataset is cross-sectional and this limits the ability to infer causal relationships or observe changes in consumer behaviour over time. Future research can use longitudinal research designs to examine the development of beliefs of risk, trust, and service quality over time with increased use and the dynamics of these changes on long-term loyalty.

Another possible opportunity to enlarge it is to combine behavioural and transactional data with perceptual measures. The current research relies on self-reported data, which can be affected by the bias (social desirability and memory errors). Using actual purchase data, delivery performance indicators, and complaint history could result in a deeper understanding of consumer behaviour and support the relationships found in this paper.

The database does not also have obvious segmentation by geographic, cultural, or platform-specific factors. Further studies would be required to examine the regional differences in India since urban and rural consumers might have different preferences and perceptions of risk factors. The effects of platform-specific strategies on consumer happiness and loyalty might be explained by comparative assessment of different e-commerce platforms.

The first opportunity to pursue in the future research is the implementation of this method of analysis to other types of products, in particular, online clothes purchasing. Compared to groceries, however, clothing involves the element of size, fit and taste, and these give rise to different forms of perceived risk and return behaviour. Applying a similar data-driven approach to apparel could help understand whether the risk, social influence, and usability prevalence rates found in the present study are similar across the categories or have a significant degree of variance.

The complex links between constructs may be studied in the future through the use of advanced analytical techniques, including Structural Equation Modelling (SEM), mediation analysis, and moderation analysis. The mediating effect of trust on the relationship between the quality of service and continuance intention and the moderating effect of the demographic factors on the risk perception may present deep theoretical implications and better managerial support.

Conclusion

This paper involves a comprehensive, data-driven research on online food and grocer purchasing behaviour in India, which explains the extent to which experiential attitudes can affect the result of recommendations and retention. The findings demonstrate that the online groceries have managed to develop a robust base on the factors of accessibility, security of transactions and efficiency of operations but these factors alone cannot assist in achieving high rates of advocacy and retention. This strategy aims at highlighting the key role of perceived risk, social influence and experience quality as influencing factors in consumer behaviour.

The adverse connotation of risk perception and effort on the likelihood of a recommendation is a pointer of the necessity to reduce the uncertainty of online grocery shopping. It is possible that customers would remain using online services because it is convenient, but their readiness to suggest the same services to other individuals will largely be influenced by their trust of the quality of the products, and quality service delivery. The difference between the usage and advocacy is paramount, and this implies that platforms must extend past the functional effectiveness in order to build more trust and credibility.

The commonness of the social dynamics that the online grocery shopping in India brings out is an indication that the online grocery shopping in India is not just an individual choice but a behaviour that is socially entrenched. The mediators of adoption and retention are the family approval and peer influence, which emphasizes the concept of digital consumption behaviour as influenced by larger social settings. The websites that are effective in implementing these social interactions will manage to achieve the high rates of client interaction and retention.

The results indicate that the various dimensions of perceptions interact in a complicated manner especially in freshness and quality. The negative relationships that were not anticipated in the multivariate models are some of the reminders of the importance of placing statistical findings into a broader conceptual framework. Such results do not indicate a direct negative impact instead, they indicate that the quality judgements are interdependent with other variables such as trust, logistics, and perceived risk, and have complicated interrelations that need in-depth research.

The study is a contribution to the body of literature on the matter of e-commerce behaviour as it gives an example of the efficiency of the data-driven approach to correlate the empirical analysis with the measured variables. It emphasizes the relevance of the application of theoretical models, including the Technology Acceptance Model, perceived risk theory, and social influence theory that can be utilized in order to comprehend consumer decision-making within the digital environment. Simultaneously, it also points to how managers should include the experiential assurance, social validation and operational excellence as in order to maintain competitive edge in the ever-changing online grocery market.

In conclusion, online grocery shopping in India will depend on the level of technological development and the capacity of the platforms to gain trust, the perceived risk, as well as on the constant value offering. This will be necessary because the online shopping of the groceries will move into a preference and a tested manner of consumption utilizing the convenience, reliability and social validation to the growing expectations of the customers.

References

Angeli, C., & Valanides, N. (2020). Developing and accessing computational thinking skills in school: A critical perspective and a research framework. Computers & Education, 146, 103728.

Google Scholar

Belsky, J. (2025). Generative AI adoption among Indian youth: Trends and digital behavior insights. Journal of Emerging Technologies in Education, 9(1), 45–59

Google Scholar

Cirillo, D., Catuara-Solarz, S., Morey, C., Guney, E., Subirats, L., Mellino, S., & Sanz, F. (2020). Sex and gender differences and biases in artificial intelligence for biomedicine and healthcare. NPJ Digital Medicine, 3(81), 1–11.

Indexed at, Google Scholar, Cross ref

Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340.

Indexed at, Google Scholar, Cross ref

Deka, R., & Parikh, M. (2025). Regional patterns of AI adoption in Indian higher education: A comparative analysis of Northeast and Western India. Asian Journal of Educational Research, 12(2), 87–104.

Google Scholar

DeLyser, L. A., & Born, K. (2021). Gender-inclusive computer science education: Practices, challenges, and opportunities. ACM Transactions on Computing Education, 21(3), 1–22.

Google Scholar

Epley, N., Waytz, A., & Cacioppo, J. T. (2004). On seeing human: A three-factor theory of anthropomorphism. Psychological Review, 114(4), 864–886.

Indexed at, Google Scholar, Cross ref

Gaumat, S., & Rani, M. (2025). Assessing student readiness for AI-based learning through the UTAUT model: Evidence from Indian universities. International Journal of Educational Technology, 8(1), 15–32.

Google Scholar

He, W., & Baxter, R. (2020). Modeling user acceptance of artificial intelligence-based technologies in education. Computers in Human Behavior, 108, 106319.

Google Scholar

Hoori, N. (2024). Awareness and perception of artificial intelligence among higher education students in India. Indian Journal of Educational Research, 10(2), 55–70.

Huang, Y., Park, M., & Choi, J. (2023). Digital familiarity and age-related divides in AI adoption. Technology in Society, 75, 102379.

Google Scholar

James, A., Gupta, P., & Thomas, L. (2016). Inclusive pedagogy for gender equity in technology education. Gender and Education, 28(7), 875–889.

Google Scholar

Kaur, M., & Garg, S. (2021). Artificial intelligence in Indian education: Challenges and opportunities for digital inclusion. Journal of Educational Technology Development and Exchange, 14(2), 55–72.

Google Scholar

Kim, J., & Kwon, S. (2024). Integrating ethical and computational literacy in AI education. Journal of Educational Technology & Society, 27(1), 60–74.

Mukherjee, P. (2022). Digital transformation and higher education in Northeast India: Emerging challenges and opportunities. Indian Journal of Education and Social Change, 9(1), 22–34.

Nass, C., & Moon, Y. (1996). Machines and mindlessness: Social responses to computers. Journal of Social Issues, 56(1), 81–103.

Indexed at, Google Scholar, Cross ref

Natarajan, R., Sen, A., & Kumar, D. (2023). Cultural diversity in AI pedagogy: Toward localized learning frameworks. Education and Information Technologies, 28(3), 4217–4236.

Park, J., Lee, S., & Kim, D. (2021). Age and technology use: Moderating effects of digital experience on AI trust. Computers in Human Behavior, 123, 106878.

Popenici, S. A. D., Kerr, S., & Wang, C. (2024). Student perceptions of AI in higher education: Opportunities and concerns. Higher Education Research & Development, 43(2), 210–227.

Posthuma, R. A., & Campion, M. A. (2009). Age stereotypes in the workplace: Common stereotypes, moderators, and future research directions. Journal of Management, 35(1), 158–188.

Indexed at, Google Scholar, Cross ref

Rai, D., & Singh, K. (2019). Bridging the digital divide: Technological access and equity in Northeast India. Journal of Rural Development, 38(4), 567–584.

Reed, J., Doty, M., & May, S. (2005). Age-related challenges in adopting digital and AI tools. Gerontechnology, 4(3), 149–160.

Rogers, E. M. (2003). Diffusion of innovations 5th.

Google Scholar

Sax, L. J., Lehman, K. J., Jacobs, J. A., Kanny, M. A., Lim, G., Monje-Paulson, L., & Zimmerman, H. B. (2017). Anatomy of an enduring gender gap: The evolution of women’s participation in computer science. The Journal of Higher Education, 88(2), 258–293.

Indexed at, Google Scholar, Cross ref

Seo, J., Li, X., & Zhang, Y. (2024). University students’ attitudes toward AI in education: A comparative study across countries. Computers & Education, 190, 104636.

UNESCO. (2022). Gender equality and bias in artificial intelligence: Global report. UNESCO Publishing.

Viberg, O., Hatakka, M., Bälter, O., & Mavroudi, A. (2023). Teachers’ trust and readiness for AI integration in education: A cross-national study. Computers & Education, 189, 104584.

Williams, K., Jang, S., & Miller, A. (2023). Gender disparities in AI-related education and employment outcomes. Journal of Educational Research, 116(4), 375–389.

Woithe, J., & Filipec, T. (2024). Gender and identity in AI classrooms: Toward inclusive education frameworks. Computers & Education, 192, 104637.

Wong, L., & Kim, S. (2023). Human-AI interaction and student engagement in intelligent tutoring systems. Computers in Human Behavior, 140, 107583.

Google Scholar

Yalcin, E., Lee, H., & Park, S. (2022). Understanding human responses to educational AI agents. Educational Technology Research and Development, 70(4), 1773–1792.

Yigitcanlar, T., Cugurullo, F., & Choi, Y. (2024). Demographic and contextual determinants of AI adoption: Evidence from smart cities. AI & Society, 39(2), 495–509.

Google Scholar

Received: 13-Apr-2026, Manuscript No. AMSJ-26-17194; Editor assigned: 14-Apr-2026, PreQC No. AMSJ-26-17194(PQ); Reviewed: 21- Apr-2026, QC No. AMSJ-26-17194; Revised: 28-Apr-2026, Manuscript No. AMSJ-26-17194(R); Published: 05-May-2026