Pandemic pains to Instagram gains! COVID-19 perceptions effects on behaviours towards fashion brands on Instagram in Sub-Saharan Africa: Tech-native vs non-native generations

ABSTRACT This study represents a novel attempt to investigate the cascading effects of COVID-19 perceptions onto behavioural patterns towards fashion brands on Instagram and across two generations tech-native vs tech non-native) in a Sub-Saharan African context. We drew our study on a sample of 338 Instagram users that experienced fashion brands on Instagram in two Sub-Saharan African countries: Uganda and Nigeria. We used partial least square structural equation modelling (PLS-SEM) to test the hypothetical model. We found that COVID-19 perception positively predicted enjoyment and usefulness, leading to more satisfaction with fashion brand accounts on Instagram and hence greater intention to follow and recommend those accounts. Finally, running a multigroup analysis (MGA), we found the effects of COVID-19 perceptions pronounced into both intentions to follow and intention to recommend via the sequence of mediators: enjoyment and satisfaction were only valid amongst the tech-native generational cohort. Our research suggested a new generational categorisation based on technology nativity – offering a new direction of generational studies in digital marketing communications.


Introduction
As the COVID-19 pandemic ravaged communities across the globe, the amount of time people spent on social media sites and the ways in which they engaged with them were dramatically altered (Molla 2021;Togun 2020). While the proliferation of social media has been an ongoing trend for some time (e.g., Hutter et al. 2013;Mohr et al. 2021), people facing social distancing measures took to social media during lockdowns and quarantine as a means of gathering information and maintaining a connection with others in ways that may fundamentally change our relationships with these technologies (Molla 2021). Pandemics and other catastrophic disruptions are ever-present threats that are only expected to increase in frequency in the future (Gill 2021). Thus, examining the generalizability of findings from extant research on social media use in the context of a catastrophic disruption is relevant to marketing communication theory and practice. Notably, the double enjoyment-usefulness approach that is regularly applied to the study of social networking sites (Casaló, Flavián, and Sergio 2017;Van der Heijden 2004;Mahmoud et al. 2021d) is worth revisiting in this extreme context that constitutes an unprecedented disruption to contemporary marketing (Mahmoud et al. 2020b(Mahmoud et al. , 2021d. Specifically, it is important to understand if individuals might be more perceptive of the usefulness of and/or the enjoyment they derive from branded social media accounts in the context of extreme disruption, given the evolution in the way social media has been used as a result of COVID-19. Moreover, given that use of social media (e.g., Instagram) tends to trend younger and established generational differences in motivations for using social media (e.g., Calvo-Porral and Pesqueira-Sanchez 2019), it would be helpful to understand if generational cohort (digital native vs non-native) impacts these behaviours. We focus on Instagram as a context for this inquiry for several reasons.
Since its creation in 2010, Instagram has become a critical platform for businesses (Elliot 2014). Allowing brands to connect with consumers more rapidly and efficiently than traditional media channels (Hassan, Nadzim, and Shiratuddin 2015) and outpacing the growth of other social media sites (Djafarova and Rushworth 2017), Instagram has begun to supplant the influence of traditional media on consumers (Duffett 2017). Instagram is unique amongst social media platforms in that it is particularly visual in nature (Bianchi 2021;Casaló, Flavián, and Sergio 2017). This allows consumers to have more meaningful interactions with brands (Serafinelli 2018) and allows the promotion of brands to generate public awareness (Geurin-Eagleman and Burch 2016). Moreover, Instagram is popular with younger audiences, an important consideration given that more than half of the global population is under the age of 34 (Statista 2020).
Further, Instagram uses increased 40% during the COVID-19 pandemic (Kantar 2020), suggesting that it was a key social media tool for keeping informed and connected during the COVID-19 pandemic. Among peer platforms, Instagram was the only one to see an increase in incoming messages, with an increase of 6.4 comments per day (Arens 2020). Given the visual nature of Instagram, however, it is unclear how this increased attention affected behaviours toward fashion brands and whether it was experienced differently for tech-native vs non-native generations. This paper contributes to the marketing communication literature by exploring these relationships in the context of extreme disruption (e.g.,  as experienced in Sub-Saharan Africa. Sub-Saharan Africa provides an interesting context to explore this research agenda for several reasons. Principally, the relationships between perceived features (i.e., usefulness and enjoyment), overall attitude (i.e., satisfaction), and behavioural intentions (i.e., intentions to follow and intentions to recommend) as well as the moderating role of generational cohorts, was established in Sub-Saharan Africa prior to the emergence of COVID-19 (Mahmoud et al. 2021d). This established baseline allows us to explore whether those recognized relationships are generalizable to the context of extreme disruptions (e.g., COVID-19 pandemic). Further, Instagram is already quite popular in the region as social media is a preferred means of communication in Sub-Saharan Africa due to the high costs associated with SMS and traditional phone calls (Gillwald et al. 2019). In Nigeria, for instance, Instagram has a penetration rate equal to 25% (Mahmoud et al. 2021d), and usage of the platform in Sub-Saharan Africa appears to be poised for an uptick given the recent launch of Instagram Lite in the region in March of 2021 (Adejumoh 2021). Further, digital marketing is increasingly being used as part of business strategy in the region (Eze et al. 2020).
Given that Instagram users trend younger (Tankovska 2021), there has been a lacuna in the research regarding the engagement of older generational cohorts with the platform. This omission is surprising since this population often represents a sizeable portion of the consumer market these brands cater to and established generational differences in social media usage (Calvo-Porral and Pesqueira-Sanchez 2019). The gap in the literature becomes particularly glaring in the context of the COVID-19 pandemic, which has seen these understudied generational cohorts adopt social media usage en masse (Togun 2020). While Mahmoud et al. (2021d) did find generational differences in Instagram usage in Sub-Saharan Africa, it is unclear whether these findings generalize to the context of COVID-19. This is important for marketing communication research and practice since attitudes towards and usage of social media networks have undergone a profound transformation during the pandemic (Togun 2020). Accordingly, the primary purpose of this study is to examine the moderating role of digital native vs non-native generational differences on a path model linking COVID-19 perceptions -defined as ones' subjective assessment of their personal concern with the negative effects of the pandemic -to intentions to follow and recommend Instagram accounts. We explore this via perceived usefulness, enjoyment of, and satisfaction with fashion brands' accounts amongst a sample of Instagram users in two Sub-Saharan countries: Nigeria and Uganda.
The present research makes several notable substantive and theoretical contributions. First, we heed calls to investigate consumer behaviours in non-idyllic contexts such as pandemics (e.g., Klein and Hill 2008;Mahmoud et al. 2021d) and in understudied populations (e.g., non-WEIRD samples; Henrich, Heine, and Norenzayan 2010), as prior findings may need to be evaluated in terms of their universality. Further, we add to the literature on the use of social media as a marketing tool (e.g., Alalwan, Rana, Dwivedi, & Algharabat, 2017) by exploring consumer behaviours on social media, rather than the more frequently studied nature of the content of social media posts (see Casaló, Flavián, and Sergio 2017).

COVID-19 and social network usage
The COVID-19 pandemic has influenced the dynamics of social media usage in many ways. While an increasing share of communications has been taking place within social network environments even prior to the pandemic (Hutter et al. 2013), the social distancing measures put in place to combat the spread of the disease accelerated this ongoing trend (Molla 2021). Not only are people spending more time than ever on social media, but they are also forced to navigate an unprecedented amount of disinformation so severe that some have declared it an 'infodemic' (Raman 2020). Moreover, users are increasingly suffering from social media fatigue and are subsequently actively avoiding screen time (Fauville et al. 2021); adjusting the content of social media posts so as not to appear tone-deaf (Wold 2020); 'doomscrolling', or immersing themselves in an information environment on social media that is relentlessly negative (Watercutter 2020); and building online communities to stave off the effects of social isolation -often around brands (Bishop 2020). Even populations that have not historically been well represented on social media (e.g., the elderly and those living in poverty) have had no choice but to engage with social networks (Togun 2020).
Brands, too, have altered their social media habits during the pandemic, often taking a more empathetic approach to content creation and shifting away from more typical humorous posts deprecating competitors (CBS NEWS 2020). Rather than posting content meant to entertain or be useful, brands have regularly posted messages of solidarity; public service announcements urging consumers to comply with preventative measures like facemask wearing, hand washing, or social distancing; and announcements meant to raise awareness of their CSR efforts (Ragavan 2020). Brands that are successfully navigating the global health crisis are increasingly focussing on engagement with consumers and less on promotional efforts on social media platforms (Salzano 2020). It appears the brands that are in the strongest position to capitalize on social media are those with established relationships with consumers as they provide a trusted source of information (Salzano 2020) in an information environment chock-full of misinformation.
Taken together, the disruptions to both user and brand social media habits in the context of the COVID-19 pandemic reinforce the need to explore the generalizability of prior research that has explored consumer interactions on social media platforms. In particular, research has established a relationship between consumers' perceived usefulness and enjoyment of social media content on downstream intentions to follow and intentions to recommend the social media account (Casaló, Flavián, and Sergio 2017). Further, it is clear that attitudes regarding COVID-19 have real-world behavioural implications. Throughout the course of the pandemic, we have witnessed two extremes of behaviour; some downplay the seriousness of the disease, carrying on as though the pandemic did not exist (e.g., covidiots; Miller 2020) and others taking extreme precautions -sometimes to the point that such behaviours are maladaptive -to avoid exposure to the virus (e.g., coronaphobia; Arora et al. 2020). Mahmoud et al. (2021a) define COVID-19 perceptions as one's perceived concern with the negative effects of the virus. Such perceptions have been shown in extant literature to impact a range of behaviours, including consumers' eating behaviours (Mahmoud et al. 2021a) and employees' customer orientations (BLINDED Mahmoud et al. 2020b). The negative impacts of social isolation and fear of contracting COVID-19 on mental wellbeing are well established and may lend to increased escapism through the use of social media (Eden et al. 2020). Given the motivation to restore personal wellbeing, anecdotal evidence suggests that people might be more eager to consume the Instagram posts of familiar, trusted brands as a method of coping or as a form of escapism (Tietjen 2020). We contend that the extreme context imposed by COVID-19 has made people more sensitive to and thus more likely to perceive the benefits of using and their enjoyment of online interactions. Formally, we predict: H1: COVID-19 perception positively predicts the perceived usefulness of fashion brands accounts on Instagram.

COVID-19 perceived usefulness and satisfaction with fashion brand account on Instagram
Perceived usefulness is defined as the extent to which technology results in user goal achievement and positive affect towards the technology (see Davis 1989;Davis et al. 1992). According to Basak and Calisir (2015), perceived usefulness is related to the extent to which a social media account provides the information users need to make specific decisions. In the context of Instagram, the perceived usefulness of Instagram pages leads to user satisfaction with an Instagram account (Alhabash and Ma 2017;Huang and Su 2018). Therefore, we hypothesize that: H3: Perceived usefulness of an Instagram account has a positive effect on users' satisfaction with using the platform during COVID-19

COVID-19 perceived enjoyment and satisfaction with fashion brand account on Instagram
In the context of social networking sites, perceived enjoyment refers to the degree of fun, relaxation, mental stimulation, and pleasure that people experience when interacting with social media (Casaló, Flavián, and Sergio 2017). This may derive from playing games, engaging with others and, in some cases, engaging with new technologies (Cheah et al. 2020; Nayal and Pandey 2020; Van der Heijden 2004) or a particular brand (Muntinga, Moorman, and Smit 2011). In the case of Instagram, fashion brands often post interesting and interactive content to drive attention and engagement. For example, they may post a poll or engage users in a contest that users may find enjoyable. This perceived enjoyment leads to a pleasant experience that creates a likelihood that the user will perceive their user experience as satisfactory. This leads us to hypothesise: H4: Perceived enjoyment of an Instagram page has a positive effect on users' satisfaction from visiting that page during COVID-19

COVID-19 satisfaction and behavioural intentions towards fashion brands' Instagram accounts
The relationship between user satisfaction and positive intentions toward fashion brands has been well established (Casaló, Flavián, and Sergio 2017). According to Mazzarolo, Mainardes, and Innocencio (2021), satisfaction with the Instagram experience can lead users to endorse brands and increase purchase intention. In the case of fashion brands, satisfied users often return to the brand's account when owners of the account share new content (Casaló, Flavián, and Sergio 2017). Since customers of fashion brands often follow the company's Instagram account and engage with it online (Casaló, Flavián, and Sergio 2017), likely, user satisfaction with a fashion brand's Instagram account will positively impact other user behaviours, such as intentions to follow the account and likelihood of positive word of mouth. We, therefore, hypothesise that: H5: Users' satisfaction with an Instagram fashion account has a positive effect on users' intention to follow that account.
H6: Users' satisfaction with an Instagram fashion account has a positive effect on users' intention of recommending that page to other Instagram users.

COVID-19 perception and behavioural intentions towards fashion brands' accounts on Instagram
Users of an Instagram brand page have positive perceptions of the experience when the site is useful, enjoyable or some combination of the two (Casaló, Flavián, and Sergio 2017). This can be the result of satisfying content (Seol et al. 2016) that meets their expectations and inspires positive behaviours with and toward the brand on and off Instagram (Mahmoud et al. 2021d). Therefore, it is presumed that a positive indirect relationship exists between the user's perception of the brand page (e.g., perceived usefulness and perceived enjoyment), their intention to follow the brand Instagram account, and their intention to recommend the brand to others. Accordingly, we hypothesise that: H7: COVID-19 perception indirectly and positively predicts users' intention to follow the brand and intention to recommend the brand to other consumers via two sequences of mediators:

H7a: Perceived usefulness of and satisfaction with fashion brands on Instagram
H7b: Perceived enjoyment of and satisfaction with fashion brands on Instagram.

Technology nativity as generational categorisation
Like the rest of the world, Sub-Saharan Africa has seen older people turning to social media in order to maintain connections with others (Togun 2020). Individuals in this region have also been forced to contend with massive amounts of disinformation disseminated through social media (Togun 2020). This onslaught may change their behaviours on the platforms -perhaps causing them to process incoming information more defensively on one hand while entrenching themselves with trusted brands on the other.
Inter-generational differences have been observed with respect to the motivations underlying technology usage by digital native (e.g., Generation Y and Generation Z) and non-native (e.g., Generation X) users. Specifically, digital natives have been shown to be motivated by the hedonic, entertainment value of technologies, whereas non-natives are more utilitarian (Calvo-Porral and Pesqueira-Sanchez 2019). It is well-established that ones' salient motives are capable of influencing consumer judgments (Wood and Hayes 2012) and perception (Ginn and Lickel 2020;Maner et al. 2005;Pusch et al. 2021). Importantly, motives direct attention and make people more sensitive to motive-related information (Schultheiss and Pang 2007). If generational cohorts are chronically predisposed to specific motivations when using social media, we should expect to see motivational biases in the perception of social media accounts. Given their divergent motivations for technology usage, it stands to reason that these generational cohorts may be asymmetrically sensitive to the usefulness of (i.e., utilitarian benefits) and enjoyment derived from (i.e., hedonic benefits) of Instagram accounts. That is, tech-natives will be attuned to enjoyment-related information, while non-natives will be attuned to information pertaining to the usefulness of social media accounts. Accordingly, we hypothesize that: H8: Users' generation (tech-native vs tech-non-native) moderates the path from COVID-19 perception and intention to follow and intention to recommend through the usefulness of and satisfaction with fashion brands on Instagram H9: Users' generation (tech-native vs tech-non-native) moderates the path from COVID-19 perception and intention to follow and intention to recommend through the enjoyment of and satisfaction with fashion brands on Instagram

Methods
Our study took place in two Sub-Saharan African countries: Nigeria and Uganda, and data were gathered through an online survey in 2021. Due to the dispersion of the targeted population, we identified different Instagram users and specifically for generations (X, Y and Z) following accounts that are sharing fashion-related content on Instagram. That is, we purposively recruited participants who were Instagram users and had interactions with accounts of fashion brands on Instagram. In that regard, two screening questions were placed at the beginning of the survey. The two yes-no screening questions were: 'Do you have at least one account on Instagram?' and 'Do you follow any accounts sharing fashion-related content on Instagram?' This was aimed at selecting those individuals who are well conversant with Instagram and fashion-related content. We projected the questionnaire and guided first-time users on how to input their responses. After the demonstration, respondents were now able to complete and submit their responses. That was majorly for generation Y and generation Z. On generation X participants, we had rapporteurs to identify and guide participants to give their responses.
At some time, a snowball method of sample selection was applied. In this case, for every recruited participant, we asked them to assist us in identifying other potential participants for the study. Out of a sample of 500 participants, we received a total of 389 responses (response rate = 78%), of which 338 were deemed valid as fifty-one responses were filtered out due to failing the challenge of the screening questions. The final sample (n = 338) had a nearly equal representation of gender and age groups:1) 18-22 years old; 2) 23-40 and 3) 41 or above and substantial size of each group (Malhotra et al. 2019). We utilised the work of Mahmoud et al. (2021a) and Mahmoud et al. (2020b)  Casaló, Flavián, and Sergio (2017) Visiting this Instagram account helps me get new ideas about fashion.
In general, visiting this Instagram account is useful.

Satisfaction
Overall, I'm satisfied with my experience with this Instagram account Casaló, Flavián, and Sergio (2017) and current study Overall, I think following this Instagram account is a good idea. Overall, this Instagram account is valuable.
Intention to Recommend I would likely recommend this Instagram account to friends and relatives interested in fashion.
Casaló, Flavián, and Sergio (2017) I would seldom miss an opportunity to tell others interested in fashion about this Instagram account on social networking sites.
Intention to Follow I predict that I will keep following this Instagram account. I will continue to look for new content published on this Instagram account. I will probably look for new content published by this Instagram account.
COVID-19 perception I believe that the effect the coronavirus pandemic has had on people is to measure COVID-19 perception and Casaló, Flavián, and Sergio (2017) alongside items designed for this study to measure usefulness, enjoyment, satisfaction and intention to follow and intention to recommend fashion brand accounts on Instagram (see Table 1). Consent to take part in the study was shown at the beginning of the survey. Due to the fact that the survey was carried out online, the signatures of the participants were not acquired (Mahmoud et al. 2021d). We told all participants that their responses would be kept confidential and anonymous. We marked all of the questions as 'Required' to avoid cases with missing data. However, all respondents were made aware of the research's purpose and processes during the surveying process. They were free to ask questions, raise concerns about the questionnaire, or leave the survey at any moment. As a result, rather than 'forcing' people to reply, the process tried to avoid missing data (Mahmoud et al. 2021d). Additionally, our study did not touch on personally sensitive questions (Sischka et al. 2020). As a result, we contend that the low dropout rate demonstrated by our survey's high response rate (78 per cent) implies that there was a slight bias related to the force answering (FA) in the current research (Stieger, Reips, and Voracek 2007;Mahmoud et al. 2021d). Conclusively, because participants do not have to disclose sensitive details in front of investigators, online surveys are less prone to socially desirable responses (Gnambs and Kaspar 2015). The majority of our sample were female (52%), millennials (42%), tech-native (71%), educated to a degree level (50%), based in Uganda (81%) and tech-native (71%).

Measures and procedure
We present a new categorisation of generation in studying marketing communications, based on technology nativity (Prensky 2001), and, thus, our study focused on two generational groups: tech-native and tech-non-native. In that respect, we identified the cut-offs based on the description of later generations, i.e., the millennials and Generation Z, as tech-savvy or native compared to older generations, e.g., Generation X, in previous research (Chakrabarti and Makhija 2021;Kitchen and Wheeler 1997;Mahmoud et al. 2020bMahmoud et al. , 2020cMahmoud et al. , 2021cMahmoud et al. , 2021dMehra, Paul, and Kaurav 2020;Palfrey and Gasser 2011;Prensky 2001;Seemiller and Grace 2019;Windisch and Medman 2008). Since our data were collected in 2021, participants aged between 41-58 years were deemed Gen X, between 23-40 years, Gen Y and 18-22 years, Gen Z.
The measures employed in the current study were multi-item, reflective and assessed on a seven-point Likert scale. We calculated the Heterotrait-Monotrait Ratio of Correlations (HTMT) and had values lower than .9 (see Table 2). Table 3 shows that all the constructs had average variance extracted (AVEs) higher than 0.5, CRs between .7 and .9 (Hair et al. 2017), and Variance Inflation Factor (VIF) values less than 3. Our results, therefore, suggest all the measures utilised in the research were admitted to a sufficient class of the discriminant validity, construct reliability, and convergent validity quality criteria. Common-Method Bias (CMB) was also assessed and returned no concerning inner VIFs (see Table 4), as all were lower than 3.3 (Kock 2015). Finally, Table 5 shows the average scores for the whole sample as well as for each tech-nativity group.

Results
The variance-based (or partial least square) method of structural equation modelling (PLS-SEM) is the primary statistical technique to test the research hypotheses. For that purpose, SmartPLS 3 (v. 3.3.3) is utilised (Ringle, Wende, and Becker 2015). Our choice of the PLS-SEM approach is based on previous research recommendations when assessing predictive models (Hair, Ringle, and Sarstedt 2014;Mahmoud et al. 2021b). Moreover, the literature (e.g., Mahmoud et al. 2020a) suggests that multivariate normality is likely to contravened by most data. Additionally, an ever-increasing body of literature has endorsed PLS-SEM for empirical research investigations where data are prone to non-normality bewilderments. (Hair et al. 2017  We test our hypotheses by running a path analysis and multigroup analysis (MGA). This includes the use of standardised betas (β: for direct effects), unstandardised betas (B: for indirect effects) and t-values associated with these coefficients in a consistent-withbootstrapping mode. Also, following the guidance by Hair et al. (2019), we use f 2 to  determine effect sizes and PLSpredict to assess out-of-sample prediction. In order to assess the model fit to our data, standardised root mean square residual (SRMR) is calculated (Henseler et al. 2014). We conduct Consistent-PLS with connections between all latent variables (LVs) activated, as suggested by Dijkstra and Henseler (2015) for the estimation of the latent variables scores, to obtain more stable results. We follow that with running a Consistent PLS Bootstrapping at 5,000 sub-samples (Preacher and Hayes 2008). We find that COVID-19 perception positively predicts both usefulness (β = .373, P < .001, f 2 > .15) and enjoyment (β = .387, P < .001, f 2 > .15). Further, both usefulness (β = .539, P < .001, f 2 > .35) and enjoyment (β = .427, P < .001, f 2 > .35) positively predict satisfaction that in turn positively predicts intentions to follow (β = .872, P < .001, f 2 > .35) and intentions to recommend (β = .822, P < .001, f 2 > .35). Building on our statistics, therefore, we judge H1, H2, H3, H4, H5 and H6 as valid. All of the additional paths are found associated with nonsignificant p-values and poor effect sizes, hence unsupported (see Table 6).
Furthermore, Table 7 demonstrates that all unstandardised betas are positive and significant at a probability value less than .001. Based on that, we conclude that H7a and H7b are fully supported. Given that the paths additionally hypothesised are not valid as concluded above, this finding means that all of the mediators (i.e., usefulness, enjoyment and satisfaction) are full transmitters of COVID-19 positive effects onto the participants' intention to follow and intention to recommend -meaning that Instagram users  can develop stronger intentions to follow and intentions to recommend fashion brands' account on Instagram because the pandemic experience makes this virtual space feel more useful and pleasing, hence satisfying. Based on Hu and Bentler (1999), our calculations show that the SRMR value equals .044 < .08, implying that our hypothetical model is a good fit for our data. Finally, Table 8 shows that the vast majority of the observed variables in the PLS-SEM analysis have lower scores of the mean absolute error (MAE) and the root mean square error (RMSE) as compared to the naïve LM benchmark . This suggests that our model has medium predictive power (Shmueli et al. 2019). Henseler, Ringle, and Sarstedt (2016) contend that performing multigroup analyses (MGAs) through PLS-SEM can be unmeaningful and posing the risk of obtaining 'misleading' results unless their measures invariance is proved. This criterion can be achieved using the measurement invariance of the composite models (MICOM) technique (Henseler, Ringle, and Sarstedt 2016). Accordingly, at least (in this instance: with no data pooling), both configural invariance and compositional invariance ought to be verified and established before performing any multigroup comparisons . Because we use a PLS-SEM approach, the measurement configural invariance is, by default, accomplished (Hair et al. 2018). Therefore, we proceed with examining whether the second criterion, i.e., compositional invariance, is established. In this regard, we run a permutation test. Table 9 shows that all of the variates have their 'Permutation P-values' above 0.05. Therefore, we judge the null hypothesis as supported, implying that the variates' original correlations are nonsignificantly different from 1, hence establishing compositional invariance (Hair et al. 2018). We run a multigroup analysis (MGA) to assess the hypothetical model invariance across tech-native and tech-non-native groups and detect if the technology nativity based generational differences moderate the hypothesised path postulated by Figure 1. We use t-values that are related to the comparisons stated in the parametric tests. The results (see Tables 10 and 11 and Figure 2) show that the direct path from COVID-19 perception to enjoyment is not invariant between the two generational groups leading to non-invariance of the indirect effects of COVID-19 perceptions onto satisfaction and hence behavioural intentions through enjoyment between tech-native and tech-non-native participants. In other words, more intense COVID-19 perception leads to higher levels of perceived enjoyment only amongst tech-native generation (β tech-native = .384, P< .001; β tech-non-native = .018, P= .853; t tech-non-native vs -native Generations = 2.079) and thus, tech-native participants are  substantially more likely to be satisfied with (B tech-native = .156, P < .001, SD = .038; B tech-non-native = .007, SD = .048, P = .861; t tech-non-native vs -native Generations = 2.168), follow (B tech-native = .122, P < .001, SD = .032; B tech-non-native = .006, SD = .033, P = .865; t tech-non-native vs -native Generations = 2.102), and recommend (B tech-native = .111, P < .001, SD = .029; B tech-non-native = .006, SD = .031, P = .867; t tech-non-native vs -native Generations = 2.098) fashion brands' accounts on Instagram because they find more enjoyment during the current pandemic than participants from tech-non-native generation. Therefore, we judge H8 as unsupported whilst H9 as supported.

Discussion
In scanning the literature, no data was found on the effects of COVID-19 perception on social media users' behaviours towards brand accounts on Instagram. Furthermore, very little was found in the literature on the tech nativity or non-nativity-based generational differences in attitudinal and behavioural patterns of Instagram users towards brands presence on the visually rich social media platform. Therefore, the primary purpose of this research was to examine the effects of COVID-19 perceptions -defined as ones' subjective assessment of their personal concern with the negative effects of the pandemic-on behavioural intentions towards fashion brands on Instagram through usefulness, enjoyment and satisfaction that are the raison d'être of Instagram users' intentions to follow and intentions to recommend as suggested by previous research (e.g., Casaló, Flavián, and Sergio 2017). Also, we aimed to assess the model invariance across two generations (tech-native vs tech non-native) in a Sub-Saharan African context. We analysed a path model where COVID-19 perception is hypothesised to indirectly predict Instagram users' intentions to follow and intentions to recommend apparel brands on Instagram via a sequence of mediators comprised of perceived usefulness, perceived enjoyment, and satisfaction with those accounts on Instagram. We did so by utilising already validated measures to assess a sample of Instagram users' COVID-19 perception, perceived usefulness, enjoyment, overall satisfaction, intentions to follow and intentions to recommend as a result of their experience with such accounts in two Sub-Saharan African countries, specifically Nigeria and Uganda. We found that COVID-19 perception positively predicted usefulness (supporting H1) and enjoyment (supporting H2). Further, both perceived usefulness (supporting H3) and enjoyment (supporting H4) positively predicted satisfaction that in its turn positively predicted intentions to follow (supporting H5) and intentions to recommend (supporting H6). Also, the indirect effects of COVID-19 perception on intentions to follow and intentions to recommend were found to be positive and significant through two sequences of mediators: usefulness => satisfaction (supporting H7a) and enjoyment => satisfaction (supporting H7b). Finally, the effects of COVID-19 perception on enjoyment as well as indirect effects on behavioural intentions via the sequence: enjoyment => satisfaction were not invariant between tech-and non-native users offering support of H9. H8 was rejected as the path: COVID-19 perception => usefulness => satisfaction => intentions to follow and intentions to recommend was invariant between the two generational groups.

Practical implications
The analysis concluded that more strongly negative perceptions of COVID-19 translate indirectly into stronger intentions to follow and intentions to recommend fashion brands' accounts on Instagram. Put another way, those who are more worried by COVID-19 and perceive its negative personal impact are likely to turn to social media and its utilitarian and hedonic value to relieve their personal distress. The pandemic context leads Instagram users to be more attuned to and thus likely to perceive features consistent with their chronically salient motives for using social media; specifically, the usefulness of (if utilitarian motives are salient) and the enjoyment they derive from (if hedonic motives are salient) fashion brands' Instagram accounts. Such motivated perception subsequently leads to positive satisfaction levels, hence more favourable behavioural intentions towards these Instagram accounts. We also found that tech-nativity-based generational differences moderate the path from COVID-19 perception to behavioural intentions towards fashion brands accounts on Instagram via perceived enjoyment and satisfaction (H9). Specifically, our results suggest that the relationship between COVID-19 perception and the enjoyment of fashion brands' accounts on Instagram is only valid amongst tech-native Instagram users, as are the COVID-19 perception indirect effects on intentions to follow and intention to recommend transmitted via enjoyment. On the other hand, one interesting finding is that COVID-19 perceptions indirect effects on behavioural intentions are transitioned via perceived usefulness for both tech-native and tech-non-native generations -meaning that both generations are likely to act favourably towards the brands they perceive as providing them benefits on a visually rich social media platform like Instagram. However, employing hedonic stimuli in the content shared by fashion brands on Instagram will be instrumental in driving engagement with tech-native generations as they are chronically motivated to use social media for its entertainment value, particularly in the context of a global health crisis and the pervasive negativity associated with it. This discovery has practical implications for marketers concerning their endeavours on social media, primarily when they aim to encourage the spread of positive word of mouth behaviours regarding their brands in the virtual spaces -activities that are especially needed given the downturn in economic activity brought by pandemics. Online word of mouth recommendations have been evidenced by previous research (e.g., Roy, Datta, and Mukherjee 2019) to have a strong influence on purchase intentions, hence actual buys. Also, favourable user-generated content on social media contributes positively to brand equity and brand attitude (Schivinski and Dabrowski 2016), especially in the fashion industry where the choice of social media channels for marketing communications would be highly critical (Anselmsson and Tunca 2019). Therefore, targeting younger generations who are tech-native can be more effective if fashion brands supplement the content shared with a proper blend of entertaining elements like music, games, riddles, etc., which are already doable given the fast-growing set of rich-media features within the Instagram environment. This is in line with the depiction of the region where youth is the majority (e.g., Ngwainmbi 2019), opening up great opportunities for fashion brands interested in the promising market, especially in a pandemic or post-pandemic context.

Research limitations and implications
The present study was devised to explore the effects of Instagram users' COVID-19 perception on their perceptions of the usefulness and enjoyment of fashion brands' accounts on Instagram and how these perceptions are translated into satisfaction and behavioural intentions towards those accounts in a Sub-Saharan context. Also, this study sought to determine whether tech-nativity-based generational differences can moderate those relationships. We built upon a cross-sectional sample response to the measures of the variables under investigation, meaning all data were gathered at one point in time. Though cross-sectional research has been blamed for the limited ability for establishing causal relationships (e.g., Langdridge 2004), it also has been argued (Spector 2019) that the ability of longitudinally designed investigations to mirror causation has been embellished and that it only offers limited advantages over the cross-sectionally designed studies in most cases in which it is utilized (Spector 2019). Furthermore, cross-sectional research findings might still be interpretable and valid if conducted based on a robust theoretical premise (Tharenou, Donohue, and Cooper 2007;Mahmoud et al. 2020b). Yet, we endorse further replications with a longitudinal design employed. Besides that, utilising force answering (FA) to limit the amount of missing data may introduce bias into the forced responses. As a result, we propose that future studies consider alternatives to FA. For example, employing 'soft reminders' (Sischka et al. 2020) in conjunction with a 'Prefer not to answer' option (Mahmoud et al. 2021d). Though focusing our study on an understudied context (i.e., Sub-Saharan Africa) contributes to the novelty of our study; however, this makes our findings less generalisable to other contexts. Therefore, future research might consider replicating our study in other regions or culturally different contexts. Moreover, replicating our study in the Sub-Saharan context with more countries sampled would be warranted for future research. Relatedly, while limiting the present research to fashion brands was consistent with previous research employing the double enjoyment-usefulness approach as the industry's reliance on visuals is congruent with the visually rich nature of Instagram, future research might look to replicate our findings using social media accounts for brands in other industries that are decidedly less visual in nature -like financial services.
Our study was based on an appropriately sized sample; however, it could be deemed relatively small when investigating generational (or any other demographic) differences. With larger sample sizes, future research can investigate more complex multigroup comparisons. That would be, for instance, through generating groups based on crossing more than one demographic variable together (e.g., gender X generation).
Finally, we measured tech-nativity-based generational groups based on the classic categorisation of generational groups. While the literature backed our approach, we suggest that future research develops and employs multi-item (or even multi-dimensional) measures to assess tech-nativity. That would offer, especially if paired with age groups, a more realistic premise for generational categorisations based on tech-nativity.

Disclosure statement
No potential conflict of interest was reported by the author(s).