(Hedonic) Shopping Will Find a Way: The COVID-19 Pandemic and its Impact on Consumer Behavior

The issue currently permeating is how COVID-19 affects our lives, including in terms of consumer behavior. For example, sales of men’s suits have fallen sharply since March 2020, while there has been high demand for jogging pants. While German online retailing was able to increase sales by double digits in 2020, downtown retailers of non-food articles (e.g., textiles, shoes, etc.) had to accept a decrease of more than 20% (HDE 2021, p. 11). Our article focuses on the questions of whether consumer behavior has been fundamentally affected by the crisis, whether previously formed shopping patterns have dissipated and led to new shopping behavior, and whether old habits will return. Using two surveys at different timestamps of the pandemic, we analyze the impact on consumers’ shopping styles and particularly discuss whether the pandemic has permanently changed online shopping tendencies and ethical behavior, and whether the desire for experience-oriented shopping has changed.


Introduction and contribution
The issue currently permeating our society is how the COVID-19 pandemic affects our lives (Malter et al. 2020, p. 147).Donthu and Gustafsson (2020) draw attention to the economic consequences of the crisis and fear "severe economic consequences across the globe, and it does not look like any country will be unaffected.This not only has consequences for the economy; all of society is affected, which has led to dramatic changes in how businesses act and consumers behave."Although vaccines offer hope that the pandemic will end in the foreseeable future, the crisis is already raising many questions about how consumer life will be changed in the long term (not only in terms of economic variables, such as income effects).Will the home office persist in the long term, permanently changing product preferences (e.g., more emphasis on furniture, less emphasis on textiles)?Will fear of further pandemics trigger an urban exodus, with corresponding consequences for the real estate market?Will the online shopping boom that has been triggered continue?Will brick-and-mortar stores give up in large numbers and will city centers therefore become deserted?Will consumers continue to cook at home, enjoy it more and avoid restaurants even after the crisis?Will personal contacts, from personal sales talks to public viewing of sporting events, become less important?Before the pandemic, we would automatically ex-tend our hand in greeting.Since March 2020, we have painstakingly untrained ourselves from this automatism, and have learned to meet other people only at a distance.Will handshaking return after the pandemic or will we continue to keep our distance?Behind all these exemplary questions is the broader question of whether the pandemic will dissolve longlived habits and trigger new mindsets and readiness for change that will endure for longer.In other words, will we return to our old habits after the crisis, because habits reassert themselves after some time, or not?Habits can be so dominant that consumers even revert to old behaviors contrary to their actual attitudes (Montano and Kasprzyk 2015).Sheth (2020, p. 282) assumes, it "is expected that most habits will return back to normal.However, it is inevitable that some habits will die because the consumer under the lockdown condition has discovered an alternative that is more convenient, affordable, and accessible.Examples include streaming services such as Netflix and Disney.They are likely to switch consumers from going to movie theatres."However, he also assumes that "when an existing habit or a necessity is given up, it always comes back as a recreation or a hobby," and he wonders which habits will return after the pandemic.As mentioned, habits can be very persistent, especially if they are linked to typical human needs, such as the desire for human connection.But will this also be the case after the COVID-19 pandemic?At present one can only agree with the view of Malter et al. (2020, p. 147): "We are just beginning to see how the pandemic is affecting consumption during the crisis and can only take wild guesses as to what its long-term influence will be." This paper attempts to address some of the issues that focus on specific shopping behavior.The findings are based on two surveys conducted in Germany in May/ June 2020 and February/March 2021.The 2020 survey took place after the first lockdown in Germany, when all stores were open again and incidence numbers were low.The second survey took place after the second wave, when consumers could expect stores to reopen.At the beginning of March 2021, the third wave with the CO-VID-19 variant B.1.1.7 (formerly the so-called British variant, today Alpha variant) had already been announced by experts.But at the time of the second survey this variant was not yet showing up in high incidence figures (as observed in Germany during the second wave around Christmas 2020), which is why politicians of the federal states as well as the courts announced or demanded the reopening of stores.In very basic terms, we will discuss the following research questions: ) Can this drastic event permanently change consumption habits, such as shopping in brick-and-mortar stores vs. online shopping?
) Does the pandemic lead to a fundamental rethinking, e.g., more concern for our environment and thus more sustainable consumer behavior, or is everyone closest to themselves?
) How much does the pandemic change the desire for experience-oriented shopping?
) To what extent do expected income losses lead to changes in consumer behavior?

Critical life events
Consumers who are experiencing, or who have recently experienced, life-changing events (Wood 2010) subsequently change their consumer behavior.Wood (2010) finds that consumers in states of high (vs.low) life change are more likely to be attracted to unfamiliar products.This finding goes against the popular belief that consumers choose familiar options in times of upheaval as a way of coping with change.In fact, according to Wood (2010), life-changing events promote a change mindset, in which consumers avoid familiar products and go for new options instead.In a similar vein, Eelen et al. (2012) show that even minor changes, such as performing an unfamiliar task, trigger consumers' openness to new consumption experiences and motivate them to choose unfamiliar options.
Kamm and Groeppel-Klein (2015) also found in their empirical studies that consumers who were experiencing a high level of life change reported more positive attitudes towards unfamiliar brands with novel product attributes (were thus very open to innovative offers).By contrast, for less unfamiliar options (unfamiliar brands with known attributes or familiar brands with novel attributes) and familiar options (familiar brands with known attributes), consumers' perceived life change has no impact on their brand attitudes.
During the pandemic, brick-and-mortar (non-food) stores were temporarily closed, so consumers were forced to buy certain products online during those weeks.After the first lockdown (May/June, summer 2020), shopping in reopened stores was relatively safe, owing to hygiene regulations, masking, and low incidence figures in Germany.Nevertheless, the downtown retail sector got off to a sluggish start.Department stores, for example, continued to suffer sales losses in August 2020 compared with 2019 (Destatis 2020).By contrast, online retail continued to grow and generated an increase in sales of approximately 14,6 % throughout the year (BEVH 2021), raising the question of whether this trend will last.Roggeveen and Sethuraman (2020, p. 169), for instance, speculated that one reason could be that after the pandemic consumers may continue to associate brick-and-mortar shopping with hygiene measures and contact restrictions and therefore maintain newfound online shopping habits.However, it is debatable whether these associations are permanent or whether consumers will forget all the inconveniences as soon as the hygiene rules are no longer necessary.
Based on current evidence regarding the influence of critical events on consumer behavior, we hypothesize that the more consumers perceive the pandemic has radically changed their lives, the more they turn to online shopping (as a novel style; note: in 2019, as a comparison year prior to the pandemic, approximately 87 % of all purchases were made through the brick-and-mortar shopping channel), compared with those who perceive the crisis as not having changed their lives as much.This relationship can be explained by the subjective impression that the crisis leads to new and persisting behavioral patterns.A similar correlation can be expected for a shift towards more ethical consumer behavior (purchase of fair-trade or sustainable/organic products), and as an indicator for changed preferences.Alexa et al. (2021) found in their survey that during the lockdown participants indicated a greater interest in sustainable products.Since a lockdown for most people is a critical life event, this finding would support our hypothesis.
In study 1 and H1, we compare the values (percentage of total purchases accounted for by online non-food shopping) after the first lockdown in May/June 2020 for reopened brick-and-mortar stores with the values from before the COVID-19 crisis.In other words, the issue is not absolute spending, but the percentage of "more or less online shopping" triggered by the crisis, taking into account the fact that even before the crisis, of course, consumers had varying degrees of affinity for online shopping.In study 2, we asked consumers about their intended future online shopping behavior and their purchases of fair-trade, sustainable, or organic products.
Thus, the dependent variable in H1 and H2 is in each case a behavioral variable, the independent variable is the general assessment of whether individual life has changed, and the mediator is the feeling that the individual consumer behavior will also change as a result of the crisis.Fig. 1 visualizes the effects hypothesized in H1 and H2.
H1: The more consumers feel that their lives have changed, the more they will turn to online shopping compared to the past.This relationship is mediated by the perception that the pandemic will alter their shopping behavior.
H2: The more consumers feel that their lives have changed, the more they will turn to ethical shopping compared to the past.This relationship is mediated by the perception that the pandemic will alter their shopping behavior.

Shopping motives
The buzzword "experience consumption" was born in the last quarter of the 20th century (Holbrook and Hirschman 1982;Pine and Gilmore 1999).According to Pine and Gilmore (1999) experience-oriented consumers want to realize themselves emotionally, do not live primarily for the future but in the present, in which they want to express individuality.
The desire for emotional experiences may be genetically programmed, and thus there may have been a need for adventure or touching events throughout human history (Gerrig 2013).However, a shift of these needs into the consumer world could be observed, since, for saturated societies living in security, experiential consumerism can improve perceived quality of life (Gilovich et al. 2015).Experiential consumerism is perceived as the "salt in the soup."Stationary stores that place a lot of emphasis on the store atmosphere or are designed in an unusual or fascinating way, triggering affective reactions such as pleasure and arousal in the consumer, can help to satisfy the desire for an exciting life (according to Mehrabian and Russell's environmental psychological model, Mehrabian 1987).Babin et al. (1994, p. 646) differentiate between utilitarian and hedonic shopping: "hedonic shopping results more from fun and playfulness than from task completion" and "reflects shopping's potential entertainment and emotional value ... Increased arousal, heightened involvement, perceived freedom, fantasy fulfillment, and escapism may all indicate a hedonically valuable shopping experience."In contrast, utilitarian shopping is viewed as a task to be planned and accomplished efficiently.In principle, both shopping styles are relevant for online shopping (Childers et al. 2001) as well, but we focus here on brick-and-mortar-stores as these were hit by the lockdown.Arnould et al. (2002) explain more broadly that consumer experiences include diverse interactions such as (1) anticipated consumption, including daydreaming and fantasizing; (2) purchase experience, referring to service encounters and atmospherics; (3) consumption experiences, including sensory experiences; (4) remembered consumption."Purchase experience" strongly resembles Babin et al.'s (1994) hedonic shopping, therefore we equate the terms "experience oriented" and "hedonic shopping." Depending on the divergent motives, in-store marketing should contribute either to increased shopping efficiency or to the subjectively experienced pleasure (via browsing through the store, interacting with others, addressing all the senses).
A few empirical studies suggest that "experience-oriented shopping" is a more "lower-class phenomenon" (Allard et al. 2009, p. 47;Sit et al. 2003).This finding is justified by the assumption of these authors that people with higher incomes would have other stimulating leisure options available to them.To counter, the desire for shopping experiences seems to depend more on consumer preferences (and the need for emotional attachment): for some, shopping is a chore; for many others -irrespective of social class -it is enjoyable.Schmitt (1999, p. 3) elevates the phenomenon of "experiential marketing" to a revolution changing the face of marketing in affluent societies forever.
Assuming that non-food stores are open, however, one condition of hedonic shopping may not be present during the pandemic: the feeling of carefreeness.We thus propose that hedonic shopping (in brick-and-mortar-stores) is less relevant for worried consumers (irrespective of economic background), as browsing through stores and enjoying the atmosphere is only pleasant if it can be done without worry.Yang et al. (2020) are not concerned with hedonic vs. utilitarian shopping styles, but with preferences concerning either hedonistic or utilitarian products and whether these purchase decisions are influenced by the COVID-19 crisis.Utilitarian/hedonic products and utilitarian/hedonic shopping styles have to be differentiated.Utilitarian products are bought for instrumental purposes, while hedonic products are bought for reasons of entertainment ("Compared to hedonic products, utilitarian products are oriented to problem-solving," Yang et al., 2020).Hedonic (vs.utilitarian) shopping, on the other hand, assumes that regardless of the product type, people want to have fun and experience something exciting when shopping, while in utilitarian shopping, again regardless of product type, they try to proceed as planned and as efficiently as possible.Of course, it is conceivable that product type and shopping style reinforce one another.However, the focus of Yang et al.'s (2020) study is whether COVID-19 involvement (e.g., browsing for information on COVID-19 in the news or media, talking about the topic, COVID-19 being closely related to life) has an impact on product-type preference.The study, conducted in China, provides some interesting evidence for this investigation, as it shows that "COVID-19 involvement makes people more likely to purchase utilitarian products" (Yang et al., 2020, p. 6) and that this relationship is mediated by perceived awe and a subsequent more problem-focused coping (taking steps to eliminate the problems induced by COVID-19)."Awe" was here operationalized following Yaden et al. (2019) and describes a feeling of powerlessness and respect in the face of an incisive event.Awe al-so leads to a higher acceptance of following social (distance) norms (during the pandemic).
In Asian cultures, awe probably plays a greater role than in European ones; in the latter, the perceived threat of the virus is more likely to be seen as an explanatory variable.However, Koch et al. (2020) found in their study corresponding results to Yang et al.'s study (2020).They were mainly interested in whether Generation Z consumers regarded online shopping as a leisure activity during the pandemic in 2020, which turned out to be the case.But they also found (because online-shopping "reduces the risk of infection, as the activity involves no direct contact with other people") that "consequently, the perceived usefulness of online shopping is an important driver during a pandemic" (p.11).Thus, this result is also emphasizing a problem-focused, utilitarian aspect.In addition, they found that being exposed to media reports about the COVID-19 crisis (e.g., reports about the urgent need to support the local economy) had a significant impact on consumers' online purchase intentions.
The authors attribute to such media reports the authority to act as an external subjective norm, which has significant power.
In our study, as a normative variable, we chose the incidence figures published by the Robert Koch Institute (Germany's national Public Health Institute).Since each respondent was asked to provide his or her zip code at the end of the survey, we were able to determine the official incidence number of the county in which the respondent resided at the time of completing the questionnaire and include it as an additional variable.Since the incidence figures are published daily and regionally, consumers can use them to estimate the local incidence of infection (we will not discuss here how meaningful the incidence figures are but use the figure as an operationalization of an external norm of the threat by the virus).We assume that the official local incidence rate is a significant moderator (amplifier).
To summarize, on the one hand we hypothesize that utilitarian shopping will be more important for consumers who feel more affected by the pandemic.On the other hand, we also assume that hedonic shopping will return more strongly once consumers believe that the pandemic is under control, because for many consumers it is a vital need, as described above.Deng et al. (2020, p. 1), who also conducted a study in China, conclude that "consumers urgently need to restore the normal level of psychological arousal through the sensory stimulation brought by consumption."We agree and propose that especially those consumers who wish for their "old" (normal) prepandemic life back will express a particularly strong desire for experience orientation.
H3a: The more consumers are worried by the pandemic, the more the desire for hedonic shopping will be dampened, while utilitarian shopping will grow in importance (compared to the time before COVID-19).H4: The desire for (H4a) hedonic shopping will be significantly reduced by the pandemic (compared to the time before COVID-19) and will return with the feeling of the pandemic being under control; with (H4b) utilitarian shopping it is the other way around.
H4c: The wish for anticipated hedonic shopping after the pandemic will be stronger as more consumers want their old lives back.
Fig. 2 visualizes the effects hypothesized in H3a and H3b.

Income losses
It can be assumed that not all respondents feel equally affected by the pandemic.Those who fear income losses may have to change purchase behavior more than others, with an increasing trend to price orientation and less to quality orientation.If hedonic shopping were a basic need, as discussed above, it would not be affected.
H5: Anticipated income loss leads to increasing price orientation (H5a) and decreasing quality orientation (H5b) but does not affect consumers' desire for hedonic shopping (H5c).

Context and sample
Two online surveys were conducted to investigate the derived hypotheses.Both studies took place in Germany and the following temporal data are therefore linked to the respective pandemic situation there.
In Germany, the first wave of COVID infections led to a first lockdown in spring 2020 that forced all non-food shops to close temporarily.Additionally, grocery stores in Germany were not allowed to sell non-food articles during the first lockdown.Study 1 took place from May 21 to June 9, 2020, after the first lockdown when incidence numbers had decreased and all stores were open again.
The pandemic hit Germany in a second wave starting in October 2020.Study 2 took place from February 26 to March 7, 2021 and therefore after the second wave, when consumers expected stores to reopen.The second lockdown began in November 2020 ("lockdown light" with open stores) and was subsequently made more restrictive (non-food stores were closed) and extended several times.On February 10, 2021, after a corresponding political decision, there was hope that the second lockdown could be ended as of March 7.It was assumed that the incidence figures would be below 50.Therefore, the 2021 study was scheduled for the late February/early March period, when consumers could assume that all stores would reopen shortly.We call this period the projected end date of the second lockdown (later referred to as t 3 ).However, the situation changed again in early March (Imöhl und Ivanov 2021).On March 4, the lockdown was extended again until March 28 owing to rising incidence numbers.On March 15, 2021, it was announced that the federal government was also discussing an "emergency lockdown" at Easter, and although this "extended rest period at Easter," which was announced on March 22, was canceled two days later, most people voluntarily avoided unnecessary contacts.Another difference between the survey periods of study 1 (2020) and study 2 (2021) was that, in 2020, incidence figures were very low everywhere in Germany, whereas in 2021 figures varied widely across the federal states.In our survey period in 2021 the German federal states differed significantly in their incidence classes (χ 2 (32) = 2106.284,p < .001).In addition, independent of the incidence figures, the COVID-19 situation was assessed differently from a political point of view.For example, Saarland and Hamburg were in the same incidence class (values between 62 and 86), but at the end of March in Saarland, for example, the "Saarland model," with wide-ranging openings and testing, was discussed, which then came into effect on April 6, whereas in Hamburg a "hard-core" lockdown was decided at the end of March.We have described the survey situation in detail here, as we have to check whether the results were affected by this development (see section 3.2."Controls").
In summary, we can state that during the survey period there was still talk of opening stores.The third wave with the variant B.1.1.7 had already been announced by experts but was not yet visible (incidence figures less than 100).In contrast to the 2020 study, vaccination was already being carried out and the hope was that with the help of the vaccines and the change of season, the pandemic could be controlled by the summer.Therefore, at the end of the second survey we also asked for consumers' projections for the summer of 2021, with questions on how they would behave once the pandemic was under control.
In both studies, data collection took place online using Qualtrics as a data collection platform.In addition to the participants' answers to the questions, the official daily COVID-19 incidence rates (according to the Robert Koch Institute) of the participants' residential areas were obtained and included in the data set.
In study 1, participants were recruited via social networks.From the initial sample we excluded 8 participants who were younger than 18 years.We considered this age threshold relevant because of the questions about income losses during the pandemic.This led to a valid sample of n = 985 (77.9 % female, M age = 41.7; ± 14.99; from 18 to 83 years).Participants represented all age and education groups from all over Germany, and 31 % belonged to the "COVID-19-risk-group" (measured via self-assessment).Participants were first asked about their shopping behavior before the crisis.In the following analyses we refer to these retrospections as time 1 (t 1 ).
Then they were exposed to some media headlines about the outbreak of the virus in Germany and were asked about their sensitivities during the first lockdown.In the third part, again after some headlines about the fact that lockdown had ended, stores were reopened, and incidence figures were low, they were asked how the pandemic had changed their lives and about their current shopping styles after the first lockdown.In the following, we consider the data on purchasing behavior after the lockdown as time 2 (t 2 ).
In study 2, participants were recruited via social networks and with the support of a professional market research agency.Again, 2 participants younger than 18 years were excluded from the sample.The final sample contained n = 1,341 valid questionnaires (56 % female, M age = 47.2; ± 15.87; from 18 to 88 years; 37 % "COVID-19-risk-group").Once again, all age and income groups were represented.We again asked consumers about changes in their life and consumer behavior.The questions were repeated from study 1 to enable a comparison.
In addition, further topics related to consumers' shopping behavior were included (e.g., ethical consumer behavior only in study 2), to cover more topics overall.Participants answered questions on their "current" perceptions shortly before the projected end of the second lockdown, which we refer to as time 3 (t 3 ).Subsequently, participants had to anticipate their consumer behavior for a time when the pandemic will likely be under control via vaccinations.In the following, we will refer to this prospective assessment as time 4 (t 4 ).
The two studies include measures of consumers' perceptions and behaviors over a longer time period, including a time point before the pandemic (t 1 , measured in retrospect), two time points within the pandemic (t 2 = after the first lockdown, 2020; t 3 = before the projected end of the second lockdown, 2021), and a "future" time (t 4 ) for which it is assumed that the pandemic is under control.

Measurement and control variables
The studies, hypotheses and results reported in the present paper were part of a larger project on the impact of the COVID-19 pandemic on consumer behavior that also addressed constructs other than those reported here.The project considered a variety of constructs in order to collect the most comprehensive data possible during the critical time windows of each pandemic phase.To make this possible (and the online interview not too long), the scope of the measured variables had to be reduced for individual constructs, and in some cases single items (Bergkvist and Rossiter 2007) were used instead of a complete scale.The following constructs will be relevant here: perceived life change, perceived change of consumption behavior, change in online shopping behavior, change in ethical consumption behavior (H1, H2), feeling of worry by the pandemic, incidence figures as proxy for external norms of the threat (H3), change of hedonic and utilitarian shopping motives (H3, H4, H5), price orientation, quality orientation (H5), desire to return to the old life (H4), income loss (H5).We will also report some further exploratory results from the larger project in the 4 th section.
Unless otherwise specified, all measurements were made on a 7-point Likert scale (ranging from 1 to 7, "strongly disagree" to "strongly agree").

Perceptions and feelings during the pandemic:
The measurement of perceived life change was conducted using a single item taken from the study of Wood (2010, p. 954) that addressed the respective period of time that we were interested in for the particular study: "I've made a lot of changes in the past two months" (study 1) and "I've made a lot of changes since Christmas" (study 2).Both periods of time refer to a situation of lockdown owing to the pandemic.Extent of worry was measured in study 1 only.Three items from the consumption emotion descriptor set (CES) by Richins (1997, p. 134) were used: "worried," "concerned," and "frightened," and participants were asked to indicate the extent to which they had felt them during the last two months.Because internal consistency reliability was satisfying (Cronbach's α = .729),items were averaged (arithmetic average) and a composite score "worry" was used.Anticipated income losses were measured with a single item: "I am afraid that my income will be reduced owing to the pandemic."Desire to have one's old life back was also measured with a single item: "I miss my former life very much." Change in consumption behavior: In both studies, change in general consumption behavior was measured by the single item: "My consumer behavior will alter permanently as a result of the pandemic" (similar to Baicu et al. 2020).
Change in the intensity of online shopping of non-food products was measured differently in the studies.In study 1, online shopping intensity was measured before (in retrospect) and after the lockdown (at the time of the study) using a 100-percent scale ("What percentage of your non-food purchases (e.g., clothing) did you/do you make online?").We used a difference score as the dependent variable for H1 (after minus before, t 2 minus t 1 ).Therefore, positive (negative) values on the difference score refer to an increase (decrease) in online shopping after the lockdown.Note that, in study 1, this measurement was conducted when "offline" shopping for nonfood articles was available again.In study 2, we were interested in how consumers assessed their prospective behavior after the second wave.Here, participants were requested to give their projection using the single item: "For me, shopping on the internet has become a matter of course, which I will continue to do frequently in the future." Change in ethical purchasing behavior was only considered in study 2. It was measured using four items adapted from Alexa et al. (2021).Participants indicated on a 7-point scale whether they buy "considerably less" (= 1) to "considerably more" (= 7) regional, organic, fair-trade and sustainable products in a comparison of the time shortly before the projected end of the second lockdown (t 3 ) to the situation before the pandemic (t 1 ).We averaged the four items to form a composite score (α = .839).
(Change in) shopping motives: Hedonic and utilitarian shopping motives were measured with four items each according to Babin et al.'s (1994) scale, adapted to European conditions by Groeppel-Klein et al. (1999).The questions were adjusted according to the respective periods of time: We asked for hedonic (utilitarian) shopping desires that participants had before the pandemic (t 1 , retrospective, study 1), that they have currently (= after the first lockdown 2020, t 2 , study 1; shortly before the projected end of the second lockdown = t 3 , study 2), and which they expect to have in summer 2021 assuming a control of the pandemic (t 4 , projection, study 2).Factor analyses were conducted using principal component analysis and varimax rotation considering the different time periods separately.As supposed, a two-factor solution was yielded for each time period.In each case, the KMO was & .779,all MSA values were & .676,and the particular two-factor solution explained & 59.11 % of the variance.All factor loadings were & .626.As a result, for each timestamp, the items measuring hedonic shopping loaded on one factor (illustrated here for t 2 ): (1) "Shopping is currently an experiential leisure activity for me that I generally enjoy and indulge in simply for the fun of it;" (2) "I currently love the variety of stimuli (people, products, decorations, scents, background music, etc.) when shopping;" (3) "I currently want to experience something new and out of the ordinary when shopping;" (4) "I think it is important for me to currently talk or exchange ideas with others when shopping."Items measuring utilitarian shopping loaded on another factor (again illustrated here for t 2 ): (1) "When shopping, I currently buy only the particular items that I really need;" (2) "Shopping is currently purely functional and a means to an end for me;" (3) "Shopping is currently a chore that I usually want to get done as soon as possible;" (4) "Before shopping, I currently make a shopping list that I then work through" loaded on a factor of "utilitarianism."The Cronbach's alpha values were as follows: α hedo_t1 = .769;α util_t1 = .693;α hedo_t2 = .731;α util_t2 = .750;α hedo_t3 = .823;α util_t3 = .666;α hedo_t4 = .850;α util_t4 = .736.Therefore, the respective hedonic items and the utilitarian items were averaged to form composite scores.
Price and quality orientation were measured in both studies.Items referred to the time when the respective study took place: t 2 in study 1 and t 3 in study 2. Price orientation was measured with two items: "I am above all interested in low-budget offers," and, "I am searching for a special price offer," according to Groeppel-Klein et al. (1999).Quality orientation was assessed by the two items: "It is important to me to buy high-quality products," and, "I will not give up high quality for a lower price" (Ailawadi et al. 2001, p. 87).Factor analyses confirmed the supposed two-factor solution.In each case, the KMO was & .556,all MSA values were & .516,and the two-factor solution explained & 80.35 % of the variance.The Cronbach's alpha values were as follows: α price_t2 = .726;α quality_t2 = .863;α price_t3 = .757;α quality_t3 = .780.Based on satisfying results of reliability analyses, the respective items of the different factors were averaged to form composite scores.

External norm of the threat of the virus:
The incidence figure (published by the Robert Koch Institute) identifiable by the participants' zip code areas on the day of questionnaire completion was obtained for each case and included as an additional variable.

Controls:
We also considered several control variables that have been found to be important in the context, such as age, education, income, and personality, such as routine seeking, emotional stability, and need for security (measured according to Costa and McCrae 1992;Rammstedt and John 2005).None of the variables correlated significantly and higher than r = .20 with the dependent variables (Spilski et al. 2018) of our hypotheses H1-H4, neither in study 1 nor in study 2. Therefore, these variables did not need to be included as covariates in the analyses (Meyvis and Van Osselaer 2018).Important for hypothesis H5 (study 1 and 2), we found that income is significantly correlated with the dependent variable quality orientation in both studies (study 1: r = .233,p < .01;study 2: r = .234,p < .01).Therefore, income was used as a covariate when analyzing H5.
We have already pointed out that it is necessary to control for any bias that may have arisen during the study 2 survey period owing to the changing COVID-19 situation and diversity across the federal states.We found (not surprisingly) a weak but significant correlation of progressing survey days and official incidence scores in respondents' federal states in this survey period (r = .112,p < .001).However, there was no significant (p = .213)negative correlation of progressing survey days (and thus possibly an increasing number of media reports on a renewed escalation of the COVID-19 situation over the course of time) to the statement, "I am optimistic about the future," so we did not detect a change in consumer sentiment over our second survey period.In contrast, the correlation (r = .371,p < .001) between optimism and the belief that we will get a handle on the pandemic through vaccines is significant and high.With the exception of Bremen (with only six respondents), optimism was not significantly different in all federal states (M Germany = 4.64, SD = 1.734), as well as the hope that the vaccine will defeat COVID-19 (M Germany = 4.17, SD = 1.601).
In summary, we therefore estimate potential bias from advancing COVID-19 cases to be low during our second survey period.

Changes in online shopping behavior H1 (study 1):
In study 1, we tested for the effect of life change on differences in online shopping behavior and considered a comparison of pre-pandemic times (t 1 ) and the period after the first lockdown in 2020 (t 2 ).The change in online shopping behavior was considered a difference score.Therefore, positive (negative) values on the difference score refer to an increase (decrease) in online shopping after the lockdown.Consumers' perceived changes in their consumer behavior served as a mediator.We used the PROCESS v3.5 macro for SPSS provided by Hayes (2018) and applied model 4 (5,000 bootstraps).
We found a significant positive effect of life change on perceived change in consumer behavior (b = 0.134, p < .001).Perceived change in consumer behavior, then, significantly increased online shopping behavior (b = 4.861, p < .001).A significant indirect effect (b = 0.651, bootstrap 95 % CI [0.403; 0.941]) indicates that the more consumers perceive general life changes, the more they perceive altered consumption behavior, leading to significantly more online purchases.These results support our hypothesis H1.The residual direct effect was also significant (b = 1.751, p < .001),which leads to the assumption of further mediators in this relationship.

H1 (study 2):
In study 2, the dependent variable was a projection regarding changes in online shopping activity in the future.Again, we assumed perceived alteration of consumption behavior to serve as a mediator.We found a significant relationship between perceived life change and perceived change in consumption behavior (b = 0.213, p < .001),which in turn influenced future online shopping preferences positively (b = 0.146, p < .001).Again, the indirect effect was significant, indicated by a bootstrapping confidence interval that excluded zero (b = 0.031, bootstrap 95 % CI [0.017; 0.046]).This supports H1 also for study 2. The residual direct effect was not significant (see Tab. 1).

Changes in ethical shopping behavior H2 (study 2):
We tested for the effect of life change on modifications to ethical shopping behavior and considered a comparison of pre-COVID-19 times (t 1 ) and the period shortly before the projected end of the second lockdown in 2021 (t 3 ).We found a significant effect of life change on perceived alteration of consumption behavior (b = 0.213, p < .001),which in turn positively influenced ethical consumption (b = 0.110, p < .001).The indirect effect was significant (b = 0.023, bootstrap 95 % CI [0.015; 0.033]).These results support hypothesis H2 (see Tab. 2).However, the residual direct effect was also significant, indicating other possible mediators.

Changes in shopping motives
The impact of worries (H3, study1): H3a proposed an impact of subjective worries on a change in shopping motives.We assumed that higher perceived worries would decrease hedonic shopping and increase utilitarian shopping motives, when considering a comparison of the situation after the first lockdown (t 2 ) with the pre-CO-VID-19 situation (t 1 ).We used the composite scores for hedonic and utilitarian shopping and formed two difference scores (t 2 minus t 1 ).Positive (negative) differences refer to increased (decreased) hedonic and utilitarian shopping motives, respectively.In addition, H3b proposed that the impact of subjective worries about the pandemic is reinforced for consumers living in areas with higher incidence figures (official figures as an external norm).We tested for H3 by using PROCESS v3.5 and applied model 1 (Hayes, 2018), which considers moderation analyses.Subjective worries served as the independent variable; the particular difference score (hedonic, utilitarian shopping) served as the dependent variable.The moderator was the official incidence figure obtained for the residential area of the participant.The variables were mean centered in order to be able to interpret main effects, too (Hayes 2018, p. 310).We estimated two moderation models (see Tab. support our hypothesis H3a.Furthermore, higher official incidence figures had neither a significant main effect on decreased hedonic shopping nor on utilitarian shopping motives.In addition, the incidence figures did not moderate the impact of perceived worries on shopping motives, neither for the hedonic nor for the utilitarian shopping motives.Thus, H3a is supported, but not H3b (see Tab. 3).

Development of shopping motives over time (H4, both studies):
Our two studies span four timestamps: t 1 = a retrospective to before the pandemic, t 2 = after the first lockdown (both measured in study 1), t 3 = shortly before the projected end of the second lockdown, t 4 = projection to when the pandemic would be under control (both measured in study 2).H4a assumes that hedonic shopping motives would decrease during the pandemic (compared to the pre-COVID-19 time) and would increase in consumers' estimation of their future shopping behavior.H4b conjectures the effect in the opposite direction for utilitarian motives.We first calculated differences within the two studies using t-tests for dependent samples.For the data in study 1 (t 1 , t 2 ), the results showed a significant decrease in hedonic shopping motives during the pandemic (M hedo_t1 = 3.36; SD = 1.39;M hedo_t2 = 2.03; SD = 1.08, t( 984 These results show that our hypothesis can be confirmed: Hedonic shopping motives decreased whereas utilitarian motives increased significantly.However, data analyses of the second survey showed only very small effects.We assume that with the advent of the vaccine and decreasing numbers of COVID-19 patients in intensive care units, the desire for experiential shopping has returned already in t 3 , and this desire intensifies once again with the prediction that the pandemic will be under control in t 4 .In other words, the results show in consumers a certain longing for hedonism when the first shock of the pandemic has been digested and hope for improvement is in sight.To test this additional assumption and to compare the different timestamps across studies, we combined the composite scores of the two samples in a new SPSS file and standardized them across all four time points (Fig. 4).An ANOVA across those four time points led to a significant overall difference for hedonic shopping (F(3, 4646) = 355.641,p < .001,partial η 2 = 0.187) as well as for utilitarian shopping (F(3, 4646) = 181.666,p < .001,partial η 2 = 0.105).The relevant contrast relates to a comparison of t 2 and t 4 and showed that hedonic shopping experienced a significant recovery from the time after the first lockdown to the projected time when the pandemic might be under control (t(4646) = 27.95,p < .001,Cohen's d = 1.25), while utilitarian shopping motives significantly decreased during this period (t(4646) = -19.20,p < .001,Cohen's d = 0.80).
Impact of desire to have one's old life back (H4c, study 2): We also found support for H4c, with a significant correlation between anticipated hedonic shopping after the pandemic and desire to have one's old life back (r = .343;p < .001).

Impact of income losses (H5, both studies):
We assumed that hedonic consumption as experience orientation is a basic need regardless of expected income losses, while price orientation will (forcibly) increase and the desire for high quality will decrease among those affected by income losses.In both studies we asked consumers whether they expect income losses owing to the crisis.The variable was z-standardized and two groups of consumers were formed: those with income loss expectations below the mean (further referred to as "no loss") and above the mean ("loss").We calculated ANOVAs comparing these two groups in terms of the particular dependent variables (hedonic consumption, price orientation), that were also z-standardized (Tab.4).For the dependent variable "quality orientation," an ANCOVA with net income as the covariate was used, as explained in section 3.2.
The results show support for H5a, H5b, and H5c in study 1, with significant differences (all ps < .001)for price and quality orientation but not for hedonic shopping/ex-

Additional results
As an indication that the questionnaires were filled out seriously and thus valid, the percentage of online purchase growth can also be seen.In study 1, the respondents indicated an average of about 13 % more online purchases than before the pandemic.This coincides with the official figures of the E-Commerce Association (BEVH 2021), indicating an increase of 14 % for 2020.
We also measured whether consumers "perceive online shopping as time-saving and convenient," according to Makhitha (2014).Over time, there is a significant increase in the importance of this item from M 2020 = 4.52 to M 2021 = 5.30 (measured on a 7-point rating scale).Similarly, we observed a significant and sharp increase in price orientation from M 2020 = 4.11 to M 2021 = 4.72, which shows that many consumers are beginning to fear possible negative economic consequences of the crisis.
In both studies, we asked consumers how much they felt harmed by COVID-19, all things considered.Here, too, we saw an increase in the mean value from 3.50 (2020) to 3.99 (2021).Interestingly, in 2020, the self-employed had the highest scores, with a plus of 0.51 scale points from the mean value while in 2021, schoolchildren (older than 18) and students reported the highest score, with a plus of 0.61 scale points from the mean.Schoolchildren and students currently feel the most aggrieved.Of all occupational groups, however, the self-employed are currently most concerned that they will have to accept a loss of income, with a plus of 0.37 from the mean 3.61.
The perceived threat of contracting COVID-19 was measured with two items based on previous studies (Bae and Chang 2021;Dryhurst et al. 2020), also slightly adjusted.In relation to these questions "I believe there is still a very high probability that I will contract COVID-19 or its variants" and "I think I have an increased risk of contracting COVID-19 (compared to other people my age)", respondents in 2021 gave a significantly higher value (composite score across both items) of M 2021 = 3.85 (SD = 1.61) than in 2020 with M 2020 = 3.09 (SD = 1.15).
It is also interesting to note that in study 2 there was no significant relationship between official incidence rates and shopping styles (similar to study 1) and incidence rates and subjectively perceived threat from the virus.The simple mentioning of these figures seems not to impress consumers.However, as the assessment of the threat posed by the virus has increased from 2020 to 2021, personal experience (sufferers in the family, in the (extended) circle of friends, or media reports) must presumably have played a role here.
At both survey time points, most respondents indicated that they found the politically imposed COVID-19 regulations appropriate.On a 7-point scale of "not strict enough" (= 1) to "just right" (= 4) to "excessive" (= 7), the 2020 mean value was 4.15 (SD = 1.63) and the 2021 means value was 4.09 (SD = 1.83).Likewise, we found very high agreement scores (scale values 5, 6, and 7 on a 7-point scale) to the questions: -"I miss my old life," with 62.7 % -"I will only visit stores that strictly follow hygiene rules and offer a lot of distance from other customers," with 55.3 % -"I will only visit stores with certified hygiene seals", with 48.4 % -"In the future, I will continue to avoid crowds in order to protect myself, and I will only visit stores in a targeted manner," with 47.4 % In addition, the mask has gained in acceptance.While in 2020 the majority of consumers still considered the mask to be "hostile to shopping", this changed in 2021: "shopping with a mouth guard is not a pleasant experience and is no fun for me", 2020: 60.0 % high agreement (M = 4.89); 2021: 42.7 % high agreement (M = 4.23), and one third of the participants also said they would not want to do shopping without the mask in the near future (regardless of incidence figures).These findings should be considered by retailers.Incidentally, agreement to continue wearing masks is not significantly correlated with sociodemographic variables but is correlated with the perception of a higher own risk of infection (r = .354,p < .001)and with the hope that the pandemic can be overcome with vaccines (r = .138,p < .001).The latter result probably shows that these consumers do not want to take any more risks shortly before vaccination and want to protect themselves with the mask until then.
Most people are aware that COVID-19 is extremely dangerous, but at the same time (even though the values are far below the mean scale value) the agreement with the questions, "I believe that the SARS-CoV-2 virus is to be assessed like a flu virus, which is why I have little to worry about," and, "I believe that the danger from CO-VID-19 is greatly overestimated," also rose from (composite score) M 2020 = 2.24 to M 2021 = 2.98.

Discussion, limitations, implications, and further research
The findings show that the pandemic is having a major impact on consumer behavior.New shopping patterns are emerging and the growth in online shopping is likely to continue for some time.Our results show that perceived life change can serve as an explanation for the increase in online shopping behavior, even when projections on future shopping behavior are considered.The pandemic is also an opportunity for consumers to demonstrate more ethical buying behavior.However, behavior will not change in all areas.People are social beings, wanting to interact and experience with all their senses.The desire for hedonic shopping/experiential shopping in brick-and-mortar-stores seems to be returning, provided we gain control of the pandemic, perhaps more strongly than ever.Interestingly, in study 2 we also found a significant negative correlation between anticipated hedonic motives and increased online shopping, as well as a significant positive correlation between utilitarian motives and online activities after the pandemic, leading us to assume that the internet will primarily be used for utilitarian shopping.
Limitations are, inter alia, that we only measured remembered shopping behavior (study 1) and respectively anticipated behavior for a post-pandemic period (study 2).We will rerun the survey when the virus is hopefully defeated.Furthermore, our mediation analyses included significant direct effects that indicate the existence of further mediators that should be addressed by future studies.
Implications: Retailers should be aware that -even if the desire for experiential orientation returns -the experiential marketing game has to be played by new rules.Customers will initially remain cautious, avoiding crowds, and are more likely to honor stores that offer high safety through hygiene standards and appropriate controls.However, this does not mean that customers cannot be inspired by fascinating store designs; this need is back.
The desire for interpersonal communication (as part of experiential shopping) also increases, and secure framework conditions must be created for this exchange.The agreement with the single question "Currently (When we have the pandemic under control) I (will) find it important to be able to talk or exchange ideas with others (salespeople, other customers) while doing my shopping" rose from M = 2.31 (SD = 1.703) in 2020 (t 2 ) to M = 4.05 (SD = 1.854) in 2021 (t 4 ).
In addition, it can be assumed that price competition in inner-city non-food retailing will increase (also because of full stocks), and that consumers will become more price sensitive.Stationary retailers must be aware that the online channel is gaining in importance; this realization has certainly matured among an overwhelming number of retailers.The question is, which channels/platforms should be used for online purchasing in the future?In 2021, our respondents said they would make 45 % of their expenditures via regional platforms (such as "Saarland-online", "OWL-online" or "Cologne-online") if all local retailers were listed there.This could also be a recommendation on how to strengthen local commerce.