Growing your own in times of crisis: the role of home food growing in perceived food insecurity and well-being during the early COVID-19 lockdown

Household food insecurity and poor well-being have increased during the coronavirus disease 2019 (COVID-19) pandemic and resulting lockdown measures. Home food growing has been associated with improved food access and well-being, but it is unknown what role it plays during food supply crises and lockdown. It is also unclear how home food growing and social restrictions may affect opinions about growing food in urban areas (i.e., urban agriculture [UA]). A cross-sectional online survey was conducted during the UK national lockdown in March-April 2020 to measure home food growing, perceived food insecurity, well-being, and opinions of UA. The participants were 477 UK-based adults (369 female, mean age 39.57 years ± 13.36); 152 participants were engaged in home food growing prior to the pandemic. Responses were compared to data collected from a separate sample of participants before the pandemic ( N = 583) to explore potential shifts in opinions about UA. Participants who engaged in home food growing had lower levels of food insecurity ( U = 19894.50, z = -3.649, p<.001, r = -.167) and higher well-being ( U = 19566.50, z = -3.666, p<.001, r = -.168) than those not engaged in home food growing. Perceived food insecurity partially mediated the relationship between home food growing and well-being; home food growing was associated with less food insecurity, which in turn was associated with better well-being. There were no differences in opinions of UA compared to the sample of participants from before the pandemic. Home food growing may have had a protective effect over perceived food security and well-being in the early stages the pandemic. Opinions of UA were positive and unchanged compared to data collected pre-pandemic. Policies that support home food growing and access to suitable growing spaces and resources may be beneficial for food system resilience and well-being.

A cross-sectional online survey was conducted during the UK national lockdown in March-April 2020 to measure home food growing, perceived food insecurity, well-being, and opinions of UA. The participants were 477 UK-based adults (369 female, mean age 39.57 years ± 13.36); 152 participants were engaged in home food growing prior to the pandemic. Responses were compared to data collected from a separate sample of participants before the pandemic (N = 583) to explore potential shifts in opinions about UA.
Participants who engaged in home food growing had lower levels of food insecurity (U = 19894.50, z = -3.649, p<.001, r = -.167) and higher well-being (U = 19566.50, z = -3.666, p<.001, r = -.168) than those not engaged in home food growing. Perceived food insecurity partially mediated the relationship between home food growing and wellbeing; home food growing was associated with less food insecurity, which in turn was associated with better well-being. There were no differences in opinions of UA compared to the sample of participants from before the pandemic.

Reviewer Status
Invited Reviewers

Introduction
The global coronavirus disease 2019 (COVID-19) pandemic triggered widespread lockdown restrictions to limit the movement of citizens and control the spread of the virus (Koo et al., 2020). These are predicted to have negative and long-lasting consequences for mental health (Daly et al., 2020;Marroquín et al., 2020;Pierce et al., 2020;Rajkumar, 2020;Sani et al., 2020;Torales et al., 2020;Zhang & Ma, 2020). Evidence from the early stages of the pandemic indicates that the incidence of mental ill health and poor well-being have increased (Daly et al., 2020;Pierce et al., 2020). Declines in well-being are likely to be compounded by pandemic-related isolation, fear, and existing health vulnerabilities experienced by some individuals (Torales et al., 2020). Estimates suggest that up to 20% of the population in the United Kingdom (UK) will require mental health support as a result of this pandemic (O'Shea, 2020). Therefore, as the pandemic continues, there is an urgent need to improve understanding of its impact on mental health and related issues.
Lockdown restrictions to slow the spread of the virus in the UK began in late March 2020. These included the closure of schools, non-essential retail and many businesses, limitations to citizens' movements, shielding of vulnerable individuals and requirements for social distancing (Brown, 2020). During the same time period, media reports highlighted disruptions to food supplies and food availability in supermarkets and shops, with purchasing restrictions placed on some items. Shoppers stockpiling food created surges in demand (O'Connell et al., 2020). This rendered many foods unavailable, increasing reliance on highly-processed, often less-healthy food options (Tan et al., 2020). Demand surges highlighted vulnerabilities in the UK food supply chain due to an over-reliance on a small number of EU countries for food imports and a reduced supplier base for supermarkets (Garnett et al., 2020). Emerging evidence suggests that levels of food insecurity quadrupled during the same time period (Loopstra, 2020). Food insecurity refers to limited access to safe, appropriate and nutritious food (Taylor & Loopstra, 2016), and is robustly associated with poor mental health (Frongillo et al., 2019;Jones, 2017;Pourmotabbed et al., 2020). In addition to food insecurity resulting from a lack of buying power, the lack of food in shops and lockdown restrictions on movement, which prevent people from obtaining food, appear to be driving new dimensions of food security (Loopstra, 2020). Lack of food in shops, in particular, may explain up to 40% of food insecurity experiences during this time (Loopstra, 2020). Thus, as intermittent tightening and relaxing of lockdown restrictions continues, there is an urgent need to understand potential factors that may mitigate the consequences of the pandemic on food insecurity and mental health.
Urban agriculture (UA), the growing of fruits and vegetables in urban and suburban areas, has historically been relied upon to bolster food supplies during times of crisis, such as war and food system shocks (Mok et al., 2014). For example, growing food at home in "Victory Gardens" was encouraged to help buffer against food shortages during the Second World War (Mok et al., 2014). Therefore, home food growing in gardens and allotments has been highlighted as a potential means of providing access to nutritious, healthy food in urban areas during the COVID-19 pandemic (Lal, 2020). Existing literature from before the pandemic indicated that home food growing can supplement household food supplies and reduce food insecurity (Algert et al., 2016;Galhena et al., 2013;Kortright & Wakefield, 2011). In addition to this, engagement in urban food growing (in general) has been associated with mental health benefits (Audate et al., 2019;Kingsley et al., 2009;Lovell et al., 2014). Home food growing in particular has been associated with improved well-being (Dobson et al., 2020;Genter et al., 2015;Soga et al., 2017a) and reduced stress (Palar et al., 2019;Van Den Berg & Custers, 2011).
Taken together, existing evidence implies that home food growing could be beneficial in mitigating some of these negative impacts of the COVID-19 pandemic. However, to our knowledge, this has not been previously investigated in academic research. This issue is timely, as reports by the UK media and other global outlets during the pandemic indicate a surge in interest in home food growing. This has included reported seed shortages in garden centres, a rise in people taking up home food growing, and large increases in applications for allotment gardens. Home food growing could therefore represent a healthy, sustainable and accessible strategy for coping with the food insecurity and well-being impacts of the COVID-19 pandemic. Thus, the primary aim of the current study was to explore relationships between home food growing, food insecurity and well-being during the early stages of the COVID-19 pandemic. The proposed relationships are illustrated in Figure 1.
Whilst UA has been proposed as a viable means of reducing food insecurity (Edmondson et al., 2019), less is known about how it is viewed by the general public. Public opinion is an important consideration in promoting the uptake of urbangrown food, particularly in household settings. Some studies have highlighted concerns about "unnaturalness" of food produced via UA (Specht et al., 2016a), and contamination from urban pollutants (Kim et al., 2014;Lal, 2020). However, increased interest in home food growing in the UK, albeit

Amendments from Version 1
Our thanks go to the reviewers for their helpful and thoughtful reviews of this manuscript. In addressing their comments and suggestions we have added more detail about participant recruitment, expanded discussions of well-being perceptions and policy implications of the findings, and focused the discussion of these results on the UK context. We have also addressed and discussed an additional limitation to the study that we are unable to confirm if participants were based in urban or rural settings, and we have condensed the Analysis section.
Any further responses from the reviewers can be found at the end of the article anecdotal, may also be accompanied by a shift in opinions of UA. A small, recent study of public opinions of urban farming in North America reported more positive than negative opinions (Grebitus et al., 2020). A secondary aim of the current study was therefore to explore changes in opinions of UA by comparing data collected at the start of the UK lockdown with data on opinions of UA from a separate sample of participants in a study conducted before the pandemic (the pre-pandemic study; Mead et al., 2021b).
We predicted that people engaged in home food growing would report less perceived food insecurity (Hypothesis 1) and better well-being (Hypothesis 2) than those not engaged in home food growing. We also predicted that individuals with higher levels of perceived food insecurity would be more likely to report intentions to engage in home food growing (Hypothesis 3). Finally, we tested a non-directional hypothesis (Hypothesis 4) that participants in the current study would show differences in opinions of UA compared to participants in the pre-pandemic sample (Mead et al., 2021b).

Participants
Participants were recruited via online adverts shared on social media and in online community groups and newsletters. We targeted social media/online groups and newsletters with either a food growing or a general focus (for example, general interest, UK-specific, or "help needed" sections of social media sites, groups and message boards). Recruitment was targeted in this way to ensure we recruited a mixture of participants who were and were not engaged in home food growing. Study adverts described the study as being about "food, well-being and the coronavirus outbreak". Individuals who were eligible to participate were aged 18 years or older and based in the UK. A sample size calculation based on data from our pre-pandemic study (Mead et al., 2021b) indicated that a sample size of 428 was required to detect small-medium effect with 95% power. Given the unprecedented nature of the pandemic, and the lack of previous research to inform a sample size calculation, we based this on a nondirectional analysis for detecting a difference in opinions of UA between the current study and the pre-pandemic study. Output of the sample size calculation is shown in the Extended data (Mead, 2021).

Measures
Demographics. Participants reported their age (years), gender, which UK country they were currently residing in, and their height and weight. Well-being. The World Health Organisation (WHO)-5 index was used to measure well-being (Topp et al., 2015). Participants responded to five items that assessed subjective well-being over the previous 2 weeks (e.g., "I have felt cheerful and in good spirits"). Responses were scored 0 ("At no time") to 5 ("All of the time"), summed and multiplied by four to give a total score out of 100. Higher scores indicate better well-being. Cronbach α value for the measure was .850. An attention check (please select "All of the time") was inserted into this questionnaire to monitor for random responding.

Experience of food insecurity.
A modified version of the eight-item Food Insecurity Experience Scale (FIES; Smith et al., 2017) was used to assess perceived experiences of food insecurity. Items were adapted to assess the prevalence of aspects of food insecurity over the previous 7 days due to lack of food available in shops (e.g., "During the last 7 days was there a time you were worried you would run out of food because of a lack of food in the shops?") as opposed to the typically assessed time period of the previous 12 months, and experiences related to lack of money (Smith et al., 2017). The modified scale is shown in the Extended data (Mead, 2021). This modification was designed to capture experiences specific to the early lockdown phase, when media reports suggested food shortages at UK shops and supermarkets. Responses were "Yes" or "No". Scores were determined by summing the number of "Yes" responses. Higher scores indicated greater perceived experience of food insecurity. Cronbach α value for the modified scale was .748.

Home food growing.
We used the following question to determine if participants were engaged in home food growing: "Is growing your own fruits and vegetables something you…". Participants selected one of the following response options: i) already do, ii) don't currently do but are planning to do, iii) are not planning to do but would consider in the future, and iv) are not planning to and would not consider doing. Participants who responded "i) already do" were categorised as engaged in home food growing (Growers); other responses were categorised as Non-growers.

COVID-19 related questions.
Participants provided the following information in relation to the UK lockdown: keyworker status 1 (yes/no), isolation status (social distancing/ self-isolating/ lockdown/working out of home but social distancing/life as normal/other). They also indicated if they were in a high-risk category for coronavirus, and if they or someonein their household had experienced symptoms of, or been diagnosed with, coronavirus.
Additional questions regarding participants' concerns about accessing food were asked to provide more detail about the types of foods that they were concerned about accessing, if any. Participants reported if they had concerns about accessing food (yes/no). Next, they indicated if they had concerns about accessing specific types of food (fruit and vegetables/ meat/dairy products/dried foods/bread/baby or infant food/pet food/other). Finally, participants were asked to indicate which extra measures they may be taking to ensure adequate food supply, if any (buying extra food/sharing food with others/monitoring or reducing the food I buy/monitoring or reducing the food I eat/growing my own fruits and vegetables/eating fruits and vegetables somebody else grows/sourcing food elsewhere/other).

UA opinion questions.
Participants read the following definition of UA: "Urban agriculture means growing fruit and vegetables in urban, suburban and surrounding areas." This was followed by the question "Do you think urban agriculture could be beneficial/not beneficial to any of the following?" and the following items: i) You, your family and friends; ii) Your community; iii) The environment; iv) The economy; v) Society in general; vi) Entertainment/alleviate boredom; vii) For education/home-schooling; viii) Alternative income because of unemployment; ix) Ensuring my family and I can eat healthily; x) Ensuring social distancing by accessing food without being in contact with others. Response options were beneficial/not beneficial/unsure. Items i-v were taken from (Mead et al., 2021b); items vi-x were added to assess opinions of UA related to the pandemic.

Procedure and design
The survey was hosted via Qualtrics (Provo, Utah) and accessed via a weblink placed in study advertisements. Participants read an information sheet (Participant Information Sheet), provided informed consent, then indicated their location (country). The survey was terminated if any country other than the UK was selected. Demographic questions, FIES, WHO-5 and COVID-related questions were then displayed in a random order with UA opinion questions always at the end. Links to the UK government website for definitions of key worker, social distancing, self-isolation and lockdown were provided next to the related questions. "Lockdown" was added as a category for isolation status alongside social distancing and self-isolation due to a change in UK law and terminology used by the UK Government shortly before the survey launched (Brown, 2020). Finally, participants were asked if they had taken part in the pre-pandemic study. This question was asked as we planned to compare data on opinions of UA between the current study and our previous study, and we aimed to avoid any overlap in participants.
A debrief sheet was displayed at the end of the study, and participants were invited to enter a prize draw to win a £20 shopping voucher as thanks for their time. The survey took approximately ten minutes to complete. The study was approved by the University of Liverpool Research Ethics Committee, ref: 5383. All data were collected between 25/03/20 and 07/04/20.

Analysis
Data were analysed using SPSS 24 (IBM Corp., Armonk, New York, USA). Scale scores were calculated in accordance with author instructions and as described above. Significance was set at p<.05 for statistical tests, except for those where it was adjusted to p<.01 or p<.005 to correct for multiple comparisons. Exploratory mediation analysis was conducted in PROCESS version 3.5 for SPSS (Hayes, 2018) with bias corrected bootstrapping (1000 samples) for analysis of indirect effects, with 95% confidence intervals reported.
For analyses that compared participants engaged in home food growing to those who were not engaged in home food growing, we defined Growers and Non-growers as described above (see Home food growing). Scores on the FIES and WHO-5 violated the assumption of normality so nonparametric analysis was used to test for differences in food insecurity and well-being between Growers and Non-growers (Hypothesis 1 and 2). A multinomial logistic regression was used to test if participants with greater experience of food insecurity would be more likely to report intentions of engaging in home food growing (Hypothesis 3).
Non-directional analyses were conducted to test if opinions of UA had changed from before the COVID-19 pandemic to the early COVID-19 lockdown (Hypothesis 4). Responses to the UA opinion questions from the current study were compared to responses to the same questions from the pre-pandemic study. Responses from the pre-pandemic study were collected in summer 2019, before the COVID-19 pandemic began. Chi-Square tests were used to test for differences between responses to UA opinion questions and studies (current study; Mead et al., 2021b). The significance threshold for these analyses was adjusted to p<.01 to correct for multiple comparisons. An additional exploratory analysis was conducted to assess if responses to UA opinion questions were associated with participants being classed as a Grower or Non-grower (p<.005 to correct for multiple comparisons). Significant Chi-Square tests were explored further by examination of standardised residuals.

Sample characteristics
In total, 684 participants entered the survey. Of these, 523 participants reached the end of the survey; 161 were excluded due to their only partial completion of the study or their location being outside of the UK. Responses with missing data or where participants failed the attention check were also excluded, leaving a final sample of 477 participants.
Most participants were female (369, 77.4% female), White (92%) and employed full time (56.8%) ( These analyses and a breakdown of demographic information by grower status are shown in the Extended data and Tables S1 and S2 (Mead, 2021).
Most participants reported engaging in some form of COVID-19-related isolation measures. Overall, 98 participants (20.5%) identified as keyworkers, 267 participants (56%) reported being concerned about their access to food, and 358 participants (75.1%) reported taking extra measures to ensure they had enough food during the pandemic. Additional sample descriptives regarding COVID-related questions are shown in the Extended data (Mead, 2021; Tables S1-5).
Home food growing, well-being and experience of perceived food insecurity

Mediation analysis
We conducted an exploratory, simple mediation analysis ( Figure 2) to test if perceived food insecurity (FIES score) explained the relationship between home food growing and well-being (WHO-5 score). Age (years) was entered as a covariate given the finding that participants in two categories of grower status differed in age. Associations between the variables are shown in Table 2. Overall R 2 for the model was .110. Being engaged in home food growing (versus not engaged) was associated with lower food insecurity, and lower food insecurity was associated with better wellbeing. There was a significant indirect effect of home food growing on wellbeing via food insecurity (IE=.914, SE = .473, CI .184 to 2.016). The direct relationship between home food growing and wellbeing was also significant, such that being engaged in home food growing was associated with better well-being. This indicates that, when controlling for age, the relationship between home food growing and well-being was partly mediated by food insecurity.
Food insecurity and intention to engage in home food growing Data from participants categorised as Non-growers (n = 325) were used to test if those with greater experience of food insecurity would be more likely to report intentions of engaging in home food growing (Hypothesis 3). Data were split by Non-grower response type: i) did not currently grow their own fruits and vegetables but were planning to do; ii) were not planning to grow their own fruits and vegetables to but would consider doing so; or iii) were not planning to grow their own fruits and vegetables and would not consider doing so. Data were entered into a multinomial logistic regression  with FIES score (food insecurity) as the predictor and grower status as the outcome, model χ 2 (2) = 1.297, p=.523. The category "iii) not planning to do and would not consider doing so" was the reference category as these participants reported having no intention of engaging in home food growing.
FIES score did not significantly predict membership of either category of intention to engage in home food growing (p's>.05). Therefore, experience of food insecurity did not predict immediate or future intentions to engage in home food growing. Hypothesis 3 is not supported. The regression table for this analysis is shown in Extended data, Table S6 ( Mead, 2021).

UA opinion questions
To test for differences in opinions of UA we compared responses to UA opinion questions i) -v) from participants in the current study to responses to the same questions by UK-based participants in the pre-pandemic study (Mead et al., 2021b). Data from participants who reported taking part (n=14) or being unsure if they took part (n=19) in the pre-pandemic study were excluded from this analysis. Data from the remaining 444 participants and the participants from the pre-pandemic study were analysed using Chi-Square tests to determine if there was an association between study (current study, pre-pandemic study) and proportion of participants indicating that they thought UA could be beneficial/not beneficial/ unsure to the following: i) You, your family and friends, ii) Your community, iii) The environment, iv) The economy, and v) Society in general. Descriptive characteristics for participants from the pre-pandemic study and results of these analyses are shown in the Extended data and Table S7-8 (Mead, 2021).
There were significant differences between study and responses to item iv) "the economy" (χ 2 (2) = 8.853, p=.002), and between study and responses to the item "society in general" (χ 2 (2) = 10.628, p=.005). We inspected standardised residuals to determine what may be driving these associations and found that these did not reach the adjusted threshold for statistical significance (z < 2.58; p's > .01). These analyses are reported in the Extended data (Mead, 2021). There was no evidence for differences between study and responses to the other items (p's>.01). These results therefore do not indicate any shift in opinion of UA between the pre-pandemic study and the current study. Overall, opinion of UA appeared to be positive across both studies: the proportions of participants endorsing UA as beneficial for items i)-v) ranged from 68.2-94.1% (Table S8).

Exploratory analyses
Chi-Square tests were used to test whether membership of the categories Grower or Non-grower was associated with responses to UA opinion questions i)-x). In addition to items i)-v) listed above, participants indicated if they thought UA could be beneficial to vi) Entertainment/alleviate boredom, vii) For education/home-schooling, viii) Alternative income because of unemployment, ix) Ensuring my family and I can eat healthily, and x) Ensuring social distancing by accessing food without being in contact with others. An adjusted p value of p<.005 was used for these analyses to correct for multiple comparisons. There was a significant association between grower category and response to the item vi) Entertainment/ alleviate boredom (χ 2 (2) = 12.575, p = .002). However, differences in proportions of participants selecting beneficial/ not beneficial or unsure between the grower categories did not reach the adjusted threshold for statistical significance (p>.005). No other associations between grower category and response to UA opinion questions were significant (p's>.005), therefore is no evidence that membership of the Grower or Non-grower category was associated with opinion of UA.
Opinion of UA appears to have been high across Growers and Non-growers: the proportions of participants endorsing UA as beneficial for items i)-x) ranged from 50.5-96.7%. Results of these analyses are shown in the Extended data, Table S9 (Mead, 2021).

Discussion
The current study provides, to our knowledge, the first data on the relationships between home food growing, perceived food insecurity and well-being during the early stages of a national COVID-19 lockdown. Those engaged in home food growing reported less food insecurity and better well-being than those not engaged in home food growing. In an exploratory analysis, we showed that engaging in home food growing was associated with better well-being. This relationship was partially mediated by perceived food insecurity, such that engaging in home food growing was associated with less food insecurity, which in turn was associated with better well-being. The analysis also showed a direct association between food growing and wellbeing which was not explained by perceived food insecurity, indicating that there are other drivers of this relationship beyond those measured in the current study. Importantly, these effects were significant when age was controlled for. We also found that food insecurity did not predict immediate or future intentions to engage in home food growing. Finally, we saw no evidence for a difference in opinions of UA between the current study and data collected before the pandemic, or for differences in opinions of UA between participants who were or were not already engaged in home food growing.
Considering our findings, we tentatively suggest that home food growing may have had some form of protective effect over food insecurity and well-being during the early stages of the UK COVID-19 lockdown. Our findings of reduced levels of food insecurity and better well-being for Growers compared to Non-growers, combined with our finding that the relationship between home food growing and well-being was partially mediated by lower food insecurity, suggest that home food growing could be a protective factor in the face of challenges to food security and well-being posed by the pandemic (or other future crises). Previous research has shown that UA/home food growing has benefits for food insecurity and well-being (Algert et al., 2016;Galhena et al., 2013;Genter et al., 2015;Kortright & Wakefield, 2011;Soga et al., 2017a). Others have theorised that these effects would be evident in the face of food system shocks (Galhena et al., 2013). The current study extends these findings and suggestions. The UK was in the early stages of lockdown when these data were collected, and levels of food insecurity due to lack of access to food had dramatically increased across the UK (Loopstra, 2020). Our observation that home food growing may provide a protective effect over well-being and food insecurity during this time provides novel evidence that home food growing may help to mitigate the negative consequences of food system shocks, such as those characterised by the disruption faced in the UK during March-April 2020. Food insecurity and poor well-being are continuing concerns, especially as the impact of Brexit and upcoming reductions to Universal Credit are predicted to exacerbate inequalities further and disproportionately impact on vulnerable groups (Barons & Aspinall, 2020;Goodwin, 2021). The current study then provides an impetus for exploration of policy mechanisms that can support adoption of home food growing, with the aim of protecting and increasing resilience to food system shocks, and supporting access to fruits and vegetables.
Reductions in well-being have been reported during the early stages of the pandemic, and are predicted to persist as lockdown restrictions continue (Daly et al., 2020;Pierce et al., 2020).
Projections of the scale of the negative impact that the pandemic will have on well-being indicate large increases in the numbers of people within the UK requiring support for mental health issues (O'Shea, 2020). This has prompted calls for innovative, efficacious interventions to support well-being in the context of the pandemic (Holmes et al., 2020). Thus, findings from the current study suggest that home food growing should be investigated as a means of supporting well-being during the pandemic, although longitudinal data are needed to confirm this and identify the underlying mechanisms that might explain this effect. The restorative impact of time spent in nature and opportunity for physical activity associated with gardening activities, alongside or in addition to the act of food growing, may have also contributed to the increased well-being reported by Growers in this study (Soga et al., 2017b). Nonetheless, our finding of greater self-reported well-being among Growers compared to Non-growers is consistent with existing research that has highlighted the potential benefits of UA and home food growing for well-being (Audate et al., 2019;Howarth et al., 2020;Kingsley et al., 2009;Lovell et al., 2014). Efforts to facilitate the uptake of home food growing to maximise its potential benefits for well-being, such as national campaigns or subsidies to provide equipment, could benefit from capitalising on the increased public interest in home food growing that has been (anecdotally) reported by UK media outlets.
It should be noted that the March-April period when these data were collected was early in the UK growing season, so large food harvests from home food growing may have been unlikely. The protective effect of home food growing that we tentatively propose may be due to home food growing providing reassurance during an acute food system shock (early lockdown), rather than an immediate food source. Participants who reported already being engaged in home food growing may have been reassured by the expectation of what they would be able to harvest in the future should purchasing restrictions in supermarkets continue. It may also be that these individuals are better networked in local food growing communities, affording them better knowledge and access to less disturbed local food sources. However, we did not measure the scale of the growing that people were engaged in, or if they engaged in home food growing for subsistence or a hobby, and if this changed in response to the pandemic. Future work needs to consider this, as the amount of food that home food growing is expected to produce may moderate these effects.
We found no evidence that participants with higher levels of perceived food insecurity had immediate or future intentions of engaging in home food growing. This has implications for intervention development because home food growing has been shown to have benefits for tackling food insecurity (Algert et al., 2016;Galhena et al., 2013;Kortright & Wakefield, 2011). Participants in this study may have been unaware of such potential benefits, or even felt discouraged from engaging in home food growing because of perceived barriers such as limited time, space, knowledge or financial resources (Mead et al., 2021a;Schupp et al., 2016). Future home food growing interventions may wish to specifically target these individuals to address any barriers, both perceived and real, and raise awareness of the potential benefits of home food growing for tackling food insecurity, as these individuals may benefit from a home food growing intervention that is accessible and easily available.
These findings have implications for the wider understanding of the consequences of the COVID-19 pandemic. Food insecurity has been identified as an underlying factor in a "global food syndemic" that may exacerbate the impact of the COVID-19 pandemic (Huizar et al., 2021). Huizar et al. highlight that food insecurity is linked to malnutrition and obesity, which in turn have been identified as risk factors for coronavirus infection and greater severity of the disease (Gao et al., 2020). Thus, effective measures to reduce food insecurity may have an indirect effect on disease vulnerability via associated levels of malnutrition and obesity. Tackling food insecurity requires a multi-actor approach and is beyond the scope of this work, however the current study highlights the potential for home food growing to have benefits for perceived experience of food insecurity during the pandemic. In addition to this, home food growing and UA in general have been associated with better diet quality and reduced likelihood of having overweight or obesity during non-pandemic times (Alaimo et al., 2008;Kamphuis et al., 2006;Mead et al., 2021b;Palar et al., 2019;Zick et al., 2013). With future lockdowns and social restrictions likely, home food growing should be investigated as a sustainable, holistic option for interventions that mitigate the negative consequences lockdown and the pandemic may have for food security and, indirectly, obesity, malnutrition and disease vulnerability.
Our secondary aim was to explore if opinions of UA may have shifted from 2019 (Mead et al., 2021b) to the early stages of the pandemic. We also tested for differences in opinion of UA between Growers and Non-Growers. In the face of global , environmental, and political (Brexit) challenges to the UK food system, urban-grown food represents an innovative, sustainable solution to safeguarding the supply of healthy, nutritious food for the population (Edmondson et al., 2019, Edmondson et al., 2020Kourmpetli et al., 2020). Furthermore, the pandemic has highlighted shortcomings and inequalities in the UK food system (Power et al., 2020). Understanding the impact on opinions of UA could help to inform future efforts to expand UA as a food systems solution. However, we saw no evidence that the COVID-19 pandemic or grower status were associated with opinions of urban grown food. Opinion of UA appeared to be generally positive both before and at the early stages of the pandemic, and regardless of participants' engagement in home food growing during the early UK lockdown. The consistently positive opinion of UA in our UK-based studies is promising for efforts that aim to increase the uptake of urban-grown food. This suggests that neither the pandemic nor participants' engagement in home food growing have had any negative effect on opinions of UA. However, specific concerns relating to urban-grown food, such as quality and safety, still need to be identified and addressed as part of efforts to upscale the UK's urban-grown food production (Kim et al., 2014;Kourmpetli et al., 2020;Specht et al., 2016b).

Limitations and future directions
Longitudinal data are needed to assess how home food growing affects well-being and perceived experience of food insecurity as the pandemic progresses. The cross-sectional nature of this study means we are unable to test the longevity of the effects reported here. Future work should consider assessing home food growing, well-being and food security at multiple time points to see if these effects persist as the pandemic continues. With future lockdowns and social restrictions likely (Brown, 2020), innovative, sustainable activities will be needed to help mitigate the negative consequences of the pandemic.
Although home food growing is considered a form of UA, it should be noted that we did not assess if participants resided in urban or rural areas. We are therefore unable to confirm if participants who reported engagement in home food growing were doing so in urban areas. Others have reported associations between home food growing and food security in rural settings (Rammohan et al., 2019), therefore, our conclusions should be considered as limited to the impact of home food growing itself, rather than the wider practice of UA. Future studies should address this limitation by considering potential differences in the experiences of urban-and rural-based participants and if this might impact on their opinions of UA.
We did not assess if Growers were engaged in home food growing prior to the pandemic, or if they had recently become engaged in home food growing response to lockdown (new Growers). It is possible that the lack of distinction may have obscured any subtle differences between the experiences or characteristics of such types of Growers. Furthermore, we did not assess the scale or longevity of participants' engagement with home food growing and if this influenced our findings. Future work may wish to consider if large-scale, established home food growing, such as allotment growing, may have a different effect on well-being and food insecurity than small-scale, home garden-based growing.
We did not measure socioeconomic status (SES) of our participants, which is an additional limitation because SES could account for the associations between home food growing and lower food insecurity/better well-being. Importantly, participants in the different categories of growing status were well-matched for other demographic characteristics and did not differ in employment status. Due to a difference in age between two categories of grower status, age was controlled for in the mediation model and the relevant significant associations remained. Furthermore, controlling for age and SES in our previous study did not attenuate associations between urban agriculture, diet quality and potential mediators (Mead et al, 2021b). Therefore, we tentatively suggest that differences in SES between participants are unlikely to fully explain our results though future work is needed to confirm this.
It should be noted that levels of perceived experience of food insecurity in our sample were low. Although we detected group differences in food insecurity between Growers and Non-growers, these findings may not be generalisable to participants or settings with high levels of food insecurity. We also adapted our measure of food insecurity (the FIES) to assess experience of food insecurity dues to a lack of food in shops over the preceding two weeks. Thus, we are unable to ascertain if the experiences of food insecurity reported by our participants were acute or chronic. The experience of food insecurity reported by participants in the current study may not be comparable to that of individuals with the highest levels of chronic food insecurity, therefore a more diverse sample is needed to extend these findings.
Finally, participants in the current study were predominantly white and female, thus limiting the generalisability of these findings to more diverse groups.  Alaimo et al., 2008), however, this research should be repeated with a more representative sample of participants to ascertain if the findings reported in the current study are applicable to other groups.

Conclusion
Home food growing may have had a protective effect on levels of well-being and perceived experience of food insecurity at the start of the UK COVID-19 lockdown. Those engaged in home food growing reported higher levels of well-being and lower levels of food insecurity than those who were not engaged in home food growing. Home food growing was associated with lower food insecurity, which, in turn, was associated with better well-being. Opinions of UA were generally positive before and during the early stages of the pandemic, and regardless of participants' engagement in home food growing. These results suggest that home food growing may have benefits for well-being and food security during lockdown, but longitudinal assessment of this as the pandemic progresses is needed to confirm this. Upscaling home food growing by increasing public interest and facilitating engagement in growing by making such opportunities more accessible could have tangible benefits for mitigating the impact of the COVID-19 pandemic on well-being and food security. Policies that provide access to land, skills, and opportunities for food growing should be developed to utilise the potential beneficial effects of home food growing for food system resilience, food security, and individual well-being.

Data availability
Underlying data OSF: Home food growing, well-being and food security during the COVID-19 UK lockdown. https://doi.org/10.17605/ OSF.IO/7EZJQ (Mead, 2021) This project contains the following underlying data: • Home food growing lockdown study data Mead.sav • pre-pandemic UA opinon comparison.sav Extended data OSF: Home food growing, well-being and food security during the COVID-19 UK lockdown. https://doi.org/10.17605/ OSF.IO/7EZJQ (Mead, 2021) This project contains the following extended data:

Open Peer Review
Overall, this is a very well written manuscript with a particularly clear method and results section. The study examines the role home food growing had on well-being and food insecurity during the UK national lockdown due to the COVID-19 pandemic. Understanding the impact home grown could have is important for both the current pandemic, potential future pandemics but also general well-being and food insecurity.

Methods
Participants -expansion on recruitment from the general online groups may be useful. For example, were they community groups from certain areas and therefore were a number of areas targeted? Did participants have to be from an urban or suburban area? Wondering if this may impact on their views of UA. May recruitment from these groups also be a limitation, leading to a predominantly white and female sample?
Analysis -there is repetition from the measure section for home food growing definition, given it is a short manuscript and the definitions are clearly articulated, one definition may be sufficient.

Results
Food insecurity and intention… -It may be worth removing "Non-grower" where you identify category iii as the reference group as here all categories are "non-growers".

Discussion
I am unsure if there are differences, but are males or females more likely to home-grow food? It is mentioned as a limitation but wondering how much this may influence findings.
It has been mentioned several times but may be worth further reiterating that food insecurity continues to be of concern due to Brexit and changes to universal credits etc. It may be worth further emphasis the continued/broader applications of this study.

General
It may just be the rendering/size of the image but the blue/white text in the middle of Figure 1 is difficult to read Is the work clearly and accurately presented and does it cite the current literature? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results?

Yes
Could any solutions being offered be effectively implemented in practice? Yes Thank you for these comments and suggestions. We have uploaded an amended version of the manuscript addressing these. Specifically, we have focussed the Discussion on the UK context and expanded our discussion of well-being perceptions. Regarding maps of study areas/participant locations and separating urban and rural participants, we are unable to produce these as we did not assess precise locations of the participants other than confirming that they were based in the UK. We have acknowledged this as a limitation of the study and expanded the Limitations and Future Directions section of the manuscript in light of this. We have also expanded the Conclusion section to include a policy implication, as suggested.
Thank you for your helpful and insightful review.

Competing Interests:
No competing interests were disclosed.