Strategizing Dinner: How American Pantry Users Think about Feeding Their Families with Limited Resources

: The ﬁnancial crisis associated with the COVID-19 pandemic has exacerbated food insecurity in the United States. The emergency provides an opportunity to re-think the American nutrition-assistance system. In this paper, we describe ﬁndings from a community-based project conducted in urban Alaska before the pandemic in collaboration with a local food pantry. We conducted semi-structured interviews with nineteen food recipients, half of them twice, about how they procure food and prepare their meals in the context of juggling other expenses and demands on their time. What participants in our study do ﬁts mainstream American patterns. Our study adds to the knowledge base by focusing on how families think strategically about their situations in context. In addition to cost, participants take nutrition and ﬂavor into account. Most importantly, they do not think about assistance programs in isolation but holistically.


Introduction
Food security is defined by the Food and Agriculture Organization of the United Nations as occurring when "all people [in a household], at all times, have physical, social and economic access to sufficient, safe and nutritious food which meets their dietary needs and food preferences for an active and healthy life" [1]. The COVID-19 pandemic has significantly exacerbated food insecurity [2,3]. Already a problem in the United States before the pandemic, as in other rich countries [4], increased food insecurity has prompted the temporary expansion of benefits for nutrition-assistance programs [5,6]. The current willingness to reconsider program guidelines provides an opportunity to improve the hunger-response system beyond the pandemic. The existing literature is limited regarding the perspectives of service recipients, and existing studies typically focus on single programs rather than the totality of American nutrition assistance options [7,8]. In this paper, we describe findings from a community-based project conducted in collaboration with a local food pantry. We interviewed food recipients about how they procure food and prepare their meals in the context of juggling other expenses and demands on their time. Although our research was conducted before the pandemic, the reasoning and strategies of our participants informs ongoing discussions.

Food Insecurity in the United States and Responses to It
It is well-established that, across lifespan and life circumstances, food insecurity is associated with poor health, including diet-related conditions [9][10][11][12][13]. Common negative health outcomes that can result from food insecurity include diabetes [14,15] and obesity [16]. Food insecurity has also been linked with decreased mental health [17,18] and congenital disabilities [19], among other adverse outcomes. In 2019, 10.5% of U.S.

Overview of the U.S. Public Safety Net for Nutrition
The U.S. government funds multiple means-tested programs that support nutrition. The USDA houses most of these programs, which speaks to a dual purpose: they stabilize agricultural prices as well as address poverty and hunger. Participation in government programs has been identified as "one of the key characteristics distinguishing VLFS [very low food secure] households from those that were more food secure," reinforcing both the degree of need and the inadequacy of these programs to meet the need by themselves [27].
The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) targets pregnant and nursing women and children up to age five. WIC is the only foodassistance program that includes nutrition risk as an eligibility criterion. WIC coupons can be used to purchase foods from specific categories consistent with the nutritional needs of program participants. WIC has been shown to improve household food security in participating families [28].
Like WIC, the Supplemental Nutrition Assistance Program (SNAP; formerly food stamps) provides food-specific income. Food options are less restrictive-most unprepared foods are acceptable-and the income guidelines are less generous (180% of the poverty line for WIC and 130% of the poverty line for SNAP). Though the food subsidy funds are federal, states administer the program and may add eligibility criteria; criminal history and work status are among them in some states. SNAP is the most extensive nutrition assistance program, serving some 40 million people in 2018 [29]. SNAP "has become one of the most effective antipoverty programs overall," and "higher SNAP benefits appear to reduce the risk of food insecurity" [30].
National school lunch and breakfast programs, and related programs such as the Child and Adult Care Food Program reimburse schools and care facilities for meals and snacks provided to income-eligible children or seniors. The USDA Food and Nutrition Service also administers several commodities programs that provide boxes of designated food items to low-income households. Some programs target specific sub-populations. Programs include the Emergency Food Assistance Program; the Commodity Supplemental Food Program, which serves senior citizens ("senior boxes"); the Food Distribution Program on Indian Reservations, which serves Native Americans and Alaska Natives.

Overview of the Charitable Safety Net for Nutrition
Soup kitchens trace their origins to the Great Depression, and the first food bank was established in 1967; however, charitable food assistance grew into an "industry" in the 1980s and 1990s as the federal government dramatically scaled back the public safety net [31].
Feeding America is a national network of some 200 regional food banks. Feeding America functions as a hub for procuring food through partnerships with national manufacturers and distributors, government commodity programs, retail partners, farmers, and through purchases funded by grant money or individual donations [32]. Food is sent on to its food bank partners, which likewise collect food from local/regional retailers, farmers, and individuals, and which serve as central warehouses for distribution partners. Pantries and meal-service sites such as soup kitchens and senior centers "purchase" food from the banks for pennies per pound, solicit donations, and ultimately distribute food to those who need it. Pantries are also typically where individuals pick up their federal commodities boxes. Strictly speaking, food banks do not distribute food directly to households, although the term is often used to refer to pantries as well. There is a growing trend among pantries to use the "client choice" model, whereby clients may select foods from within different categories: rice or pasta, for example, or two cans of fruits from the fruit shelf, rather than being given a pre-fabricated box with items family members may not want [33].
Originally conceptualized as emergency assistance, the charitable safety net has become a regular source of food for many food-insecure families [34]. In 2017, 26% of food-insecure households obtained food from a food pantry [25]. Nationally, although lowincome families are the most frequent users of the charitable food safety net, increasingly, families with relatively higher incomes are turning to pantries as a resource because they are unable to make ends meet [35].

Personal Safety Net
Although not a focus of this research, the personal safety net warrants a brief mention. The personal safety net is the one everyone relies on to weather difficult times: personal resources and the family and friends who can provide financial or other aid. As documented by Edin and colleagues in particular [22,36,37], the social networks of low-income people may well include people who are eager to help but have little left to share.

Eating in Poverty: Navigating Limited Food Resources
It is widely established that low-income families juggle monthly expenses. Household budgets are typically inelastic. Thus, rising costs for other basics needs can increase food insecurity, including health care [2,38] and housing [39]. Families typically prioritize housing: rent and utilities [40]. In the most recent national report of a periodic survey, 59% of U.S. pantry users said they must choose between paying for food or paying for their rent/mortgage, and 72% must choose between food and utilities [41]. Some low-income individuals buy lower-cost, higher-calorie, and less-healthy food in an attempt to stretch their food budget [42]. Sometimes parents, and even older siblings, forego eating altogether to feed their children, as indicated by higher rates of food insecurity among adults than children of various ages within the same households [43]. Similar strategies have been identified among food-insecure families who are not poor and do not use pantries [44].

Materials and Methods
The first author has a long-standing relationship with the Food Bank of Alaska (FBA, Anchorage, AK, USA). Inspired by a previous collaborative project [45], staff and volunteers at FBA-partner St. Francis House (SFH, Boston, MA, USA) food pantry at Catholic Social Services (CSS) approached her for assistance in exploring how their clients experience housing insecurity, how their pantry services support housing needs, and how SFH could better support housing stability for its participants [46]).

Research Site Description and Setting
Conditions in Alaska reflect national patterns. In 2018, the food insecurity rate in Alaska was 12.9% [47]. In fiscal year 2019, 12% of Alaskans utilized SNAP [48]. In Anchorage, the state's largest city and where St. Francis House is located, an estimated 11.5% of residents are food insecure [49].
The Food Bank of Alaska is part of the Feeding America Network and acts as a statewide hub for food distribution to non-profit and tribal partners. Before the pandemic, the FBA distributed food to over 300 food pantries, soup kitchens, senior centers, and children's programs throughout Alaska-and through these organizations, to about 155,000 people, according to the most recent statewide survey [41]. According to the same survey, 32% of food recipients in Alaska are children under 18, and 13% are seniors. Sixty percent of clients of the Food Bank of Alaska network have at least one adult in the household who has worked for pay in the last year. Fifty-three percent of this group of working FBA-network clients still have incomes below the federal poverty level. Of those not working, 21% are retired, and 69% cannot work due to disability [41]. FBA also distributes perishable food directly in Anchorage via a mobile food pantry.
St. Francis House is the largest food pantry in Alaska. It uses the client choice model. In 2018, 88% of St. Francis House clients were food insecure [50]. St. Francis House is a member of the Food Bank of Alaska network and, by extension, part of the Feeding America network. SFH distributes senior boxes (commodities) and food donated to Feeding America, FBA, or directly to SFH/CSS and its partners. Both FBA and SFH also purchase food with donated money. Clients typically visit the pantry once each month. There are no rules restricting clients' use of other pantries. In 2017, according to pantry administrative data provided by the program manager, SFH served 7197 households, or 15,570 people (D. Rittenberg, personal communication, 9 February 2018).

Recruitment
Using procedures approved by the university's Institutional Review Board, we recruited participants by posting flyers about the project at the pantry and invited clients to leave their names with pantry staff; we contacted them later to further explain the project and, if they were willing, to set appointments for the first interview. We also made verbal announcements in the pantry waiting room on several occasions and made appointments with people who approached us; in some cases, we conducted first interviews that same day. SFH offered interview space and also tokens allowing families an additional pantry visit as an interview incentive.

Data Collection
Our team conducted paired semi-structured interviews with a convenience sample of nineteen SFH pantry clients through the spring and early summer, 2017. Rigor in qualitative research is as much a function of data collection-obtaining thick description from participants-as it is of analysis [51]. To ensure high-quality data, the team roleplayed with the interview guide, and the students observed faculty conducting interviews before conducting interviews themselves.
The first interview consisted of five questions, each with several probes, addressing the family's experiences with the pantry, sources of food besides the pantry, worries about food, overall financial situation, and health. Participants also completed face sheets for demographic information and facts about assistance programs utilized and sources of food. Interviewers used the face sheet information to tailor interview questions.
Participants were asked to participate in a second interview. They were to take three photographs of meals, preferably three of the same meal on different days. The team had disposable cameras available, however, everyone who agreed to take photos had cameras on their phones. The use of photos minimized recall bias, and through both the pictures themselves and the ensuing reflections, we elicited additional details about more meals per family [52,53]. Approximately half of the initial participants (n = 10) followed through with taking pictures and engaging in the second interview, resulting in thirty photos.
Photos usually showed plates of food, sometimes food in pots before it was served, but not people. Follow-up interviews focused on those meals: where foods (ingredients) came from, who prepared them, who ate them, how satisfying they were. The second interviews finished with a description of what participants considered to be a "good meal" ("not a special meal, like a birthday or holiday, but one you would feel good about serving your family").
CSS had interpretation services available for their clients and offered these services for the pantry interviews; however, no volunteers for the study wanted to use them. The audio-recorded interviews were professionally transcribed. Voice recordings were erased once team members had the opportunity to listen to them alongside the transcripts, thereby both verifying the transcripts and hearing the spoken tones and feelings not captured in the written documents.

Data Analysis
The previous study and St. Francis House's particular interests provided sensitizing concepts [54,55]. Analysis began with the first interviews insofar as the faculty members and the observing students discussed what stood out as part of the debriefing. As transcripts became available, they were analyzed thematically, primarily by one faculty member and the students. We used a constant comparative process borrowed from constructivist grounded theory [55,56]. Identification of initial themes began while data collection was still in progress so subsequent interviews could probe more deeply into emerging ideas. Although each team member had a particular interest area, each viewed the photos and read the entirety of each transcript carefully and repeatedly. The group discussed ideas collectively for consensus [51] and to ensure reflexivity regarding researchers' social location and disciplinary biases [56]. Identifying and evaluating themes was done without specialized software. Our analytic notes consisted more of shared marginalia on transcripts than separate documents; however, the team did communicate and build on each other's ideas. We reached data saturation, linking "similar concepts and processes in different instances, experiences, contexts, and events [in a way that] develops concordance within the dataset" [51].

Participants
The convenience sample of people who approached us to participate consists of people collecting food from the pantry. Interviews indicate that they typically were also those who prepared food for their families. Our sample is more racially and linguistically privileged than the St. Francis House clientele at large. In particular, more than half of the individuals in our small sample are white, far outnumbering any other racial/ethnic group, whereas only 25% of 2017 SFH clients (heads of household) are white. Furthermore, English is the language of the home of all but one participant. This outcome is in contrast to the 34% of SFH heads of households in 2017 who identified their primary language as something other than English (SFH administrative data provided by D. Rittenberg, personal communication, 9 February 2018).
Seventeen of the nineteen participant households-89%-received one or more forms of means-tested public assistance, indicating that they are poor or near-poor (Table 1). In all but one case, the assistance included at least one program directly related to food. Participant households also received ongoing help with rent, heat, and childcare, and one household utilized the state-funded, event-specific emergency General Relief Assistance. No family now received Temporary Assistance for Needy Families (TANF; locally: Alaska Temporary Assistance Program, ATAP), although one family had done so until they reached their 60-month limit on benefits.

Indications of Vulnerability
None of the participants in our sample self-identified as homeless. However, four of the nineteen participant families (21%) were in precarious housing situations. Indeed, one family would qualify as homeless under official U.S. definitions. The three other families were not homeless when we met but described circumstances that suggested their housing would be at risk soon, such as having almost depleted an insurance payout from which they were paying rent. Notably, almost a third of the sample (n = 6) received housing assistance; for four of them, housing was now stable, however, the other two struggled to pay even their reduced rent. One family was without hot water in their mobile home for six months because they could not afford to replace the heater.
Almost all families in our sample experienced health challenges. Seven of the nineteen (36.8%) participants lived in households where someone received Disability income, sometimes Supplemental Security Income, sometimes Social Security Disability Insurance. Of the remaining twelve families, participants from eleven mentioned their own or family members' health issues during their interviews: high blood pressure, diabetes and prediabetes, depression, COPD, and a recent aneurism. Some of these conditions are directly linked to diet quality.

Themes
Themes that emerged from the interviews included (1) Sources of food utilized, and three themes related to strategies for feeding their families: (2) Paying other bills: Food is where the wiggle room is, (3) Procuring food: Pantry first, and (4) Stretching, preparing, and serving food.

Sources of Food Utilized
By definition, everyone in the sample utilized the SFH pantry, and all but one participant also engaged with other public and/or charitable food safety net programs. More than half of our sample received SNAP; however, some households received minimal dollar amounts, sometimes less than $20/month. Seven of nine families with children in the home full-time utilized WIC and/or the federal school meal program, depending on the children's ages. Five of seven households that include seniors received the senior boxes at St. Francis House each month; a few non-senior families indicated that they received other commodities elsewhere. All but two of the participant households utilized the St. Francis House pantry every month. Of the two who did not come every month, one was coming to the pantry for the first time when we met, accompanying a relative who did come regularly. The other person had a family member who was intermittently employed, and when working, earned enough to make the pantry unnecessary-however, this family still used the pantry several times a year. Every household, with one exception, used at least one other pantry regularly as well; some used several others. The other pantries varied in size and selection; from a small food shelf at the participant's church, to the city's next-largest pantry, to the mobile pantries sponsored by Food Bank of Alaska. Two participants indicated that they utilized community kitchens whose typical clients are homeless. All of the participants shopped at local grocery stores, however, several SNAP recipients indicated that SNAP comprised the only money they spent on food.
In contrast to the findings of the earlier study [45], wild foods procured through hunting, fishing, and gathering do not play the prominent role for this urban Alaskan sample that they do for rural Alaskans, or even non-Native people living on the road system. However, a few SFH families did eat small amounts of wild foods. Our participants were gifted small quantities of game meats, received gifted seafood, engaged in limited fishing, and picked berries in season. More than half of the sample cultivated foods to varying degrees: growing herbs in a window box, maintaining a single indoor container of tomatoes, participating in community gardens, and raising vegetables in the backyard.
Participants' stories of their lives and their meals routinely included other people, their personal safety nets. For example, one participant regularly combined foods with a neighbor (also a pantry client), and their families ate together weekly. Some participants described sharing rides to the pantry. Another woman had a neighbor who sometimes shared leftovers; someone else had received some food from a fellow church member who was moving away. People mentioned shopping at Sam's Club and Costco with others who had memberships. Illustrating how limited the resources of many participants' friends and families were, one woman said the main person she could rely on for extra help was her mother-however, it was her mother who had introduced her to the pantry, having been a long-time user herself. Likewise, one woman said, "There were a few times I just cried, and then it got to the point where I had to call my [adult] daughter" and ask for help, but "I mean, she's not got a ton of money either, so it's not something I would want to do to her often."

Paying Other Bills: Food Is Where the Wiggle Room Is
In line with the literature reviewed above, SFH pantry clients prioritize paying their housing costs before paying for food. A majority of participants stated clearly that their rent was what they pay first; rent was "non-negotiable," in one person's phrase. Two participants were paying mortgages-one on a trailer for which she also paid rent on the space where it sat, another on a house bought before a family business failed and for which foreclosure was an ongoing concern.
Utilities were another priority for the people we interviewed. They handled this cost in different ways. One woman said she had made sure to get an apartment that included utilities. The most common strategy was to make sure to pay utilities next after rent. Some juggled utility costs from month to month; for example, skipping a bill one month but then paying most or all of it the following to prevent having the service canceled. "There's always times I can't pay all my bills," said one person, summarizing what several others also said. Someone else who managed to pay all her bills said she had 46 cents left the previous month. The one participant who said he did not pay rent first, explained that he paid the electric bill first "because it takes longer to get evicted than get the power turned off." This man was behind on his rent but judged that he was paying enough to stay housed. One of the residents in subsidized housing recounted how it was only since her rent was significantly reduced that she could stay current on her other bills-and she still used the pantry regularly.
Transportation costs also came before food for some participants. Transportation usually meant a vehicle, typically an older model in need of repairs. One woman specified that her car was seventeen years old; other vehicle owners explained that their vehicles were paid off in wealthier times. One man recounted with frustration that he had bought and paid for a truck when he had a well-paying job, however, he subsequently lost the job and then was seriously injured. Having no income, he borrowed money against the truck's value, but with the loan, now he had to carry full insurance coverage, which was a significant burden. In contrast, some others described cutting back on vehicle insurance coverage as part of reducing costs. For all vehicle owners, gasoline was a significant expense, and many could not travel to all the places they wanted. One family specified that they could not access a mobile food pantry because they could not afford the extra gas.
Often, the participants who owned vehicles were those who talked about working during their interviews-they or a family member was currently employed, or they were looking for work. In the context of limited local public transportation and the long Alaskan winter, these individuals saw their cars as a factor in staying financially afloat. To the extent that pantry food and other assistance enhanced families' capacity to keep their cars, it contributed to the possibility of work and self-sufficiency as paired goals.

Procuring Food: Pantry First
The photos and accompanying descriptions of the meals' component items, where the components were obtained, and how the food was consumed provide a window into how pantry clients think about creating meals. Three significant sub-themes arose in how participants procure their foods. First, not surprisingly, was cost. Given that many families had little-sometimes no-money left after paying other bills, they prioritized food they could obtain for free or at a reduced cost. Nutrition was second. Participants referenced family health challenges as well as general public health messages about nutrition as they described their selection of foods at the pantry and the store; and third, flavor: condiments and spices were valued for livening up the monotony.
One mother described her careful planning of pantry use. When she lost her job, she got a list of all the pantries in town. She saw that several were open at different times on Tuesday, and for over a year, she had reserved the first Tuesday of the month for visiting five pantries. Early in the month, she obtained the best selection of food, she reasoned, and then she could use her SNAP benefits to fill in the gaps. Perishables from the pantries were eaten immediately or frozen. Sometimes she spent some SNAP dollars before her pantry round; for example, recently, the first of the month (when SNAP benefits become available) fell on a Thursday, and she was out of food so had to shop. Then, "WIC checks come late in the month to help carry us through." This combination of resources was adequate for meeting her needs in most months. Yet, during some months, she had to send her children to eat with relatives in other households, while she and other adults left at home ate scanty meals at the month's end.
Features of this mother's overall approach were apparent in others' as well. Going to the pantry early in the month was a common strategy. People who received federal commodities were eligible to pick up their new box, and they often made one trip for both commodities and the allotted monthly pantry visit. Items in the commodities box were pre-determined; however, the client choice model allowed for some discretion over other items. One man said he made a point to come first thing in the morning to get the best choices.
Going to the store after the pantry visit(s) and building purchased food around the free items was also common. Several people said that they shopped for food at commercial stores "very little," and the one mother was not alone in using only designated public assistance for purchases. Participants took advantage of sales and expiration date-related discounts. One woman said she went to the store weekly to catch the unadvertised markdowns, although she bought very little in any one trip.
The foods that participants purchased were typically things that they did not receive, or rarely received, from pantries. Many participants mentioned proteins specifically. St. Francis House had recently been distributing salmon steaks, whole chickens, and liquid eggs, and these proteins were present in people's photographed meals. Whereas, for several families, the proteins from SFH were their only meats for the month, others mentioned meats from the store; for example, a half-price steak filet. Another mother specified that her family's milk and eggs were purchased with WIC coupons.
Some participants used the pantries during the second half of the month after they had used their SNAP or other money for food. In these narratives, there was a tone of hope that maybe they wouldn't need the pantries this month-however, there was also a pattern of consistent pantry use.
As with the one careful mother, those who relied exclusively on assistance for their food struggled by the end of the month. One person figured that she and her son could eat on SNAP for about two weeks, and the pantries-SFH and two mobiles, in her case-got them through another ten days if they were prudent. The last days of the month were tenuous. "It's about the 20th or 21st when the panic sets in" said someone else about when both SNAP benefits and pantry foods were used up. In both cases, there was no set plan for those final days. Similarly, when we interviewed one mother during spring break, she was anxious because her children were not receiving their lunches at school and instead were eating through the month's food more quickly.
Most participants were also attuned to nutrition, despite knowing that they were not eating as well as they wished. One participant commented, "We probably get far too much junk food, just because that's what's easily available." She described avoiding the donated sweets when she could, as when she recently had the choice of an artisan cheese or pastries from the pantry extras shelf and took the cheese.
Also widely available were starches-pasta, rice, potatoes, and bread-so much that some people were tired of them. All participants indicated that they had received pasta and bread at their last pantry visit unless they intentionally declined it. They were given "pasta out the yinyang," as one person expressed it. Several said the starches they were offered served as "filler," but did not provide real sustenance. All but one of the photographed meals included grains or potatoes; arguably, half were dominated by them.
Several caretakers of young children explained their decision-making at the pantry and the store as a combination of wanting protein and vegetables and the key question: will the kids eat it? One Alaska Native woman explained that she had access to some traditional foods from family in rural communities, which she was confident were healthy; however, her relative-foster children had not previously been raised with such foods and at first refused to eat them. Although "now they're starting to eat Native foods" after some time with her, her pictures were still of western foods: cereal, shepherd's pie, and a grilled cheese sandwich.
Another participant made a point of including a protein, a starch, and a fruit/vegetable in each of the meals we discussed. "I try to have one good meal a day," she said, and the meals in the photos were her good meals for three days. The portions were quite small, and the produce in particular was minimally present; seven cubes of canned pineapple in one. For other participants, fruits and vegetables were present in some meals only as a carrot, onions, pickles, or fresh chives. In some photos, produce was conspicuously absent. In the 23 photos with a fruit or vegetable, about a third came from cans, and some participants expressed reservations about the sodium. The group of participants who took pictures of their meals included some of those who raised some of their own food, but because we collected data during the spring, the only home-grown foods being consumed at the time were herbs.
In contrast to limited actual intake of fruit and vegetables, however, most "good meals"-the hypothetical meals participants would want to provide for their families if they could-included salads or other vegetables. One woman with diabetes was particular about the fresh produce items she would choose to snack on through the day to stabilize her blood sugar.
Commentary about the photographed meals and the desired good meals makes clear that participants were aware of the compromises involved in eating diets heavy on starch and low on vegetables. Although grateful for the free food, some people with medical issues expressed frustration that they knew their diets exacerbated their conditions. Some participants spoke with enthusiasm about odd items they had received at the pantry that livened up their menus, such as Gorgonzola cheese, a curry dish, and barbeque sauce. For example, dried jalapeños added to scrambled eggs turned a "same old-same old" breakfast into a "rare treat." Other people also described their purchased food in terms of added flavor: sour cream, pepperoncini, spices, salad dressing, vinegar, teriyaki sauce, butter, mayonnaise. Two families who regularly shared foods made a large pot of spaghetti for about ten people and were pleased that, the week of the photographs, they could enliven this routine dinner with a package of Italian sausages from the discount bin. Someone else who had made clear that she grew tired of the repetitious meals described her response: "just season the dickens out of it." Two participants grew herbs-chives, rosemary, thyme, "and stuff like that"-in window boxes. In general, it was evident in many interviews that clients wanted to make their meals interesting, and they appreciated items that added flavor.

Stretching, Preparing and Serving Food
Sometimes the way foods were packaged was inconvenient, but people were reluctant to let anything go to waste. For example, at the time of data collection, SFH was distributing whole chickens, and some participants had to learn how to cut them up. The salmon was distributed in relatively large packages of frozen steaks that often stuck together, such that they needed to be thawed together, and then used quickly. One photo showed all of one household's salmon steaks cooked at once; the participant explained that she and her family ate some steaks for one meal, then she put the rest into a casserole for a few days later. Other small households said that they ate salmon for multiple days in a row. Someone else purchased a package of pork chops for half price; he cooked all the chops, then he ate some with his family and used the leftovers to add flavor to a bean soup.
Some of the fresh food needed to be frozen before it spoiled. For example, SFH was distributing blueberries consistently for a period the previous fall, and a participant who collected a lot of them froze them and was still putting them on her oatmeal several months later.
Several participants emphasized that their ability to cook, "to get creative with food," as one expressed it, made the difference for them. One woman had a crockpot and used it regularly. A woman whose family refused to eat any more cereal for breakfast invented corn flake cookies so as not to waste the cereal they received. Another participant commented that he observed people decline pasta "because they don't know how to cook it [besides adding red sauce] but sometimes the Korean sauces, like they've got this noodle sauce . . . ".
In line with national trends, adults in our sample said that children in their households were fed before adults when there was not enough for everyone. Several parents and caretaking-grandparents described skipping meals occasionally so that children could eat. A woman with three teenagers in the house skipped meals weekly. One man, whose daughter was with her mother the week he took his pictures, had no fruits or vegetables in his three photographed meals at all. He explained that he saved food and ate better when his daughter was present.

Discussion: Strategizing Dinner
This research has its origins in a university/community-pantry partnership, and our sample is comprised of people who use the pantry. It is not surprising that food from the charitable safety net plays a central role in our participants' photos and narratives. Indeed, although we know that almost all of our pantry participants also relied on other forms of assistance, it is striking that for nine-almost a third-of the meals for which we have pictures, the ingredients came exclusively from St. Francis House and/or another pantry. The predominance of pantry food may be attributed in part to the clients' having recently visited SFH when we asked them to take the pictures. Nonetheless, that people were eating so soon the food they had just received speaks to bare shelves in people's cupboards.
It is a strength of the qualitative project that we capture the totality of how pantry users think about feeding their families, which necessarily includes public and personal safety nets. That so many of our participants so carefully plan their pantry visits and use them as the starting point for building the month's supply of food points to how most of our research participants saw their situations as chronic. As the financial aftermath of the pandemic increases food insecurity, understanding how food-insecure people make sense of their situations can provide guidance for program revision.
Food insecurity for our participants is part of a larger experience of financial insecurity. The food which participants accessed from St. Francis House and other sources of assistance allowed them to divert more of their meager resources to rent and other non-negotiable expenses. What participants in our study do to make ends meet is consistent with existing research [41,57]. Without this food, some of these families would spend more of their limited earned money on food, perhaps jeopardizing their ability to pay for housing; others would eat less and less well than they already do, possibly risking their and their family members' health. Others have called for broader policy action addressing overall inequality as a response to food insecurity [4]. We endorse this call, for as our participants demonstrate, more latitude for paying other bills will ease their navigation of food resources.
The importance of food assistance to how people in poverty organize their household finances also has implications for program eligibility criteria and regulations. For example, St. Francis House and other Anchorage pantries did not require that clients use only one pantry, and indeed, almost all participants used more than one. This was important for families' stretching their benefits. Pantry users who live in regions with more restrictive policies may well be more desperate.
Increasing the amount of food provided per pantry visit (and increasing SNAP and WIC allotments) is an idea worth exploring insofar as it is remarkable how much time is required to procure free and low-cost food. The amount of food currently provided is based on the assumption that these are emergency programs. Pantry food is expected to help people get through a few bad days, not feed a family for a month. To spend an entire day each month going to pantries, or to devote several mornings to pantries each month, or to visit the store weekly without buying much just to ensure that no sales were missed, all takes extensive time. Allowing people to take more could reduce the number of visits necessary.
Food-specific measures will remain necessary in the U.S., and it is important to note that, although the individual programs have their own histories and rules, they are experienced as a whole. As program managers of public and charitable safety net programs respond to long-term financial effects of the pandemic, they are encouraged to think across systems. It has been argued that charitable food exacerbates poverty and food insecurity because it undermines support for public safety net programs [58][59][60]. Recent policy changes prompted by COVID-19-for example, temporarily expanding SNAP benefits-as well as increased pressure on the charitable food system [61,62], provide an opportunity to re-think the long-term relationship among and between the types of safety nets, and to increase income assistance-not just nutrition assistance-that will make people less dependent on pantries.
Future research, too, might examine the use of multiple programs at once to tease out how they enhance (or undermine) each other in the twin goals of reducing food insecurity and poverty.

Limitations
Arguably, it is a limitation of our study that our sample is not representative of St. Francis House. Our disproportionately white and English-speaking sample may well under-represent the challenges faced by the SFH population. It is possible that the Alaska Native and various immigrant and refugee cultural communities whose members use the pantry but who are mostly missing from the sample provide a stronger personal safety net for their community members. We cannot take this for granted, however, given the disproportionate rates of poverty for these communities. Likewise, sampling from a single pantry limits the extent to which our findings can be generalized in the strict sense to other pantry users in Anchorage and elsewhere. However, in this qualitative study, we seek not to provide a statistically accurate portrait of the population of pantry users or people in poverty, but rather a sense of the experience of being a pantry user. Our findings should be transferable to pantry users in other mid-size cities, who face similar programmatic and environmental opportunities and constraints.
Indeed, our paper reinforces the applicability of findings about how people in urban, developed settings respond to food insecurity. Much of the food security research in Alaska has been conducted in rural Alaska and/or with Alaska Natives peoples, for whom food security is a broader concept linked to traditional foods and cultural practices [63,64]. Food insecurity in urban Alaska with mostly non-Native people has been less studied. Our previous research with a rural, mostly non-Native sample found different patterns than we find here [45]. That residents of a distant, sub-arctic city experienced food insecurity in ways comparable to residents in other U.S. cities suggests that knowledge of patterns can be applied across similar settings.

Conclusions
The participants in our study are thoughtful and intentional in how they use the safety net programs available to them to feed their families. As the aftermath from the COVID-19 pandemic prompts food security program managers and policymakers to re-think their services, we can learn from these pantry users. Thinking about their strategies adds a layer of concreteness to thinking about system improvements: we seek not just to make things "better" in some abstract sense, but to make the lives of these people less stressful, to allow these families to eat as well as they know they ought. Changes should be readily integrated into the strategies already employed.  Institutional Review Board Statement: The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of the University of Alaska Anchorage (protocol code 996827-8, approved 2 February 2017).
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
Transcripts with pseudonyms may be shared on an individual basis upon request. IRB-approved participant consent forms made no provision for public archiving of transcripts.