The Impact of Excluding Food Guarding from a Standardized Behavioral Canine Assessment in Animal Shelters

Transcription

1 animals Article The Impact of Excludg Food Guardg from a Stardized Behavioral Cane Assessment Animal Shelters Hear Mohan-Gibbons 1, Emily D. Dolan 1, Pamela Reid 2, Margaret R. Slater 1, Hugh Mulligan 1 Emily Weiss 3, * 1 Strategy, Research Development, American Society Prevention of Cruelty to Animals (ASPCA ), New York, NY 10018, USA; hear.mohan-gibbons@aspca.org (H.M.-G.); emily.dolan@aspca.org (E.D.D.); margaret.slater@aspca.org (M.R.S.); hugh.mulligan@aspca.org (H.M.) 2 Anti-Cruelty Behavior Team, Anti-Cruelty Group, American Society Prevention of Cruelty to Animals (ASPCA ), New York, NY 10018, USA; Pam.Reid@aspca.org 3 Eque Welfare, Anti-Cruelty Group, American Society Prevention of Cruelty to Animals (ASPCA ), New York, NY 10018, USA * Correspondence: emily.weiss@aspca.org; Tel.: Received: 23 November 2017; Accepted: 5 February 2018; Published: 8 February 2018 Simple Summary: Recent research has called to question value of food guardg assessment as a predictive tool determg safety of shelter dogs. This study examed effect of elimatg food guardg assessment ne U.S. animal shelters. It was found that when food guardg assessment was removed, bites or or juries to staff or adopters did not crease. However, dogs exhibitg food guardg were less likely to be adopted, had a longer shelter stay, were more likely to be euthanized than dogs general population. Based on previous research this study s fdgs, authors recommend that shelters discontue food guardg assessment. Abstract: Many shelters euthanize or restrict adoptions dogs that exhibit food guardg while animal shelter. However, previous research showed that only half dogs exhibitg food guardg durg an assessment food guard home. So, dogs are often misidentified as future food guarders durg shelter assessments. We examed impact of shelters omittg food guardg assessments. Ne shelters conducted a two-month basele period of assessg food guardg followed by a two-month vestigative period durg which y omitted food guardg assessment. Dogs that guarded ir food durg a stardized assessment were less likely to be adopted, had a longer shelter stay, were more likely to be euthanized. When shelters stopped assessg food guardg, re was no significant difference rate of returns of food guardg dogs, even though more dogs were adopted because fewer were identified with food guardg. Additionally, number of juries to staff, volunteers, adopters was low (104 cidents from a total of 14,180 dogs) did not change when food guardg assessment was omitted. These results support a recommendation that shelters discontue food guardg assessment. Keywords: food guardg; shelter assessment; dogs; aggression; animal shelter; euthanasia; SAFER; assessments; ASPCA 1. Introduction Food guardg (FG) is a natural Canids, yet can result adverse consequences when domestic dogs exhibit this animal shelters. FG can occur when a dog has a consumable Animals 2018, 8, 27; doi: /ani

2 Animals 2018, 8, 27 2 of 11 item that it wants to reta. Guardg s, such as growlg snappg towards a perceived competitor, can be seen puppies as young as two to three weeks of age [1]. Most, if not all, stardized assessments currently used shelters [2 4] clude an evaluation of FG. These assessments typically evaluate extent of FG by usg a fake h, made of rubber or plastic on a dowel, to pull food bowl away from dog touch ir cheek while eatg. Different assessments vary time degree of contact between dog fake h. In 2005, The American Society Prevention of Cruelty to Animals (ASPCA ) [5] conducted a survey of US animal shelters. Eighty-ne percent of respondg organizations reported y conducted a assessment ir shelter almost all assessed guardg around a food bowl. For sake of concision this manuscript, specific FG component of assessment is referred to as FG item. Two recent studies dicated that assessg FG shelter was a poor predictor of FG home. Marder et al. [6] looked at 97 dogs adopted from an animal shelter compared ir FG durg assessment -home. All dogs were assessed usg Match-Up II Shelter Dog Rehomg program [4]. Twenty (21%) dogs exhibited FG assessment of those, 11 (55%) showed FG -home. Interestgly, of 77 dogs that did not display FG durg assessment, 17 (22%) did display FG -home. Wher or not dogs guarded food durg shelter assessment had little bearg on FG y displayed -home. The second study [5] followed 96 dogs that exhibited FG durg Safety Assessment Evaluatg Rehomg (SAFER ) assessment [2]. Dogs that exhibited FG were allowed free-access to dry dog food ir kennel, n subsequently adopted to homes. Three weeks after adoption, only one of 96 dogs was reported to exhibit FG home. By three months, no dogs were reported showg FG -home. The discrepancy between se two studies rate of FG -home could be a result of free feedg regimen dogs experienced while shelter, substantially different durations of follow-up, support given to adopters, fundamental differences shelter dog populations, or or unidentified variables. However, conclusions regardg ability of FG item to reliably predict FG -home is consistent across two studies. Both of se studies call to question value of assessg FG while dog is shelter. It appears that dogs showg FG guardg durg assessment are no more or less likely to exhibit FG a home. Equally as important, Marder study [6] found that adopters did not perceive FG as a problem keepg ir dog, which suggests that stardized FG assessments may be unnecessary irrelevant. In 2016, Patronek Bradley [7] considered usefulness of stardized assessments shelters noted that reliably predictg problematic s future adoptive homes is vanishgly unlikely. They argued that if a dog bit or threatened to bite durg a stardized assessment that wher or not dog would do so an adoptive home would be at best, not much better than flippg a co. Although authors were referrg to aggression assessments general, not just FG portion of assessment, ir primary argument is that stardized assessments result an unacceptably high cidence of false positives (which are those dogs that exhibit aggression durg assessment not -home). A high rate of false positives is particularly alarmg when it comes to FG, because FG was one of top reasons shelter staff reported a dog would be deemed unadoptable [5] euthanized [6]. Even when a shelter chose to not euthanize dogs with FG, those dogs were more likely to have adoption restrictions (such as experienced owners only or families without children) reby reducg pool of potential adopters dogs [6]. Sce publication of Marder et al. [6] Mohan-Gibbons et al. [5] studies, members of animal welfare field have shared ir terest concerns with ASPCA about omittg FG item from ir assessment process. Their primary concerns were that juries to staff or adopters would crease that dogs would be returned more often. Thus, primary goals of this study were to compare outcomes, juries, returns after FG item was removed from assessment.

3 Animals 2018, 8, 27 3 of Materials Methods 2.1. Inclusion Criteria In order to be cluded this study, shelters had to already be usg a stardized assessment that cluded a FG item be willg to discontue FG item vestigation phase. The FG item volved presentg dog with a bowl of food n usg a fake h on a dowel to touch dog bowl while dog was eatg food. Only FG toward people was addressed; FG directed towards or dogs was not a part of any stardized assessment this study. Any dog exhibitg FG was cluded, regardless of breed or age (as long as y were at least 6 months old, per assessment guideles). Shelters refraed from makg any or program or adoption changes durg study that would have affected research Recruitment of Shelters Animal shelters were recruited this study through a convenience sample. The authors had prior relationships with some of shelters spoke directly with staff to ask ir participation. Or organizations volunteered mselves study followg a presentation on guardg s shelter given by Drs. Emily Weiss Amy Marder at Society of Animal Welfare Associations conference The authors also sought recommendations from or leaders field to identify organizations that might be terested participatg. Any organization that expressed terest participatg was sent an providg a summary of project. A total of 11 shelters enrolled study. In first month, 2 shelters were excluded: 1 shelter pre-screened dogs FG as part of ir relocation program, renderg m admissible or shelter withdrew because y felt y could not meet data collection requirements. The shelters were all non-profit humane societies except one large municipal animal control organization. These shelters were located seven states: Colorado, Florida, Mississippi, New York, North Carola, South Carola. There was one shelter each of se states except Florida New York, which each had two. Seven shelters used SAFER assessment two shelters used ir own stardized assessment Protocol Ne shelters completed 5-month study from 1 September 2015 to 31 January In first 2 months (basele period), shelters were structed to contue with ir regular process assessg, n y were to omit FG item next 2 months (vestigation phase). In vestigation phase, shelters were advised to conduct all or components of ir assessment (mus FG item) to proceed with adoptions as per normal. A data collection m was issued to each shelter to collect followg mation each month of basele phase vestigation phase entire dog population: (1) total take, (2) total returns of dogs adopted with past 30 days, (3) average length of stay, (4) total number of dogs adopted, transferred, euthanized; total number of bites juries (5) observed while -shelter (6) reported -home post-adoption. Those data were also collected dogs that were classified as food guarders both basele vestigation phases. Durg last month of study, January, only data collected were number of returns reports of bites or juries that occurred -home (#2 #6 above). Bites juries were collected as a sgle category, defed as any type of jury to a human, any context, while shelter or after adoption. This mation was gleaned from reports shelter from conversations when owners contacted shelter to discuss ir dog post-adoption. Shelters were asked to report all cidents any reason, as a result, an cident could occur durg any teraction with dog were not restricted to teractions around food. For example, a dog FG group may have caused harm by pullg someone off ir feet while on a walk was

4 Animals 2018, 8, 27 4 of 11 still counted as an jury, even though no food was volved. This ensured a conservative approach that all cidents were captured this research. On a monthly basis, completed ms were collected from each shelter by . At end of study, authors conducted an mal conference call with all of shelters to learn more about ir experiences durg study Identification of FG Shelters were structed to identify FG one of three ways: (1) durg FG item of assessment basele phase, (2) from terviews conducted with person who brought animal to shelter, or (3) through observations made by staff, volunteers, potential adopters while dog was housed at shelter (separate from assessment). As noted Section 2.3, FG item was permed only durg basele phase but or ways of identifyg FG were utilized over entire study period. Once staff saw a dog exhibitg FG, y recorded on stardized data collection m how that was identified. The researchers provided shelters with descriptions of FG to enable m to categorize level of severity of guardg. These descriptions were labeled as mild, moderate or severe (Appendix A). The s could be seen various contexts cludg durg assessment, ir kennel, or reported from owner. Staff documented level of severity each dog regardless of how was identified Grants Once enrolled study, each shelter was issued a $2500 grant to support work needed to complete study. Grant funds were used to tra staff, create literature on FG adopters staff, start a hotle, or recruit shelter volunteers Statistical Analysis Characteristics of groups were described usg frequencies percentages calculated with Microsoft Excel. Chi-square tests were permed to compare basele vestigation phase rates of bites/juries returns entire general dog population, as well as specifically group of FG dogs. Chi-square was also used to compare percentage of food guarders non-food-guarders euthanized durg four-month study period. StataSE 13.1 (StataCorp LP, College Station, TX, USA) was used se analyses, p < 0.05 was considered to be statistically significant. 3. Results 3.1. Population A total of 14,180 dogs entered ne shelters across four-month study period. The highest take at a participatg shelter was 3247 dogs lowest take was 596 dogs. Durg basele phase, 7112 dogs entered shelters 7068 dogs entered durg vestigation phase. Durg basele phase, only 49% of dogs were assessed upon take. The or dogs were returned to ir owners bee an assessment could be completed or were considered adoptable without beg assessed, due to ir desirable appearance, size, or breed. Over whole study period, 5% (778/14,180) of dogs eir exhibited FG or had a history of FG. Furr details of which dogs were identified with FG durg each phase can be found Figure 1.

5 Animals 2018, 8, 27 5 of 11 Figure 1. Total take of dogs between two study phases: percentage of dogs who were assessed, those dogs identified with food guardg, cidence of jury/bites each phase. Non-FG groups are orange food guardg (FG) groups green, with correspondg p-values displayed between basele vestigative phases. In basele phase, re are blue boxes dotted les that show percentage of dogs that had stardized shelter assessment of those assessed, how many were identified with FG. This was added so reader can visualize large percentage of dogs that did not have a stardize assessment permed to clarify two ways a shelter can arrive at cidence of FG (17% is only of those assessed while 8% is all take) Risk of Injury Figure 1 shows total take of dogs between two study phases color groupgs those who displayed FG (green) vs non-fg (orange). The number of bites/juries was low both phases.

6 Animals 2018, 8, 27 6 of 11 FG dogs, eir shelter or home, were no more likely to flict juries once assessment item was stopped. A with-phase study analysis was also conducted. In home, adopters were significantly more likely to report bites/juries from dogs FG group compared to non-fg group, both basele (p = 0.006) vestigate stages (p = 0.01). In home, re were only 8 reported bites/juries FG group 35 non-fg group. In shelter, re was no association of bites/juries related to FG or non-fg groups, basele (p = 0.5) or vestigative phases (p = 0.6) Outcomes FG Dogs Out of 14,180 dogs enterg shelter, 14% (2051/14,180) were still shelter at termation of study could not be assigned an outcome. Over course of study, 56% (7961/14,180) of all dogs were adopted 18% (2576/14,180) were transferred to or organizations adoption. Fifty-four percent of dogs (3867/7112) were adopted durg basele phase 58% (4094/7068) durg vestigation phase. Adoption of FG dogs was 39% basele (223/571) 45% vestigation phase (93/207) (p = 0.14, X 2 = 2.17, 1 df). Across both study phases, dogs that showed FG stayed shelter longer, a mean of 13.6 days (range = days), than general dog population, a mean of 9.9 days (range = days). A total of 11.2% of dogs (1592/14,180) were euthanized durg study. Significantly more FG dogs (15.4%, 120/778) were euthanized than non-fg dogs (10.9%, 1472/13,402), p < 0.001, X 2 = 0.09, 1 df. Of 120 dogs euthanized FG group, 74% (n = 89) had listed as a reason euthanasia 45% of those had FG listed as a reason euthanasia. The or 55% were euthanized or al issues such as aggression toward or animals or people, high arousal, or problematic kennel. Degree of FG Impact If dogs exhibited FG -shelter, majority of time it was mild severity (Table 1). Anor key fdg is percentage of dogs identified with severe FG was unchanged, even once FG item was omitted. Table 1. Severity of Food guardg (FG) identified all FG dogs comparg basele (8%, 571/7112) vestigation (3%, 207/7068) phases. There was an overall significant difference categories of severity between basele vestigative phase (p = 0.001, X 2 = 14.86, 2 df). In basele, 20% of FG was identified by history 80% by assessment. Durg vestigation phase, 71% of FG was identified by history 29% from shelter observations. Severity of FG Behavior Basele Investigation N % N % Total Dogs with FG Behavior 571 8% 207 3% Mild % % Moderate 85 15% 55 27% Severe 95 17% 35 17% Total % % Figure 2 shows outcomes of adoption, euthanasia, transfers, categorized by severity of FG. Durg both basele vestigation phases, dogs with severe FG were less likely to be adopted more likely to be euthanized than moderate or mild food guarders.

7 Animals 2018, 8, 27 7 of 11 Animals 2018, 8, x 7 of 11 OUTCOMES BY FG SEVERITY [BASELINE PHASE] Adopt Euthanize Transfer No outcome 23% 14% 9% 24% 9% 15% 44% 45% 26% 36% (a) 41% 14% MILD MODERATE SEVERE OUTCOMES BY FG SEVERITY [INVESTIGATION PHASE] Adopt Euthanize Transfer No outcome 7% 27% 12% 54% 11% 22% 29% 29% 9% 37% 40% 23% MILD MODERATE SEVERE (b) Figure 2. Outcomes of FG dogs, by severity of FG, are shown (a) Basele phase Figure 2. Outcomes of FG dogs, by severity of FG, are shown (a) Basele phase (b) Investigation phase. Some dogs each severity category did not have an outcome by end of (b) Investigation phase. Some dogs each severity category did not have an outcome by end study period (no outcome). In basele, 20% of FG was identified by history 80% by of study period (no outcome). In basele, 20% of FG was identified by history 80% by assessment. Durg vestigation phase, 71% of FG was identified by history 29% from shelter assessment. Durg vestigation phase, 71% of FG was identified by history 29% from observations. shelter observations Returns Dogs FG group were no more likely to be returned (p = 0.7, X 2 = 0.11, 1 df) vestigation phase (11%, 10/93) than y were durg basele phase (9%, 21/223). There were significantly

8 Animals 2018, 8, 27 8 of 11 more returns of all dogs (p < 0.001, X 2 = 18.33, 1 df) durg vestigation phase (13%, 526/4094) than durg basele phase (10%, 379/3867). However, even though statistically significant, a difference of 3% returns is too small to be meangful to field. 4. Discussion This study documented impact of omittg FG item from stardized shelter assessments. Durg basele phase, 8% of dogs were classified as food guarders. Durg vestigation phase, when dogs were no longer assessed FG, only 3% were classified as food guarders. Or studies [5,6] reported higher percentages of FG (14 21%), however y looked at number of FG dogs based on number assessed, not on overall take numbers. If this study were to calculate FG dogs based on those assessed (which could only be done durg basele phase), percentage aligns with or studies (16.5%). However, sce many shelters assess less than half ir dog population, se reported percentages likely over-represent prevalence of FG shelter dog population. When designg this study, it was not anticipated that such a large number of dogs would not be assessed durg basele. While it is not known what each dividual staff person did with each shelter (relative to dogs not assessed), shelters did not always assess dogs who were easily adoptable (i.e., healthy, small, obviously social, friendly dogs). It is not clear how much bias this brgs what direction bias would be, due to lack of assessments among a subset of population of dogs. One key factor used to improve validity was each shelter was its own basele. This study was designed to be applied research functiong shelters, constraed by shelter realities. Sce this was a convenience sample, re could be a bias those shelters that chose to participate. The shelters varied size (number of dogs staff), one was a municipal shelter while most were non-profit organizations, y varied ir location around country. We deliberately aimed a large sample size from a variety of shelters to mimize biases results improve external validity. The safety of staff, volunteers, adopters was identified as a critical concern when shelters considered removg FG item from ir assessments. However, few bites or juries were reported durg this study, wher or not dogs were assessed FG. It is terestg that rate of bites juries was so low, even basele phase, when a significant number of dogs were not assessed. This furr confirms conclusion that testg FG does not improve safety a shelter settg. Reported juries home did not crease once FG item was stopped, however re was a higher percentage of juries by FG dogs compared to non-fg dog group. However, re was a low prevalence of bites/juries home all dogs (Figure 1) reportg cluded all bites juries, even when not volvg food. Sce many shelters advised adopters that se dogs might FG home, it is possible that new owners were more likely to report bites or juries thus biasg numbers. It is plausible that dogs who FG shelter may also have or problems that make m more likely to cause juries post-adoption, but this study did not yield that data. Anor limitation of this study, which was also beyond scope, was a lack of systematic follow-up home, like protocol used Mohan-Gibbons study [5]. As a result, we could not ask more mation on situations where bite or jury occurred home. We did not ask shelters to separate actual bites from or types of juries, nor did we ask detail about cidences. Although collectg that mation would have provided more al context, concern was it would troduce unwanted bias because staff adopters would need to correctly terpret every situation if FG was present or not. By trackg all cidences of bites or juries, our conclusions are more conservative. In this study, dogs with food guardg stayed shelter longer than general population were more likely to be euthanized. An earlier survey [5] found that FG was one of top reasons why shelters did not place dogs to adoption. In this study, while only 15.4%

9 Animals 2018, 8, 27 9 of 11 of FG dogs were euthanized, over half of se were euthanized concerns unrelated to FG (i.e., medical problems or severe challenges). Most of dogs that showed FG both basele vestigation phases were classified as mild FG. In vestigation phase, even though fewer dogs were identified with food guardg, percentage of dogs classified as severe remaed unchanged (17% both phases). Thus, it appears that dogs exhibitg more serious FG are likely to be identified without usg a stardized assessment. Relquishg owners may report that ir dog displays FG when it is more serious staff may be more likely to observe severe guardg durg daily teractions shelter. This confirms assertion made by Patronek & Bradley [7] that more serious ms of aggression are highly likely to be identified without puttg dog through a assessment. The animal welfare community had expressed concerns that returns to shelter would crease if FG item was elimated. Our study showed that return rates FG dogs were not significantly different from basele phase, however return rate general population was. Even though those returns presumably cluded dogs that guarded food home, a difference of 3% is not large enough to be meangful context of shelter returns. The statistical significance is driven by large number of dogs adopted from ne shelters. A dog returned from FG group may not have been returned food guardg. The reasons given returns were sufficient to provide any level of detail. It is challengg to gar quality data when those data are based on owners relayg details to shelter personnel who n capture that mation to restrictive shelter software. Future work could consider learng more about reasons dogs beg returned to shelter. If shelters wish to furr reduce returns, we recommend that y have conversations with all adopters about normal cane, such as FG, that any dog might exhibit this home. Staff encourage adopters to call shelter if y encounter problems need help. Or research has shown that when pet owners are given realistic expectations, y are more likely to keep ir pets [8]. Perhaps providg adopters with basic mation about FG offerg post-adoption support will prove to be a key to successful adoptions this at-risk population. Shelters did not make large organizational changes durg course of study. Shelter staff was aware when FG was stopped. We encouraged shelters to talk with owners relquishg dogs to obta a history of FG, some shelters changed ir take ms to reflect this. We encouraged adoption counselors to tell potential adopters that any dog might guard ir food some added this mation to ir adoption paperwork. Many shelters reported that staff became more diligent about recognizg reportg FG vestigative phase because dogs were not gog to be assessed FG. One shelter even uncovered that staff would not report observations of a dog guardg food kennel (durg basele phase) because y believed staff would catch it FG item durg assessment. It is crucial shelters to establish a clear communication system staff volunteers to share pertent mation about dogs ir care, especially when considerg omittg FG item from ir assessment. In previous research, a significant number of dogs that were observed guardg food durg assessment did not exhibit this home [5,6]. These false positive dogs are more likely to stay longer shelter to be euthanized. In addition, some dogs do not guard food while shelter but will at home. These false negative dogs are often correctly deemed as safe by shelter staff sent home with adopters who are not adequately prepared to encounter this. Any dog can exhibit food guardg. Our research, paired with previous research referenced this manuscript, substantiate that existence of food guardg shelter does not always predict same home. The cidence of jury to shelter staff adopters is low assessg dogs FG does not furr reduce already low jury rate.

10 Animals 2018, 8, 8, x of Conclusions Recommendations 5. Conclusions Recommendations Previous research suggested that many dogs exhibitg FG shelter can be safely placed Previous research suggested that many dogs exhibitg FG shelter can be safely placed to homes because FG is often not exhibited home nor was it a problem adopters when it to homes because FG is often not exhibited home nor was it a problem adopters when was seen. In light of current results, authors recommend that shelters need not conduct it was seen. In light of current results, authors recommend that shelters need not conduct FG item of a stardized assessment because it results a longer stay shelter, creased FG item of a stardized assessment because it results a longer stay shelter, creased likelihood of euthanasia, cidence of false positives is likely unacceptably high. FG was not likelihood of euthanasia, cidence of false positives is likely unacceptably high. FG was not prevalent participatg ne shelters, of dogs that did exhibit, most prevalent participatg ne shelters, of dogs that did exhibit, most showed showed only mild FG. Once FG item was omitted, more of those dogs were adopted only mild FG. Once FG item was omitted, more of those dogs were adopted fewer fewer were euthanized. Severe FG was found regardless of assessment. Discontug were euthanized. Severe FG assessment FG item was did found not result regardless an of crease assessment. juries Discontug to shelter staff FG or adopters, assessment item re did was not no result meangful an crease change juries returns to post-adoption. shelter staff or adopters, re was no meangful Our recommendation change returns is post-adoption. shelters to collect as much mation about dog from all sources Our (such recommendation as prior owner is staff shelters observations) to collect to as best much support mation each dog about fdg dog a home. from For all example, sources (such if a dog as shows prior owner guardg staff observations) over his food to while best support kennel, each dog shelter fdg a home. have procedures For example, to if ensure a dog shows staff are guardg traed so that y over his can food stay while safe while teractg kennel, shelter with dog. However, have procedures y to ensure not assume staff are traed dog will so that show y this can stay safe while his new teractg home. Shelters with dog. not However, ignore y FG. not Rar, assume staff dog will be show transparent this with adopters his new about home. what Shelters y know from not ignore prior FG history. Rar, staff -shelter. Staff be transparent set adopters with adopters up about success what by y explag know from that FG prior is a history normal -shelter. dogs, it has Staff a low occurrence, set adopters if up seen, y success by explag contact that shelter FG promptly is a normal assistance. There dogs, are it has techniques a low occurrence, managg if seen, modifyg y FG contact shelter home when promptly owners assistance. are both There concerned are techniques as well as able managg to implement modifyg se techniques. FG home when owners are both concerned as well as able to implement se techniques. Acknowledgments: The authors deeply appreciate followg shelters ir participation study Acknowledgments: The authors deeply appreciate followg shelters ir participation study ir ir time time data data collection collection process; process; Humane Humane Society Society of Boulder of Boulder Valley Valley (CO), (CO), SPCA SPCA Tampa Tampa Bay Bay (FL), (FL), Pet Alliance Pet Alliance of Greater of Greater Orlo Orlo (FL), (FL), Sourn Sourn Pes Pes Animal Animal Shelter Shelter (MS), (MS), SPCA SPCAof ofwake County (NC), Nebraska Humane Society (NE), Erie County SPCA (NY), Animal Care Centers of New York (NY), Charleston Animal Society (SC). Processg costs this paper were covered by Special Issue sponsors, Maddie s Fund, Found Animals, The Humane Society of United States (HSUS). Author Contributions: Emily Weiss, Margaret R. Slater, Pamela Reid, Hear Mohan-Gibbons, Hugh Mulligan Author Contributions: Emily Weiss, Margaret R. Slater, Pamela Reid, Hear Mohan-Gibbons, Hugh Mulligan conceived conceived designed designed project; project; Margaret Margaret R. R. Slater, Slater, Hugh Hugh Mulligan, Mulligan, Hear Hear Mohan-Gibbons Mohan-Gibbons analyzed analyzed data; Hear Mohan-Gibbons, Emily EmilyD. D. Dolan, Dolan, Margaret Margaret Slater, Slater, Emily Emily Weiss, Weiss, Pamela Pamela Reid Reid wrote wrote manuscript. Conflicts of Interest: The authors declare no conflict of terest. Appendix A Below are three categories with descriptions of guardg s that may be beobserved durg food portion of of an an assessment. Please Pleasechoose choose broad broadcategory that thatis is best bestfit fit each eachdog. Vocal s, such as growlg or barkg, are not listed, as y can be observed any category. Figure A1. Behavioral Categories Food Guardg. Figure A1. Behavioral Categories Food Guardg. A.1. Mild Appendix A.1. Mild Dog s head, neck /or body become stiff/rigid but dog does not escalate to Moderate or Severe Dog s categories. head, neck May /or show one body or more become of se stiff/rigid s: but dog does not escalate to Moderate or Severe categories. May show one or more of se s: eats more quickly as food assessment progresses holds eats more a crouched quicklyposition as food assessment over bowl when progresses person is near moves holds abody crouched /or position head to over block bowl when bowl person from is assessor s near approach or fake h momentarily freezes or stops eatg (without liftg head from bowl) when assessor approaches or reaches with fake h

11 Animals 2018, 8, of 11 moves body /or head to block bowl from assessor s approach or fake h momentarily freezes or stops eatg (without liftg head from bowl) when assessor approaches or reaches with fake h Appendix A.2. Moderate The guardg remas or near food bowl. Dogs may show any of s Mild category plus one or more of se s: may lift lips/bare teeth bites direction of food or at fake h while head is bowl /or while touchg dog s cheek with fake h or person is near although dog may bite fake h, dog does not lift his head from bowl to bite toward assess-a-h or pursue h as it is retracted Appendix A.3. Severe The guardg is directed towards assessor or hler. Dogs may show any of Mild or Moderate s, plus at least one of followg: multiple bites to fake h /or bites that move up fake h leaves bowl to bite fake h (eir while h is advancg towards bowl or as it is withdrawn) any attempts to lunge, snap, or bite towards person s arm or body or hler durg food bowl item dogs that exhibit guardg eir bee or after food guardg assessment ( example, bowl is placed down dog immediately shows any guardg bee assessor has attempted ir itial approach to bowl or dogs that contue to guard area after food bowl is removed). References 1. Scott, J.P.; Fuller, J. The Social Behavior of Dogs Wolves. In Genetics Social Behavior of Dog; University of Chicago Press: Chicago, IL, USA, 1965; pp ISBN American Society Prevention of Cruelty to Animals. Safety Assessment Evaluatg Rehomg (SAFER ). Available onle: (accessed on 29 December 2017). 3. Assess-a-Pet. Available onle: (accessed on 29 December 2017). 4. Animal Rescue League of Boston. Match-Up II Shelter Dog Rehomg program. Available onle: (accessed on 29 December 2017). 5. Mohan-Gibbons, H.; Weiss, E.; Slater, M. Prelimary Investigation of Food Guardg Behavior Shelter Dogs United States. Animals 2012, 2, [CrossRef] [PubMed] 6. Marder, A.R.; Shabelansky, A.; Patronek, G.J.; Dowlg-Guyer, S.; D Arpo, S.S. Food-related aggression shelter dogs: A comparison of identified by a evaluation shelter owner reports after adoption. Appl. Anim. Behav. Sci. 2013, 148, [CrossRef] 7. Patronek, G.J.; Bradley, J. No better than flippg a co: Reconsiderg cane evaluations animal shelters. J. Vet. Behav. 2016, 15, [CrossRef] 8. Houpt, K.A.; Honig, S.U.; Reisner, I.R. Breakg human-companion animal bond. J. Am. Vet. Med. Assoc. 1996, 208, [PubMed] 2018 by authors. Licensee MDPI, Basel, Switzerl. This article is an open access article distributed under terms conditions of Creative Commons Attribution (CC BY) license (