Роль эмоционального состояния в принятии решения о пожертвовании тема диссертации и автореферата по ВАК РФ 00.00.00, кандидат наук Шепеленко Анна Юрьевна

• Шепеленко Анна Юрьевна
• кандидат наук
• 2024

Оглавление диссертации кандидат наук Шепеленко Анна Юрьевна

Введение

Актуальность исследования

Проблема исследования

Состояние разработанности проблемы исследования

Цель и задачи исследования

Теоретико-методологические основы исследования

Методы исследования

Выборка и эмпирическая база исследования

Теоретическая и практическая значимость исследования

Основные результаты исследования и положения, выносимые на защиту

Основные результаты исследования

Положения, выносимые на защиту

Заключение

Апробация результатов исследования

Список использованной литературы

Приложение

Приложение 2. Статья «How Emotions Induce Charitable Giving: A Psychophysiological Study»

Приложение 3. Статья «The relationship between charitable giving and emotional facial expressions: Results from affective computing»

Приложение 4. Статья « How the emotions evoked by homeless pets induce online charitable giving»

Введение

Введение диссертации (часть автореферата) на тему «Роль эмоционального состояния в принятии решения о пожертвовании»

Актуальность исследования

Благотворительные организации существенно влияют на социально-экономические процессы в обществе, реализуя программы, направленные на улучшение различных сторон жизни населения [1]. Являясь инструментом продвижения субъективного благополучия общества [2], благотворительные организации сами сильно зависят от поддержки населения. Например, в Соединенных Штатах Америки за 2018 год около 70% всех пожертвований в благотворительные организации были сделаны частными лицами [3].

В условиях экономической нестабильности и глобальных проблем, с которыми сталкиваются люди во всем мире, смягчающий эффект, оказываемый некоммерческими организациями (НКО), имеет решающее значение. Однако сбор средств для исполнения своих обязательств является сложной задачей для многих благотворительных организаций [4].

Фонды, занимающиеся помощью бездомным животным, оказались в особенно сложной ситуации, поскольку они сталкиваются с недостатком финансирования не только из-за внутриотраслевой конкуренции, но также из-за неоднозначного отношения в обществе к их подопечным [5]. В частности, бродячие собаки часто воспринимаются как угроза из-за распространения ими зоонозных заболеваний [6-8] или возможного агрессивного поведения [9], что часто приводит к жестокому обращению с этими животными со стороны людей [5, 10]. В этом отношении особенно важна роль благотворительных фондов, так как они помогают решить проблему бродячих собак гуманными методами, такими как стерилизация, вакцинация, пропаганда ответственного владения и т. д. [11, 12].

Изучение причин, лежащих в основе процесса принятия решения о пожертвовании, может оказать дополнительную поддержку некоммерческому сектору в целом и благотворительным фондам помощи бездомным животным в частности. Однако просоциальное поведение, то есть готовность отдавать свои

ресурсы ради блага другого [13], является сложным психологическим явлением [14], что затрудняет его изучение и иногда приводит к противоречивым выводам.

Эмоциональное состояние — один из ключевых факторов, влияющих на благотворительное поведение [15-17]. Среди эмоций, влияющих на готовность сделать пожертвование, выделяют печаль, гнев, страх и радость [17-20]. Однако эффективность использования конкретных дискретных эмоций в контексте призывов к благотворительности до сих пор остается предметом дискуссий. Противоречия в результатах исследований касаются не только типа эмоций, но и их валентности (знака). В частности, некоторые исследования свидетельствуют о том, что отрицательные (неприятные) эмоции более эффективно привлекают пожертвования [21, 22], тогда как другие показывают обратное [23, 24]. В то же время исследований связи между интенсивностью (силой) эмоций, вызванных призывами к благотворительности, и готовностью пожертвовать представлено недостаточно.

В целом исследования эмоций и просоциального поведения показали значительный уклон в сторону внутривидовой помощи, а не межвидовой. Другими словами, большая часть научных публикаций посвящена призывам к благотворительности в поддержку людей [17, 21-25], в то время как влияние эмоций на просоциальность по отношению к животным остается недостаточно изученным.

Что касается методологических подходов, следует отметить, что в большинстве исследований благотворительного поведения и эмоций используются декларативные методы (опросы, шкалы отношения), которые имеют определенные ограничения, поскольку участники не всегда способны осознавать свои чувства или не готовы их раскрывать. Особенно это касается спорных и табуированных тем, таких как сексуальная ориентация, венерические заболевания, поддержка социально отвергаемых групп населения [26], а также может быть применимо и к бродячим собакам, так как отношение к ним зачастую неоднозначное [5].

Использование объективных методов измерения эмоционального состояния

может помочь в преодолении ограничений, связанных с субъективной оценкой

5

эмоций. Одним из таких методов являются программы автоматического распознавания эмоций на основе лицевой экспрессии, такие как FaceReader, iMotions и другие, также называемые аффективными вычислениями [27, 28]. Данный метод позволяет комплексно и дискретно оценить эмоциональное состояние благодаря возможности в режиме реального времени распознавать шесть базовых эмоций (радость, печаль, страх, гнев, удивление и отвращение), а также знак и силу эмоций.

Электромиография (ЭМГ) лицевых мышц также является объективным и широко распространенным методом оценки эмоционального состояния [29], который может быть эффективен для валидации данных относительно знака эмоций, полученных с помощью аффективных вычислений.

Для обеспечения наиболее точной оценки эмоционального состояния в дополнение к регистрации активности лицевых мышц используют измерения вегетативных реакций, таких как электрическая активность кожи (ЭАК) и частота сердечных сокращений (ЧСС). Данные методы являются наиболее чувствительными для оценки силы эмоций [30, 31] и тоже могут быть использованы для валидации данных аффективных вычислений.

Вместе с тем в научной литературе наблюдается нехватка исследований, изучающих эффективность призывов к благотворительности объективными методами. В частности, отсутствуют исследования привлекательности некоммерческой рекламы с использованием систем автоматического распознавания эмоций. Также не было обнаружено исследований благотворительного поведения в поддержку бездомных животных, основанных на физиологических данных (ЭМГ, ЭАК, ЧСС), хотя, учитывая неоднозначное отношение к бродячим собакам в обществе, именно в этом направлении объективные методы могли бы помочь сделать вклад в изучение просоциального поведения.

Таким образом, актуальность данного исследования обусловлена

потребностью некоммерческих организаций в увеличении финансирования их

деятельности за счет разработки эффективных маркетинговых коммуникаций,

6

направленных на сбор пожертвований. Следовательно, имеется необходимость в создании теоретико-методологической базы исследований взаимосвязи просоциального поведения и эмоций, основанной на объективных методах оценки эмоционального состояния (включая аффективные вычисления, ЭАК, ЭМГ лицевых мышц и регистрацию сердечной деятельности), которая может быть использована для прогнозирования успешности благотворительной рекламы НКО, в особенности фондов помощи животным.

Проблема исследования

Данное исследование направлено на изучение взаимосвязи эмоционального состояния и благотворительного поведения в поддержку бездомных животных. Кроме того, отдельное внимание уделяется изучению характеристик бездомных животных, таких как возраст, состояние здоровья, признаки бездомности и присутствие человека поблизости, как факторов, влияющих на эмоциональное состояние и размер пожертвования.

Для изучения данной проблемы используются методы регистрации поведенческих и психофизиологических проявлений эмоций (самоотчет, аффективные вычисления, ЭМГ, ЭАК, ЧСС), что позволяет комплексно и объективно оценить роль эмоций в процессе принятия решения о пожертвовании.

Состояние разработанности проблемы исследования

Эмоции и просоциальное поведение

Оценка эмоционального состояния широко используется для изучения различных типов поведения и принятия решений. Так, в контексте просоциального поведения эмоциональное состояние влияет на готовность отдавать свои ресурсы другому [15-17]. Было показано, что пожертвования связаны с эмоциями, концептуализированными как в категориальном [32-34], так и в многомерном подходе [35, 36].

Вместе с тем до сих пор неясно, какие именно эмоции наиболее эффективны

в контексте благотворительного поведения. Так, некоторые дискретные эмоции в

зависимости от контекста могут инициировать или, наоборот, подавлять

7

просоциальные действия. В частности, при просмотре социальной рекламы с изображением печальной жертвы под воздействием эмоционального заражения усиливается склонность к благотворительности [17]. Схожие результаты были получены, когда потенциальными бенефициарами помощи были домашние животные: люди готовы были пожертвовать больше, когда видели рисунки с печальной собакой, чем когда видели веселую [35]. С другой стороны, в контексте медицинской помощи печаль и размер пожертвования имеют криволинейную зависимость, т.е. умеренное выражение печали в призывах к благотворительности увеличивает готовность пожертвовать, а чрезмерная печаль, наоборот, снижает [37]. В то же время, когда благотворительность направлена на образование и оказание помощи при стихийных бедствиях, печаль положительно и линейно связана с пожертвованиями [37].

Столь же неоднозначные эффекты возникают, когда призывы к просоциальному поведению вызывают страх. В контексте экологии Хайн и Гиффорд [18] показали, что вызывающие страх сообщения о загрязнении окружающей среды больше способствуют благотворительному поведению, чем неаффективные сообщения. Однако, когда целью сбора средств была борьба с изменением климата, вызывающие страх призывы не оказали существенного влияния на пожертвования [38]. Пугающие сообщения также оказались неэффективными в контексте сбора онлайн-пожертвований для малоимущих и оказания помощи при стихийных бедствиях [37].

Другие чувства и эмоции, такие как гнев, шок, тревога и сочувствие, в зависимости от контекста и дизайна исследования также показывают разную эффективность для привлечения пожертвований [37, 39-42]. Таким образом, на данный момент в научной литературе нет единого представления о том, как именно дискретные эмоции влияют на готовность сделать пожертвование, что во многом обусловлено различием целей, на которые направлена благотворительность.

Противоречия в результатах исследований касаются не только конкретных

эмоций, но и их знака. Например, Берт и Стронгмен [21] обнаружили, что

изображения, демонстрирующие отрицательные эмоции, увеличивают количество

8

пожертвований, включая деньги, предметы и время. Кроме того, исследование ЭЭГ показало, что реклама с отрицательной валентностью вызывает большее внимание к потенциально угрожающим стимулам, о чем свидетельствует амплитуда вызванного потенциала Р1 [43].

В то же время частое использование отрицательных эмоций в призывах к благотворительности подвергается критике [44, 45] и может привести к неприязни по отношению к благотворительным организациям [46]. На этом фоне появляется все больше доказательств того, что положительные (приятные) эмоции могут быть эффективным инструментом для привлечения пожертвований [24, 47]. В частности, данные функциональной магнитно-резонансной томографии (фМРТ) показали, что участники эксперимента были более склонны делать пожертвования, когда стимульные изображения повышали активность прилежащего ядра — области мозга, связанной с положительным возбуждением [24]. В контексте аффективных вычислений также было показано, что пожертвования положительно коррелируют с радостью и валентностью, зарегистрированной перед принятием решения о пожертвовании [23].

Особое внимание следует уделить влиянию силы эмоций на пожертвования. Результаты, полученные с помощью айтрекинга, показали, что по мере увеличения силы эмоций увеличивается и интерес к просмотру благотворительной рекламы, что, в свою очередь, может увеличить вероятность пожертвования [48]. Согласно опросу, проведенному Амато [39] после крупного пожара в Австралии, люди, испытавшие более высокий уровень эмоционального возбуждения, например ужас или шок, чаще делали пожертвования в благотворительные организации и проявляли готовность помочь жертвам пожара [39]. Однако, насколько нам известно, связь между размером пожертвования и силой эмоций изучена недостаточно.

Таким образом, можно сделать вывод, что результаты исследований взаимосвязи эмоций и благотворительного поведения противоречивы. В частности, одни исследования показали эффективность положительных эмоций при сборе

пожертвований, тогда как другие выявили обратное. Роль отдельных дискретных

9

эмоций в целом также остается неясной, поскольку различные исследования

показали эффективность целого ряда эмоций, что может быть связано с различиями

в исследовательских парадигмах и условиях, при которых принимались решения о

пожертвованиях. Также остается недостаточно изученным влияние силы эмоций на

благотворительное поведение. Следовательно, имеются предпосылки для

дальнейших исследований взаимосвязи эмоций и благотворительного поведения.

Влияние характеристик животных на принятие решения о пожертвовании

Учитывая многообразие деятельности благотворительных организаций,

важно понимать специфику каждой из них. Однако на сегодняшний день

наблюдается нехватка академических публикаций, посвященных просоциальному

поведению людей, обусловленному призывами к благотворительности

(социальной рекламой) в поддержку животных. Фонды помощи бездомным

животным часто используют в рекламных сообщениях фотографии своих

подопечных для привлечения внимания. Условия содержания и состояние

представленных животных могут быть связаны с благотворительным поведением,

поскольку способны вызвать определенное эмоциональное состояние, а также

продемонстрировать проблему, для решения которой требуются пожертвования.

Возраст животного может являться одним из факторов, влияющих на

благотворительное поведение. Ранее было показано, что готовность сделать

пожертвование в пользу детей выше, чем в пользу взрослых [49]. Это может быть

связано с тем, что маленькие дети, а также щенки или котята воспринимаются как

более уязвимая категория по сравнению со взрослыми людьми или особями

животных. Кроме того, вполне вероятно, что щенки могут провоцировать у людей

паттерны родительского поведения по аналогии с младенцами [50]. Таким образом,

вероятно, объем пожертвований на поддержку щенков будет больше, чем на

поддержку взрослых собак.

Состояние здоровья и признаки бездомности также могут быть факторами,

влияющими на готовность пожертвовать деньги. Так, было показано, что

изображения, демонстрирующие страдания и печаль людей, вызывают

высочайший уровень сострадания [51], что также может проявляться и в

10

отношении животных. Таким образом, вполне вероятно, что изображения бездомных или больных собак наглядно показывают проблемы, на решение которых направлен сбор средств, а также вызывают больше сочувствия, что, в свою очередь, увеличивает размер пожертвования.

Отсутствие потенциального опекуна — еще один признак, который может подчеркнуть тяжелое положение животного. Вполне вероятно, что изображения собак, рядом с которыми присутствуют люди, могут спровоцировать эффект свидетеля или диффузию ответственности [52, 53], что может снизить вероятность пожертвований в поддержку таких животных по сравнению с одинокими.

Следовательно, такие характеристики собак, как состояние здоровья, возраст, признаки бездомности и присутствие человека поблизости, могут быть связаны с размером пожертвований.

Цель и задачи исследования

Цель исследования — изучить взаимосвязь между эмоциональным состоянием и размером пожертвования в поддержку бездомных собак, а также определить, как характеристики состояния животных связаны с эмоциями и благотворительным поведением.

К задачам данного исследования относятся следующие теоретические, методологические и эмпирические задачи:

— проанализировать основные теоретические подходы к изучению эмоций, а также результаты эмпирических исследований, посвященных взаимосвязи эмоций и просоциального поведения;

— проанализировать результаты эмпирических исследований, посвященных изучению характеристик получателей помощи, влияющих на просоциальное поведение;

— подготовить надежный инструментарий для исследования взаимосвязи эмоций и размера пожертвования в поддержку бездомных собак;

— подготовить надежный инструментарий для исследования взаимосвязи характеристик бездомных собак и размера пожертвования в их поддержку;

— собрать данные лабораторных и онлайн-эксперимента, направленных на изучение взаимосвязи размера пожертвований с эмоциями и характеристиками бездомных собак;

— проанализировать собранные данные на предмет наличия взаимосвязи между размером пожертвования и эмоциями;

— проанализировать собранные данные на предмет наличия взаимосвязи между размером пожертвования и характеристиками бездомных собак;

— сравнить полученные результаты с данными ранее опубликованных эмпирических исследований, посвященных взаимосвязи эмоций и благотворительного поведения;

— сравнить полученные результаты с данными ранее опубликованных эмпирических исследований, посвященных взаимосвязи характеристик получателей помощи и благотворительного поведения;

— систематизировать полученные результаты и подготовить основные выводы относительно взаимосвязи размера пожертвования с эмоциями и характеристиками бездомных собак.

Теоретико-методологические основы исследования

Теоретическую и методологическую базу исследования составили работы по следующим направлениям:

— подходы к описанию эмоций (Дж. А. Расселл, Л. Ф. Барретт, П. Экман, У. В. Фризен);

— эмпирические исследования взаимосвязи эмоций и просоциального поведения (П. М. Гомер, Д. А. Смол, Н. М. Веррочи, С. Д. Б. Берт, К. Стронгмен, Ц. Лян, Ц. Чен, А. Женевский, Б. Кнутсон, П. Р. Амато, Х. Сабато, Т. Когут);

— измерение эмоций с помощью регистрации активности лицевых мышц (П. Дж. Ланг, М. К. Гринволд, М. М. Бредли, А. О. Хамм, И. Б. Мосс, М. Д. Робинсон, У. Димберг, Б. Карлссон);

— измерение эмоций на основе реакций вегетативной нервной системы (П. Дж. Ланг, М. К. Гринволд, М. М. Бредли, А. О. Хамм, Х. П. Санчес-Наварро, Х. М. Мартинес-Сельва);

— распознавание эмоций по лицевой мимике, аффективные вычисления (П. Экман, У. В. Фризен, В. Хадинежаде, Л. Лойенс, О. Крипс, Д. Мэйсон, Б. Павловска, Б. Д. Мойл).

Методы исследования

В ходе данного исследования было проведено два лабораторных эксперимента (первый — с регистрацией периферических физиологических данных, второй — с распознаванием лицевой мимики с помощью программы FaceReader) и один поведенческий онлайн-эксперимент. Каждый из трех экспериментов состоял из двух частей. В первой части участники совершали добровольные пожертвования в поддержку бездомных собак, а во второй — оценивали свои эмоции, вызванные изображениями животных. Все процедуры исследования проводились в соответствии с Хельсинкской декларацией и были одобрены комиссией по внутриуниверситетским опросам и этической оценке эмпирических исследовательских проектов НИУ ВШЭ.

Задание на пожертвование

Данное задание было схожим для всех трех экспериментов. Испытуемые получали 320 рублей за участие в исследовании, которые они могли добровольно пожертвовать на нужды собак, представленных на изображениях, из расчета от 0 до 10 рублей на каждое из 32 животных. Сумму пожертвования испытуемые определяли самостоятельно, используя шкалу, появляющуюся после каждого изображения. Рядом со шкалой для выбора размера пожертвования (от 0 до 10 рублей) участники могли видеть номер изображения (например, 3 из 32) и остаток денежных средств от суммы вознаграждения (остаток = 320 руб. - сумма всех ранее сделанных пожертвований). Итоговый остаток после предъявления всех изображений испытуемые могли забрать себе. Все собранные пожертвования были

переданы в фонд помощи бездомным животным, о чем участники были проинформированы заранее.

Методы оценки эмоций

В данном исследовании был использован комплексный подход для оценки эмоций в процессе принятия решения о пожертвовании, включающий в себя как субъективные (самоотчет эмоций), так и объективные методы.

Самоотчет эмоций

Во второй части эксперимента участники повторно просматривали 32 изображения животных и давали оценку своим эмоциям, используя 9-балльную шкалу. В первом лабораторном и в онлайн-эксперименте участники оценивали только знак и силу эмоций. Во втором лабораторном эксперименте участники оценивали знак и силу эмоций, а также шесть базовых эмоций. При оценке знака эмоций на 9-балльной шкале значение «1» означало очень неприятные эмоции, а значение «9» — очень приятные. При оценке дискретных эмоций и силы эмоций «1» означало очень слабые эмоции, «9» — очень сильные эмоции. Объективные методы оценки эмоций Эксперимент 1

В рамках данного эксперимента для оценки эмоционального состояния

участников использовались следующие методы:

1. Электромиография лицевых мышц: большой скуловой мышцы (zygomaticus

major) и мышцы, сморщивающей бровь (corrugator supercilii). ЭМГ соп-ugator

supercilii является достоверным маркером отрицательных эмоций, тогда как

ЭМГ zygomaticus major отражает положительные эмоции [54, 55].

Регистрация ЭМГ-активности мышц осуществлялась с помощью

биполярного размещения 4-мм поверхностных электродов Ag/AgCl [56] на

правой стороне лица. Необработанный сигнал ЭМГ был усилен, а частоты за

пределами диапазона 10-350 Гц отфильтрованы. Сигнал также подвергался

двухполупериодному выпрямлению и интегрированию в автономном

режиме с постоянной времени 500 мс. Для анализа брали среднее значение

ЭМГ-активности за 1 секунду до предъявления стимула и вычитали его из

14

среднего значения ЭМГ-активности, зарегистрированного в течение всего периода предъявления стимула — 6 секунд. Такой подход позволил нейтрализовать влияние на результаты эмоционального состояния испытуемых до предъявления стимула [57].

2. Регистрация электрической активности кожи. Проводимость кожи коррелирует с силой эмоций, поэтому этот метод чувствителен к аффективным стимулам любой валентности [54, 58]. Данные ЭАК регистрировали с помощью биполярных поверхностных электродов Ag/AgQ, размещенных на втором и четвертом пальцах левой руки (недоминантной для всех испытуемых). Необработанный сигнал был настроен на обнаружение активности в диапазоне 0-100 мкс. ЭАК рассчитывали как величину всех пиков, появляющихся в интервале от 0,9 до 4 секунд после начала предъявления стимула. Данные были преобразованы (log10[SCR + 1]) для исправления искаженного распределения [30, 57].

3. Фотоплетизмограмма использовалась для регистрации частоты сердечных сокращений. Данный параметр чувствителен при измерении силы эмоций. Было показано, что стимулы, вызывающие сильные эмоции (как приятные, так и неприятные), замедляют ЧСС в первые секунды предъявления больше, чем нейтральные [54, 59]. Фотоплетизмограф регистрировал амплитуду объема крови в пальце с помощью фотоэлемента, расположенного на дистальной фаланге третьего пальца левой руки (недоминантной для всех испытуемых). ЧСС в ударах в минуту (уд/мин) рассчитывали по пикам амплитуды, а замедление ЧСС считали как разницу между ежесекундной ЧСС и базовой линией (1 секунда до предъявления стимула). Для анализа брали максимальное замедление от базовой линии в первые 3 секунды (1-3 сек) и в последние 3 секунды (4-6 сек) предъявления стимула [57, 60].

Данные методы использовались для измерения эмоций в первой части

эксперимента, т.е. в момент первого предъявления стимула, непосредственно перед принятием решения о пожертвовании. Физиологические сигналы регистрировали,

усиливали и фильтровали с помощью системы сбора данных ActiChamp (Brain Products, Германия) [57].

Эксперимент 2

В рамках данного эксперимента для оценки эмоционального состояния участников использовались аффективные вычисления — FaceReader (версия 8.0, Noldus Information Technology). Данная программа имеет один из самых высоких показателей точности распознавания среди восьми наиболее распространенных программ автоматического распознавания эмоций [27]. FaceReader одновременно оценивает два измерения эмоций (силу и знак), шесть базовых эмоций (радость, гнев, отвращение, страх, печаль и удивление) и «нейтральное» состояние. Значения знака эмоций варьируются от -1 (крайне отрицательные эмоции) до 1 (крайне положительные эмоции). Значения силы эмоций, а также шести основных эмоций лежат между 0 (неактивно) и 1 (активно). Данные FaceReader записывались с частотой дискретизации 10 Гц и разрешением 720 пикселей (1280*720). Для анализа брали среднее значение по каждой эмоции в течение 1 секунды до предъявления стимула и вычитали его из среднего значения эмоции в течение всего периода предъявления стимула (6 секунд) [61].

Данный метод аффективных вычислений использовался для оценки эмоций в первой части эксперимента, т.е. в момент первого предъявления стимула, непосредственно перед принятием решения о пожертвовании.

Эксперимент 3

Данный эксперимент проходил онлайн для валидации данных, полученных в ходе лабораторной части исследования, чтобы исключить влияние лабораторных условий на принятие решения о пожертвовании. Таким образом, в данном эксперименте был использован только субъективный метод — самоотчет силы и знака эмоций.

Характеристики стимулъного материала

В качестве стимульного материала использовались 32 фотографии собак,

которые потенциально могли влиять на эмоциональное состояние испытуемых.

Фотографии животных были отобраны так, чтобы они были похожи на типичные

16

материалы фондов помощи бездомным животным, призывающие к благотворительности. Каждое изображение содержало четыре бинарные характеристики состояния собаки, которые были определены как потенциально значимые при принятии решения о пожертвовании: состояние здоровья (здоровая/больная), признаки бездомности (домашняя/бездомная), возраст (щенок / взрослая особь) и присутствие рядом человека (с человеком / без человека). Стимульные материалы были предварительно протестированы в рамках онлайн-опроса (N=17) на соответствие изображений указанным характеристикам (уровень согласованности ответов респондентов по каждому изображению в среднем по всем четырем характеристикам составлял 84-100%). Предварительное тестирование стимулов на уровень силы и знака эмоций, а также шесть базовых эмоций не проводилось. Общий набор картинок содержал 16 категорий с комбинациями четырех характеристик, по две картинки на каждую категорию. Таблица характеристик каждой категории представлена в дополнительных материалах (Приложение 1).

Список литературы диссертационного исследования кандидат наук Шепеленко Анна Юрьевна, 2024 год

Список использованной литературы

1. Ott J.S., Dicke L. The Nature of the Nonprofit Sector // The Nature of the Nonprofit Sector. 4th ed. / ed. Ott J.S., Dicke L.A. New York: Routledge, — 2021. 1-604 p.

2. Ressler R.W., Paxton, P., Velasco, K., Pivnick, L., Weiss, I., & Eichstaedt, J. C. Nonprofits: A Public Policy Tool for the Promotion of Community Subjective Well-being // Journal of Public Administration Research and Theory. — 2021. — Vol. 31, № 4. — P. 822-838.

3. Giving USA [Electronic resource] // Giving USA 2019 : Americans gave $427.71 billion to charity in 2018 amid complex year for charitable giving. 2019. URL: https://givingusa.org/giving-usa-2019-americans-gave-427-71-billion-to-charity-in-2018-amid-complex-year-for-charitable-giving/ (date of the application: 16.09.2023).

4. Gratton P.C. Organization development and strategic planning for non-profit organizations // Organization Development Journal. — 2018. — Vol. 36, № 2. — P. 27-38.

5. Whyke T.W., Lopez-Mugica J. Content and Discourse Analysis of Cruelty towards Stray Dogs as Presented in Chinese Social Media // Society & Animals. — 2020. — Vol. 31, № 2. — P. 139-158.

6. Kalthoum S., Ben Salah C., Rzeigui H., Gharbi R., Guesmi K., Ben Salem A., Ferchichi S., Zammel F., Fatnassi N., Bahloul C., Seghaier C. Owned and free-roaming dogs in the North West of Tunisia: estimation, characteristics and application for the control of dog rabies // Heliyon. — 2021. — Vol. 7, № 11. — P. e08347.

7. Parry N.M.A. COVID-19 and pets: When pandemic meets panic // Forensic Science International: Reports. — 2020. — Vol. 2. — P. 100090.

8. Razali K., Kaidi R., Abdelli A., Menoueri M.N., Ait-Oudhia K. Oral flora of stray dogs and cats in Algeria: Pasteurella and other zoonotic bacteria // Vet World. — 2020. — Vol. 13, № 12. — P. 2806-2814.

9. Katica M., Gradascevic N., Hadzimusic N., Obradovic Z., Mujkanovic R., Mestric E., Coloman S., Dupovac M. Widespread of Stray Dogs: Methods for Solving the Problem in Certain Regions of Bosnia and Herzegovina // International Journal of Research -GRANTHAALAYAH. — 2017. — Vol. 5, № 6. — P. 414-422.

10. Gupta N., Gupta R.K. Animal Welfare and Human Health: Rising Conflicts over Stray Dogs in Chandigarh // South Asia Research. — 2019. — Vol. 39, № 3. — P. 339-352.

11. Mota-Rojas D., Calderón-Maldonado N., Lezama-García K., Sepiurka L., Garcia R.D.C.M. Abandonment of dogs in Latin America: Strategies and ideas // Vet World. — 2021. — Vol. 14, № 9. — P. 2371-2379.

12. Zumpano R., Tortosa A., Degregorio O.J. Estimación del impacto de la esterilización en el índice de crecimiento de la población de caninos // Revista de Investigaciones Veterinarias del Perú. — 2011. — Vol. 22, № 4. — P. 336-341.

13. Eisenberg N., Sadovsky A. Prosocial Behavior, Development of // Encyclopedia of Applied Psychology, Three-Volume Set. 1st ed. / ed. Spielberger C. Elsevier Science Ltd., — 2004. — P. 137-141.

14. Tusche A., Bockler A., Kanske P., Trautwein F. M., Singer T. Decoding the Charitable Brain: Empathy, Perspective Taking, and Attention Shifts Differentially Predict Altruistic Giving // Journal of Neuroscience. — 2016. — Vol. 36, № 17. — P. 4719-4732.

15. Homer P.M. When sadness and hope work to motivate charitable giving // Journal of Business Research. — 2021. — Vol. 133. — P. 420-431.

16. Liang J., Chen Z., Lei J. Inspire me to donate: The use of strength emotion in donation appeals // Journal of Consumer Psychology. — 2016. — Vol. 26, № 2. — P. 283-288.

17. Small D.A., Verrochi N.M. The Face of Need: Facial Emotion Expression on Charity Advertisements // Journal of Marketing Research, — 2009. — Vol. 46, № 6. — P. 777-787.

18. Hine D.W., Gifford R. Fear appeals, individual differences, and environmental concern // The Journal of Environmental Education. — 1991. — Vol. 23, № 1. — P. 36-41.

19. Telzer E.H., Fuligni A.J. Daily Family Assistance and the Psychological Well-Being of Adolescents From Latin American, Asian, and European Backgrounds // Developmental psychology. — 2009. — Vol. 45, № 4. — P. 1177-1189.

20. van Doorn J., Zeelenberg M., Breugelmans S.M. The impact of anger on donations to victims // International Review of Victimology. — 2017. — Vol. 23, № 3. — P. 303-312.

21. Burt C.D.B., Strongman K. Use of images in charity advertising: Improving donations and compliance rates // International Journal of Organisational Behaviour. — 2005. — Vol. 8, № 8. — P. 571-580.

22. Sabato H., Kogut T. Happy to help—if it's not too sad: The effect of mood on helping identifiable and unidentifiable victims // PLoS One. — 2021. — Vol. 16, № 6. — P. e0252278.

23. Fiala L., Noussair C.N. Charitable Giving, Emotions, and the Default Effect // Economic Inquiry. — 2017. — Vol. 55, № 4. — P. 1792-1812.

24. Genevsky A., Västfjäll D., Slovic P., Knutson B. Neural Underpinnings of the Identifiable Victim Effect: Affect Shifts Preferences for Giving // Journal of Neuroscience. — 2013. — Vol. 33, № 43. — P. 17188-17196.

25. Septianto F., Tjiptono F. The interactive effect of emotional appeals and past performance of a charity on the effectiveness of charitable advertising // Journal of Retailing and Consumer Services. — 2019. — Vol. 50. — P. 189-198.

26. Maison D., Pawlowska B. Using the Facereader Method to Detect Emotional Reaction to Controversial Advertising Referring to Sexuality and Homosexuality // Springer Proceedings in Business and Economics. — 2017. — P. 309-327.

27. Dupre D., Krumhuber E.G., Küster D., McKeown G.J. A performance comparison of eight commercially available automatic classifiers for facial affect recognition // PLoS One. — 2020. — Vol. 15, № 4. — P. e0231968.

28. Stockli S., Schulte-Mecklenbeck M., Borer S., Samson A.C. Facial expression analysis with AFFDEX and FACET: A validation study // Behavior research methods. — 2018. — Vol. 50. — P. 1446-1460.

29. Mauss I.B., Robinson M.D. Measures of emotion: A review // Cognition and emotion. — 2009. — Vol. 23, № 2. — P. 209-237.

30. Boucsein W. Electrodermal activity. 2nd ed. Springer US, — 2012. 1-618 p.

31. Schretzmayer L., Kotrschal K., Beetz A. Minor immediate effects of a dog on children's reading performance and physiology // Frontiers in veterinary science.

— 2017. — Vol. 4. — P. 257186.

32. Paramita W., Septianto F., Tjiptono F. The distinct effects of gratitude and pride on donation choice and amount // Journal of Retailing and Consumer Services. — 2020. — Vol. 53. — P. 101972.

33. Vitaglione G.D., Barnett M.A. Assessing a New Dimension of Empathy: Empathic Anger as a Predictor of Helping and Punishing Desires // Motivation and Emotion.

— 2003. — Vol. 27, № 4. — P. 301-325.

34. Urbonavicius S., Adomaviciute K., Urbutyte I., Cherian J. Donation to charity and purchase of cause-related products: The influence of existential guilt and experience // Journal of Consumer Behaviour. — 2019. — Vol. 18, № 2. — P. 89-96.

35. Haynes M., Thornton J., Jones S.C. An exploratory study on the effect of positive (warmth appeal) and negative (guilt appeal) print imagery on donation behaviour in animal welfare // Proceedings of the Marketing Accountabilities and Responsibilities: ANZMAC 2004 Conference. Wellington, New Zealand: Australian and New Zealand Marketing Academy, 2004.

36. Zarghamee H.S., Messer K.D., Fooks J.R., Schulze W.D., Wu S., Yan J. Nudging charitable giving: Three field experiments // Journal of behavioral and experimental economics. — 2017. — Vol. 66. — P. 137-149.

37. Ge R., Zhang S., Zhao H. Do expressions of sadness, anxiety and fear have different impacts on attracting donations? Evidence from a Chinese online charitable crowdfunding platform // Information Technology and People. — 2022. — Vol. 36. №. 7. — P. 2869-2893.

38. Palosaari E., Heme K., Lappalainen O., Hietanen J.K. Effects of fear on donations to climate change mitigation // Journal of Experimental Social Psychology. — 2023. — Vol. 104. — P. 104422.

39. Amato P.R. Emotional Arousal and Helping Behavior in a Real-Life Emergency // Journal of Applied Social Psychology. — 1986. — Vol. 16, № 7. — P. 633-641.

40. Basil D.Z., Ridgway N.M., Basil M.D. Guilt and giving: A process model of empathy and efficacy // Psychology & marketing. — 2008. — Vol. 25, № 1. — P. 1-23.

41. Cockrill A., Parsonage I. Shocking People Into Action: Does It Still Work? // J Advert Res. Journal of Advertising Research, — 2016. — Vol. 56, № 4. — P. 401413.

42. O'Loughlin Banks J., Raciti M.M. Perceived fear, empathy and financial donations to charitable services // Service Industries Journal. — 2018. — Vol. 38, № 5-6. — P. 343-359.

43. Zubair M., Wang X., Iqbal S., Awais M., Wang R. Attentional and emotional brain response to message framing in context of green marketing // Heliyon. — 2020. — Vol. 6, № 9. — P. e04912.

44. Ash S. The Barnardo's Babies: Performativity, Shame and the Photograph // Continuum. — 2005. — Vol. 19, № 4. — P. 507-521.

45. Beswick D., Dasandi N., Hudson D. International development NGOs, representations in fundraising appeals and public attitudes in UK-Africa relations // Britain and Africa in the twenty-first century. Between ambition and pragmatism / ed. Beswick D., Fisher J., Hurt S.R. Manchester University Press, — 2019. — P. 196-213.

46. Stayman D.M., Aaker D.A. Are All the Effects of Ad-Induced Feelings Mediated by AAd? // Journal of Consumer Research. — 1988. — Vol. 15, № 3. — P. 368373.

47. Cavanaugh L.A., Bettman J.R., Luce M.F. Feeling Love and Doing More for Distant Others: Specific Positive Emotions Differentially Affect Prosocial

Consumption // Journal of Marketing Research. — 2015. — Vol. 52, № 5. — P. 657-673.

48. Sciulli L.M., Bebko C.P., Bhagat P. How Emotional Arousal and Attitudes Influence Ad Response: Using Eye Tracking to Gauge Nonprofit Print Advertisement Effectiveness // Journal of Marketing Management. — 2017. — Vol. 5, № 1. — P. 1-11.

49. Lee D.I., Park J.H. The relationship between a charity crowdfunding project's contents and donors' participation: An empirical study with deep learning methodologies // Computers in Human Behavior. — 2020. — Vol. 106. — P. 106261.

50. Lehmann V., Huis in't Veld E.M.J., Vingerhoets A.J.J.M. The human and animal baby schema effect: Correlates of individual differences // Behavioural processes.

— 2013. — Vol. 94. — P. 99-108.

51. Mercadillo R.E., Barrios F.A., Díaz J.L. Definition of Compassion-Evoking Images in a Mexican Sample // Perceptual and motor skills. — 2007. — Vol. 105, № 2. — P. 661-676.

52. Wiesenthal D.L., Austrom D., Silverman I. Diffusion of Responsibility in Charitable Donations* // Basic and Applied Social Psychology. — 1983. — Vol. 4, № 1. — P. 17-27.

53. Garcia S.M., Weaver K., Moskowitz G.B., Darley J.M. Crowded minds: The implicit bystander effect // Journal of personality and social psychology. — 2002.

— Vol. 83, № 4. — P. 843-853.

54. Bradley M.M., Codispoti M., Cuthbert B.N., Lang P.J. Emotion and Motivation I: Defensive and Appetitive Reactions in Picture Processing // Emotion. — 2001. — Vol. 1, № 3. — P. 276-298.

55. Dimberg U., Karlsson B. Facial reactions to different emotionally relevant stimuli // Scandinavian Journal of Psychology. — 1997. — Vol. 38, № 4. — P. 297-303.

56. Fridlund A.J., Cacioppo J.T. Guidelines for Human Electromyographic Research // Psychophysiology. — 1986. — Vol. 23, № 5. — P. 567-589.

57. Shepelenko A., Kosonogov V., Shestakova A. How Emotions Induce Charitable Giving. A Psychophysiological Study // Social Psychology. — 2023. — Vol. 54, №2 5. — P. 261-270.

58. Zielinski P. An Arousal Effect of Colors Saturation // Journal of Psychophysiology.

— 2016. — Vol. 30, № 1. — P. 9-16.

59. Sánchez-Navarro J.P., Martínez-Selva J.M., Torrente G., Román F. Psychophysiological, Behavioral, and Cognitive Indices of the Emotional Response: A Factor-Analytic Study // The Spanish journal of psychology. — 2008.

— Vol. 11, № 1. — P. 16-25.

60. Messerotti Benvenuti S., Buodo G., Dal Bo E., Palomba D. Attention and affect in dysphoria: Insights from startle reflex modulation and cardiac deceleration // Behaviour Research and Therapy. — 2020. — Vol. 131. — P. 103626.

61. Shepelenko A., Shepelenko P., Obukhova A., Kosonogov V., Shestakova A. The relationship between charitable giving and emotional facial expressions: Results from affective computing // Heliyon. — 2024. — Vol. 10. — P. 23728.

62. Benjamini Y., Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing // Journal of the Royal Statistical Society: Series B (Methodological). — 1995. — Vol. 57, № 1. — P. 289-300.

63. Faul F., Erdfelder E., Lang A.G., Buchner A. G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences // Behavior research methods. — 2007. — Vol. 39, № 2. — P. 175-191.

64. Shepelenko A., Shepelenko P., Panidi K., Kosonogov V., Shestakova A. How the emotions evoked by homeless pets induce online charitable giving // Journal of Philanthropy and Marketing. — 2024. — Vol. 29, № 2. — P. e1842.

Приложение 1

Таблица П1

Характеристики стимулов

Категория стимула Бездомный (да = 1, нет = 0) Больной (да = 1, нет = 0) Щенок (да = 1, нет = 0) Присутствие человека(да = 1, нет = 0)

1 0 0 0 1

2 0 1 1 1

3 0 0 1 1

4 1 1 1 0

5 1 0 0 1

6 0 0 1 0

7 1 1 1 1

8 1 0 0 0

9 1 0 1 1

10 1 1 0 1

11 1 1 0 0

12 0 0 0 0

13 0 1 0 0

14 0 1 1 0

15 1 0 1 0

16 0 1 0 1

Примечание: каждая категория имеет разное сочетание 4 характеристик: 1) бездомный / не бездомный; 2) больной / не больной (здоровый); 3) щенок / не щенок (взрослый); 4) с присутствием человека на картинке и без него.

Приложение 2. Статья «How Emotions Induce Charitable Giving: A

Psychophysiological Study»

Авторы: Шепеленко А.Ю., Косоногов В.В., Шестакова А.Н. Аннотация: Emotions play a pervasive role in determining advertising effectiveness. However, it is still a controversial question as to whether pleasant or unpleasant advertisements are more effective in attracting donations. We recorded self-reported valence and arousal, as well as physiological activity (skin conductance, heart rate, and facial electromyography), while 54 participants watched affective pictures of dogs and made decisions with regard to possible donations. More unpleasant pictures provoked larger donations. Corrugator EMG was the most closely related to donations. A change in heart rate and zygomaticus EMG also correlated with charitable giving, but to a lesser extent. Hence, corrugator EMG could be useful in studies of emotional influence on prosocial behavior. With regard to stimulus characteristics, homeless and sick dogs provoked a greater affective response and larger donations.

Original Article

How Emotions Induce Charitable Giving

A Psychophysiological Study

Anna Shepelenko, Vladimir Kosonogov, and Anna Shestakova

Institute for Cognitive Neuroscience, HSE University, Moscow, Russia

a |

I

1 f. I

Abstract: Emotions play a pervasive role in determining advertising effectiveness. However, it is still a controversial question as to whether pleasant or unpleasant advertisements are more effective in attracting donations. We recorded self-reported valence and arousal, as welt as physiological activity (skin conductance, heart rate, and facial electromyography), while 54 participants watched affective pictures of dogs and made decisions with regard to possible donations. More unpleasant pictures provoked larger donations. Corrugator EMG was the most closely related to donations. A change in heart rate and zygomaticus EMG also correlated with charitable giving, but to a lesser extent. Hence, corrugator EMG could be useful in studies of emotional influence on prosocial behavior. With regard to stimulus characteristics, homeless and sick dogs provoked a greater affective response and larger donations.

Keywords: charity, advertisement, emotions, valence, arousal, psychophysiology

Prosocial behavior and charity play a significant role in socioeconomic processes, allowing societies to resolve crises collectively and to significantly improve the lives of socially unprotected segments of the population. However, for many charities, fundraising is extremely difficult (Buteau et al., 2013; Gratton, 2018), especially in the context of the pandemic and the global economic crisis. Specifically, in UK, charities have requested government assistance (Pro Bono Economics, 2020), and in the United States, 70% of charities reported a decrease in donations (CAF America, 2020). A better understanding of the behavioral and particularly emotional reactions associated with charitable giving can increase the effectiveness of communications aimed at attracting individual donations. In the United States, 70% of all contributions were donated by individuals (Giving USA Foundation, 2019).

Emotions play a pervasive role in advertising effectiveness (Poels & Dewitte, 2019), which may be due to the fact that emotional content (including images, words, etc.) attracts more attention and causes a deeper cognitive processing than emotionally neutral information (Huang et al., 2021; Schindler & Kissler, 2016). Various discrete emotions, such as pride, disgust, fear, guilt and shame, can affect the effectiveness of advertising messages in different ways (Brennan & Binney, 2010; Northey et al., 2020; Poels & Dewitte, 2019; Septianto & Tjiptono, 2019). Emotional appeals are widely used to attract attention for both commercial and social marketing campaigns. In particular, 95%

of prescription drug advertisements on American television contain emotional appeals (Frosch et al., 2007). However, 90% of advertisements fail to reach their target because they are never seen by people (Trott, 2013). In the battle for consumer attention, marketers use shock and incongruent advertising (Waller, 1999). This approach is justified by the fact that negative emotioas can contribute to attracting consumer awareness and action (Banyte et al., 2014).

Valence is one of the affective factors that influence perception. As in commercial advertisement, emotional appeals are commonly used in social campaigns. Negative and shocking messages typically dominate (Albouy, 2017; Chang & Lee, 2009). Moreover, negative emotions prevail in appeals to induce donations to charities (Bagozzi & Moore, 1994). Previous research has shown that shock advertisements are effective for increasing attention and memorability and positively influence the awareness of social campaigns (Dahl et al., 2003). In addition, shock content was perceived as more justifiable for nonprofit organizations than for commercial ones (Parry et al., 2013). This effect of shock advertising is achieved by violating social and personal norms (Dahl et al., 2003). At the same time, the use of religious taboos or morally offensive images has been shown to be perceived as unacceptable for both commercial and social advertising and provoke the opposite reaction (Parry et al., 2013).

However, the frequent use of negative emotions has been criticized (Septianto & Tjiptono, 2019). It has been

I

1 f. I

I

«B 8.

shown that repeated exposure of negative images causes wear out (extinction) in terms of emotions felt and attention paid (Thornton & Rossiter, 2001). Moreover, repetitive use of negative appeals can lead to a reduction in empathy and may even cause aversion to a charitable advertisement (Stayman & Aaker, 1988). In some cases, negative emotions were shown to cause distraction from the information and trigger a defensive reaction in the target audience. For example, d'Adda et al. (2018) found that moral commitment leads to information avoidance, as evidenced by the positive correlation between environmental information avoidance and the use of air conditioners. In a study by Chen (2016), the participants who were asked to read the low-fear appeal text had a stronger intention to participate in proenvironmental activities in comparison to those who read the high-fear appeal text.

Positive emotions are less frequently used in social campaigns. For instance, combining the positive emotion of strength and the negative emotion of sadness has helped to effectively persuade people to donate (Liang et al., 2016). However, in some cases, the positive emotions might negatively affect the willingness to help. Using images of the happy faces of the beneficiaries has been shown to reduce empathy and donation size (Small & Verrochi, 2009). A comparison of positive and negative affective stimuli showed that negative stimuli are usually better remembered (Lang et al., 1995). Haynes et al. (2004) showed that negative appeal was more successful at increasing the intention to donate money and time. It has also been shown that positive and emotionally neutral priming, not related to the beneficiary of the donation, provokes a smaller amount of donations than negative (Huang et al., 2021).

At the same time, in recent years, calls for using positive appeals in charity advertising have increased. In a study by Genevsky et al. (2013), conditions that elicited more positive arousal also elicited higher donations, whereas conditions that elicited higher negative arousal did not. In addition, the interest in using positive emotions in donations is growing rapidly, with an ever increasing exploration of how positive emotional appeals can be effective for motivating prosocial behavior (Cavanaugh et al., 2015; Piff et al., 2015).

Arousal can also influence prosocial behavior. It has been shown that arousing (i.e., provoking any emotion) stimuli, compared to neutral ones, are more attention-grabbing (Citron et al., 2014). In addition, arousal affects the information memorization processes; in particular, objects with a negative or positive valence are better remembered than neutral objects (Kensinger & Corkin, 2004). Amato (1986) interviewed people shortly after a major fire and showed that those who experienced a higher level of emotional arousal (such as terror, horror, or

shock) made donations to relief funds more often and showed a desire to help victims of the fire. Nevertheless, it should be noted that in that research, the focus was mainly on negative emotions. Arousal also affects proecological behavior even when its source is not ecologically related (Pu et al., 2018). It was shown that after watching a 20 s video of an exciting roller coaster, which causes positive arousal, participants were more likely to be concerned about the environment and were more likely to sign an environmental petition than when watching a nonexciting 20 s video with a moving train. However, viewing the exciting video did not have any effect on the willingness to donate. A study of visual attention using eye tracking showed that the stronger the emotions evoked by the ad and the more favorable the attitudes toward the ad, the more engaged a viewer may become, and, as a result, the greater the propensity for that individual to donate to the advertised charity (Sciulli et al., 2017). Thus, it can be noted that emotional arousal affects the perception of information, attention, and interest in viewing ads.

It is also important to understand why emotions lead to donations. The donor's emotional state can be influenced by the characteristics of the beneficiaries (e.g. age; Lee & Park, 2020), the charity's context (e.g. singularity, when a single person in need elicits more compassion than a group; Vastfjall et al., 2014), or the image used, such as the background (Choi et al., 2020) or image quality (Lee & Park, 2020). It has been shown that more donations are made for children than for adults (Lee & Park, 2020). In particular, puppies may trigger a greater willingness to help because they can evoke parental patterns by analogy with infants (Lehmann et al., 2013). Pictures showing suffering in social contexts and expressions of sadness elicited the highest level of compassion response (Mercadillo et al., 2007). Thus, pictures of lonely, homeless, young, and sick dogs can be expected to evoke feelings of compassion and sadness, which in turn might increase donation rates compared to images of domestic and healthy dogs. Therefore, our study opted to focus on pet charities, where the most significant characteristics may be age, home presence, presence of a human nearby, and the health status of an animal. It was expected that there would be more donations for puppies than for adult dogs. We also supposed that the presence of a human in pictures could provoke the bystander effect or responsibility diffusion (Darley & Latane, 1968), which was also demonstrated in donations to charities (Garcia et al., 2002; Wiesenthal et al., 1983).

It is also important to note that most of the research on the influence of emotions on donations is based on declarations - advertising performance measured with surveys - not real donations. However, emotional reactions, especially negative ones, can occur subconsciously.

I

1 f. i

l

«B

8.

Some common methods to physiologically measure valence and arousal are skin conductance response (SCR), heart rate (HR), and electromyography (EMG). Corrugator supercilii EMG is a valid marker of negative emotions, whereas zygomaticus major reflects positive emotions (Bradley et al., 2001; Dimberg & Karlsson, 1997). Skin conductance correlates with arousal level, so it is sensitive to affective stimuli of any valence (Bradley et al., 2001; Zielinski, 2016). Both negative and positive stimuli were shown to decelerate HR more than neutral ones (Bradley et al., 2001; Sanchez-Navarro et al., 2008).

The goal of this study was to examine the relationship between the emotional state of contributors and their donations to pet charities. For this purpose, we complemented the results obtained using self-reports with the psychophysiological measurements. We registered self-reported valence and arousal, as well as physiological activity (SCR, HR, and facial EMG), while participants viewed pictures of dogs and made decisions about giving donations to them. We hypothesized that negative emotions would provoke larger donations. We also wanted to identify the crucial characteristics of pictures that would have the greatest effect on donations.

Methods

Participants

We ran a power analysis in which we opted for a medium effect size of .40 and a power of 80%, which gave us a sample size of 52. Therefore, 54 healthy participants (63% females, Mage = 24.5 years, SD = 5.7) participated in the study. All procedures were conducted in accordance with the Declaration of Helsinki and approved by the local ethics committee. Prior to the experiment, the participants were given 320 rubles (the equivalent of eight metro tickets) to have the ability to donate throughout the experiment.

Materials and Design

Ninety-six colored photos of dogs were selected from freely accessible sources on the Internet. Of the 96 photos, 32 were selected for the experiment based on responses from a survey, two photos for each of the 16 categories. These 16 categories were the result of shuffling four binary characteristics of dogs that were considered potentially significant for donations: presence of a home, health status, age, and presence of a human nearby. The relevance of each stimulus to a specific category was tested in

a remote survey in which 15 respondents answered four questions: "Is the dog in the picture homeless?", "Is the dog sick?", "Is the dog a puppy?", and "Is there a human in the picture?" Photos with the highest level of consistent responses were selected for the experiment.

Procedure

First, participants viewed 32 pictures of dogs. The pictures were presented on a 19-in computer screen for 6 s. The intertrial interval was between 12 and 18 s to avoid anticipation in physiological variables. We used five different presentation orders to reduce the sequential effect of affective valence (Kosonogov, 2020). Pictures were qua-sirandomly distributed so that pictures from the same categories were not presented more than twice in a row. During this process, we recorded the participants' emotional states via corrugator and zygomaticus EMG, HR deceleration, and SCR After each picture, they were asked to donate 0-10 rubles to the fund that had sent the photograph. In the second part, they rated the valence and arousal of each picture (from 1 to 9) without physiological sensors in the free time setting. After the study, the collected money was actually sent to a pet charity.

Data Collection and Reduction

The PsychoPy software (Peirce et al., 2019) was used to run the experiment. The physiological signals were acquired, amplified, and filtered with an ActiChamp data acquisition system (Brain Products, Germany). The signals were recorded at a 1,000 Hz frequency.

Bipolar placement of Ag/AgCl surface electrodes on the second and the fourth fingers of the participant's left hand (nondominant for all of them) was used to obtain SCR values. The raw signal was set to detect activity in the 0-100 nS range. The SCRs were computed as the magnitude of peaks appearing between 0.9 s and 4 s after the onset of picture presentation (Venables & Christie, 1980). The data were transformed (loglO[SCR + 1]) to correct the distorted data distribution (Boucsein, 2012).

The corrugator supercilii and zygomaticus major EMG activity was collected using bipolar placement of 4-mm Ag/AgCl surface electrodes (Fridlund & Cacioppo, 1986) on the right side of the face. The raw EMG signal was intensified, and frequencies outside of the range of 10-350 Hz were filtered out. The signal was also full-wave rectified and integrated offline with a time constant of 500 ms. For each picture onset, the EMG activity was acquired by subtracting activity in the 1 s before the picture onset from the average activity occurring during the 6 s of picture

Table 1. The descriptive statistics of average self-reported responses to pictures

I

1 f.

■8

«a

8.

Measure M SD

Valence, 1-9 4.45 1.66

Arousal, 1-9 5.19 0.54

Donation, 0-10 6.41 1.03

presentation, which made it possible to neutralize the effect of the emotional state of the subject before the presentation of the stimulus.

A photoplethysmograph was used to record the amplitude of blood volume in the finger using a photocell from the distal phalange of the third finger of the left hand (nondominant for all participants). Heart rate (HR, in beats per minute, bpm) was calculated from amplitude peaks, and its deceleration equaled the difference between the HR at each second of the image presentation and the baseline (1 s before the stimulus onset). Using methods mentioned in a recent study by Benvenuti et al. (2020), we calculated the maximum deceleration from the baseline in the first 3 s (1-3 s) and in the last 3 s (4-6 s) of the picture presentation.

Data Analysis

The normality of the distributions was measured as the Kolmogorov-Smirnov d statistic. Internal consistency of the donation task was calculated as Cronbach's a. Pearson's correlation analysis and quadratic regression were performed to reveal relationships between all dependent variables (with pictures as cases). The multiple correlations were corrected using the false discovery rate (Benjamini & Hochberg, 1995). To study the possible effects of dog characteristics, we conducted f-tests for dependent samples independently for each binary characteristic and each dependent variable. The multiple comparisons were also corrected using the false discovery rate. As a measure of effect size, Cohen's d was calculated for each comparisoa

Results

Manipulation Check

The descriptive statistics of self-reported responses to the pictures are presented in Table 1. The distributions of valence, arousal, and donations were normal (ds equaled 0.14, 0.07, and 0.10, respectively, all ps > .20). Internal consistency of the donation task was excellent (a = .98),

8 7

w r

3 6

o

a 5 4

J 3

v. M. V

• •

12 3 4 5 6 unpleasant va)ence

7 8 9 pleasant

Figure 1. The relationship between valence and arousal of the presented pictures of dogs.

which allows us to believe that all items measured the same construct.

The relationship between valence and arousal showed a typical quadratic pattern, F(2,29) = 31.38, p < .001, R2 = .68 (Figure 1). Valence negatively correlated with corrugator EMG activity, r(32) = -.92, p < .001. Zygomaticus EMG positively correlated with valence, r(32) = .40, p = .024. Valence also positively correlated with HR deceleration (4-6 s), r(32) = .39, p = .027. That is, the more unpleasant pictures evoked larger corrugator and smaller zygomaticus major activities, as well as larger HR deceleration. Arousal positively correlated with SCR, /-(32) = .56, p = .001. In other words, more arousing pictures evoked larger SCRs. Arousal also negatively correlated with HR deceleration (4-6 s), r{32) = -.38, p = .030. More arousing pictures evoked larger HR deceleration.

Correlations Between Donations and Other Measures

Pearson's correlation analysis showed that the size of a donation negatively correlated with valence measured by self-reporting, r(32) = -.81, p < .001 (Figures 2, 3). We also found both a linear, r(32) = .48, p = .006, and a quadratic relationship between arousal and donations, F(2,29) = 9.08, p = .001 (Figures 2, 4). A positive correlation was found between corrugator EMG and donations, r(32) = .84, p < .001 (Figures 2, 5). Zygomaticus EMG also correlated with donations, r(32) = -.39, p = .029 (Figures 2, 6). Therefore, pictures attract larger donations if they evoke larger corrugator and smaller zygomaticus major activity. A change in heart rate (4-6 s) negatively correlated with donations, r(32) = - .36, p = .042 (Figures 2,7). Pictures that evoked larger deceleration attracted larger donations. The SCRs did not correlate with donations, r(32) = .17, p = .37.

a

I

■8 s.

3

c 2

0

S 1

V

1 0 o

•o

at -1 33

-2 -3

© o

donations

Figure 7. The relationship between heart rate deceleration and donations.

Table 2. The descriptive statistics of self-report responses to pictures with different dog characteristics

Donations, 0-10 Valence, 1-9 Arousal, 1-9

Characteristics M SD M SD M SD

Homeless 6.60" 2.65 3.46e 0.74 5.19 1.45

Domestic 6.23* 2.50 5.43e 0.95 5.19 1.35

Sick 7.28" 2.68 3.31" 1.01 5.51' 1.49

Healthy 5.55" 2.66 5.58" 0.78 4.83' 1.40

Puppy 6.35 2.51 4.72« 0.80 5.118 1.32

Adult 6.47 2.58 4.17« 0.75 5.35« 1.32

With a human 6.43 2.53 4.43 0.81 5.17 1.36

Without a human 6.40 2.51 4.46 0.72 5.20 1.35

Note. Means (M) with the same superscripts differ at p< .05. SD= standard deviation.

Discussion

The present study investigates the relationship between the human emotional state and the size of donations to homeless pets. We are the first to measure emotional state not only with self-reported scales (valence and arousal) but also using objective physiological correlates (SCR, HRand facial EMG), making this a unique contribution to the literature. Participants viewed pictures of dogs and made decisions about possible donations to them. We hypothesized that emotional valence would have an effect on the donation size. We also attempted to identify those characteristics of dogs and scenes that would influence donations.

First, we found a clear negative correlation between valence and donations. In other words, our participants donated more when they experienced more negative emotions. This relationship seems to be linear but not quadratic ([/-shaped). Very unpleasant pictures (2 in the 1-9 scale) still provoked a great amount of donations. This is in accordance with the studies that also showed the effectiveness of unpleasant advertising material used by

charities for generating awareness. For example, similar results were found in an intergroup study where subjects viewing a painting of a sad dog were potentially willing to donate more than subjects viewing a happy dog (Haynes et al., 2004) and in a study with children where sad children were given more donations than smiling children (Small & Verrochi, 2009).

In contrast to Genevsky et al. (2013), who showed that pleasant pictures could provoke larger donations, we demonstrated the opposite relationship. This result could be explained by the fact that the positive emotions of the subjects negatively correlated with the required amount of donation. In the Genevsky et al. (2013) study, the participants did not choose the donation amount themselves; they only agreed or refused to donate the amount that was automatically offered. Their positive emotions could be elicited by a mere decrease in the required donation amount but also by selecting the recipient of the donation.

We did not find a quadratic ([/-shaped) relationship between valence and prosocial behavior, as in the study by Borawska et al. (2020), who concluded that the most negative emotions (provoked with shocking scenes), in comparison with mildly negative emotions, did not increase the intention to drive in a safer manner. Nevertheless, we need to note that they drew this conclusion after a moderation analysis with three levels of independent variables but not a correlation analysis regarding the whole spectrum of emotion values. Our finding of the association between prosocial behavior and emotional valence calls for further investigation.

As for the relationship between donations and arousal, both linear and quadratic relationships were significant. If we accept the quadratic one, it will mean that some highly arousing pictures elicit large donations, but at the same time, other highly arousing pictures generate small donations. Medium-size donations were provoked by non-arousing (neutral) pictures. However, the left tail of this curve consists of five arousing (4.87-5.37) pictures attracting the least donations (4.63-5.17). These pictures turned out to be of pleasant valence (6.87-7.24), all depicting healthy domestic dogs. This leads us to the conclusion that it is not arousal itself that provokes larger donations, but it strengthens the influence of negative valence. This is in accordance with the data of Amato (1996), who showed that greater arousal of negative emotions prompted a higher intention to engage in prosocial behavior.

Of note, our unpleasant pictures were closer to the negative pole of valence than our pleasant pictures were close to the positive pole. In other words, our unpleasant pictures were more unpleasant than our pleasant pictures were pleasant. At the same time, our pleasant pictures

Table 3. The descriptive statistics of physiological responses to pictures with different dog characteristics

Characteristics Corrugator EMG, nV Zygomaticus EMG, HV SCR, |iS HR 4-6 s, bpm

M SD M SD M SD M SD

Homeless 0.18' 0.49 0.13 0.14 0.24 0.26 -0.30 1.76

Domestic -0.06" 0.31 0.16 0.37 0.22 0.28 0.04 1.89

Sick 0.23» 0.56 0.11 0.17 0.26" 0.31 -0.49« 1.75

Healthy -O.îl" 0.32 0.17 0.36 0.20" 0.23 0.24* 1.69

Puppy 0.02e 0.28 0.15 0.30 0.23 0.29 -0.03 1.75

Adult 0.10« 0.44 0.13 0.20 0.23 0.25 -0.23 1.68

With a human 0.06 0.35 0.16 0.23 0.21 0.27 0.04 1.82

Without a human 0.05 0.38 0.12 0.28 0.24 0.27 -0.03 1.68

Note. Means (M) with the same superscripts differ at p < .05. SD = standard deviation.

a f

I

1 f. i

I

¿a 8.

8 &

were less arousing than our unpleasant pictures. In International Affective Picture System (IAPS; Lang et aL, 1997), a database of affective pictures, pleasant pictures of dogs were evaluated at the same level: around 7 (on a scale from 1 to 9) of valence and 5 (out of 9) of arousal (the scales were the same as in our study). In the EmoMadrid database (Carretie et al., 2019), the pattern is similar. Hence, we are prone to believe that this could be considered a reflection of the negativity bias (Cacioppo & Berntson, 1994; Taylor, 1991). Our results corroborate the above-mentioned observations showing that emotional reactions were biased toward the negative pole. In our study, unpleasant pictures were rated as more arousing than pleasant pictures.

This fact could also influence our zygomaticus EMG data. We find a correlation between zygomaticus activity and self-reported valence, as well as donations; however, compared to corrugator EMG, the association between muscle activity and the size of the donation was weaker. Also, this metric was not sensitive to dog characteristics. It is well known that activity of the zygomaticus muscle is stronger in response to pleasant stimuli (Lang et al., 1993). Perhaps our pleasant pictures were not pleasant enough to strongly activate this muscle. In addition, comparing different pleasant content, Bradley et al. (2001) showed that aside from erotica, pictures of nature provoked the least zygomaticus activity. Hence, zygomaticus EMG would be recommended as a tool for studying pet charity advertisements, but with some limitations.

In contrast, corrugator EMG showed the largest correlation with donations in our study. This measure is typically interpreted as a marker of negative emotion (Bradley et al., 2001; Mohammed et al., 2021). The high level of correlation between the activity of this muscle and donations obtained in our study (.84) makes it very informative for predictions of the effectiveness of the advertisements built for attracting donations for pet

charities. Both corrugator EMG and valence ratings differed depending on three of the dog characteristics (homeless/domestic, sick/healthy, and puppy/adult). However, only two of them (homeless/domestic and sick/healthy) influenced donations. From this, it can be inferred that the differences between puppies and adult dogs were not strong enough to be reflected in donations.

Our manipulation check showed that SCRs correlated with self-reported arousal, as in previous studies (e.g. Zielinski, 2016). If we admit that arousal itself did not influence donations, then it is not surprising that SCRs did not correlate with donations. Therefore, SCRs would not provide much information for studying donations depending on valence.

Heart rate deceleration is also a measure of emotional state. Bradley et al. (2001) showed that negative pictures decelerated HR more than positive pictures did, whereas both affective categories decelerated HR more than neutral pictures. Sanchez-Navarro et al. (2008) found a correlation between HR change and arousal ratings. Our results are consistent with those of previous studies. In particular, our negative and more arousing pictures decelerate HR more than positive and less arousing ones. In our study, we find a negative correlation between HR deceleration and donations, meaning that pictures that decelerated HR more provoked larger donations. Thus, HR deceleration would be recommended as a tool for studying pet charity advertisements.

As for stimulus characteristics, images of homeless/ domestic and sick/healthy dogs yielded differences in donations. The sick/healthy dichotomy had the greatest effect sizes and was significant throughout all measures except zygomaticus EMG. Altogether, these behavioral and psychophysiological findings can have important practical implications for pet charity advertisers. They may make an informed choice based on our findings to accordingly predict the effectiveness of the campaign.

a

I

I

1 f. I

This study is the first to use a comprehensive approach to measuring emotions in the context of prosocial behavior, including data on autonomic and behavioral responses. The data obtained will make it possible to identify the most convenient and reliable method for assessing the effectiveness of advertising materials, which can be used for practical tasks of nonprofit organizations. Future studies could be performed to take into account other characteristics of images, such as image quality, the dogs' emotions, their sickness levels, and additional elements in the photos. Moreover, in future studies, the personal attitudes of subjects could be monitored using a survey about their propensity for donations and prosocial behaviors as well as their attitudes toward animals, especially having experience owning a pet (dog, cat, or other). An additional variable to be considered in future studies is the emotional state of the subjects before the start of the experiment, which, as previously shown, can increase or decrease the emotional state caused by charity appeals (Huang et al., 2021). An important limitation of this study which should be mentioned is the effects of laboratory conditions on the behaviors of subjects. In particular, under the conditions of the experiment, the subjects were in a confined space and were forced to view images, even unpleasant ones, which, as a result, could provoke a donation. However, in real life, unpleasant pictures could cause avoidance or distraction, and therefore, donation behavior would differ.

To conclude, first, our experiment showed the negative linear relationship between valence and donations to pet charities. That is, unpleasant pictures of dogs provoked larger donations. Second, higher arousal was associated with larger donations. Third, among physiological measures, corrugator EMG was the most closely related to the effect on donations. Pictures that provoked higher corrugator activity attracted larger donations. Hence, corrugator EMG could be used in future studies of emotional influence on prosocial behavior and be recommended for use in the applied research of advertising for charity campaigns.

References

Albouy, J. (2017). Emotions et comportements prosociaux: Étude de l'efficacité des campagnes humanitaires émotionnelles [Emotions and prosocial behaviours: A study of the effectiveness of shocking charity campaigns]. Recherche et Applications en Marketing (French Edition), 32(2), 5-26. https://doi.org/10. 1177/0767370116676242 Amato, P. R. (1986). Emotional arousal and helping behavior in a real-life emergency. Journal of Applied Social Psychology, 76(7), 633-641. https://doi. org/10.1111/j.1559-18l6.1986.tb01164.x Bagozzi, R. P., & Moore, D. J. (1994). Public service advertisements: Emotions and empathy guide prosocial behavior. Journal of Marketing, 58(1), 56-70. https://doi.org/10.1177/002224299405800105

Banyte, J., Paskeviciute, K., & Rutelione, A. (2014). Features of shocking advertising impact on consumers in commercial and social context. Innovative Marketing, 70(2), 35-46.

Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B (Methodological), 57(1), 289-300. https://doi.Org/10.1111/j.2517-6161.1995. tb02031.x

Benvenuti, S. M„ Buodo, G„ Dal Bo, E., & Palomba, D. (2020). Attention and affect in dysphoria: Insights from startle reflex modulation and cardiac deceleration. Behaviour Research and Therapy, 737, Article 103626. https://doi.Org/10.1016/j.brat.2020. 103626

Borawska, A, Oleksy,T., & Maison, D. (2020). Do negative emotions in social advertising really work? Confrontation of classic vs. EEG reaction toward advertising that promotes safe driving. PLoS One, 75(5), Article 0233036. https://doi.org/10.1371/journal.pone. 0233036

Boucsein, W. (2012). Electrodermal activity. Springer.

Bradley, M. M., Codispoti, M„ Cuthbert, B. N.. & Lang, P. J. (2001). Emotion and motivation I: Defensive and appetitive reactions in picture processing. Emotion, 7(3), 276-298. https://doi.org/10. 1037/1528-3542.1.3.276

Brennan, L„ & Binney, W. (2010). Fear, guilt, and shame appeals in social marketing. Journal of Business Research, 63(2), 140-146. https://doi.Org/10.1016/j.jbusres.2009.02.006

Buteau, E„ Brock, A, & Chaffin, M. (2013). Nonprofit challenges. What foundations can do. The Center for Effective Philanthropy, http:// www.effectivephilanthropy.org/portfolio-items/nonprof it-challenges

Cacioppo, J. T., & Berntson, G. G. (1994). Relationship between attitudes and evaluative space: A critical review, with emphasis on the separability of positive and negative substrates. Psychological Bulletin, 775(3), Article 401. https://doi.org/10.1037/ 0033-2909.115.3.401

CAF America (2020). The voice of charities facing COVID-19 worldwide, https://www.cafamerica.org/wp-content/uploads/ CV19_Report_CAF-America.pdf. Accessed 21 June 2021.

Carretie, L., Tapia, M„ Lopez-Martin, S„ & Albert, J. (2019). Emo-Madrid: An emotional pictures database for affect research. Motivation and Emotion, 43(6), 929-939. https://doi.org/10. 1007/sH031-019-09780-y

Cavanaugh, L. A, Bettman, J. R„ & Luce, M. F. (2015). Feeling love and doing more for distant others: Specific positive emotions differentially affect prosocial consumption. Journal of Marketing Research, 52(5), 657-673. https://doi.org/10.1509/jmr.10. 0219

Chang, C. T„ & Lee, Y. K. (2009). Framing charity advertising: Influences of message framing, image valence, and temporal framing on a charitable appeal 1. Journal of Applied Social Psychology, 3902), 2910-2935. https://doi.org/10.1111/j.1559-1816.2009.00555.x

Chen, M. F. (2016). Impact of fear appeals on pro-environmental behavior and crucial determinants. International Journal of Advertising, 35(1), 74-92. https://doi.org/10.1080/02650487. 2015.1101908

Choi, J., Li, Y. J., Rangan, P., Yin, B„ & Singh, S. N. (2020). Opposites attract: Impact of background color on effectiveness of emotional charity appeals. International Journal of Research in Marketing, 37(3), 644-660. https://doi.Org/10.1016/j.ijresmar. 2020.02001

Citron, F. M., Gray, M. A., Critchley, H. D„ Weekes, B. S., & Ferstl, E. C. (2014). Emotional valence and arousal affect reading in an interactive way: Neuroimaging evidence for an approach-withdrawal framework. Neuropsychologic, 56(100), 79-89. https://doi.Org/10.1016/j.neuropsychologia.2014.01.002

a

I

ji I I

s

0

1 &

d'Adda, G„ Gao, Y„ Golman, R„ & Tavoni, M. (2018). It's so hot in here: Information avoidance, moral wiggle room, and high air conditioning usage. FEEM Working Paper No. 07.2018. https:// doi.org/10.2139/ssrn.3149330 Dahl, D. W„ Frankenberger, K. D„ & Manchanda, R. V. (2003). Does it pay to shock? Reactions to shocking and nonshocking advertising content among university students. Journal of Advertising Research, 43(3), 268-280. https://doi.org/10.1017/ S0021849903030332 Darley, J. M„ & Latane, B. (1968). Bystander intervention in emergencies: Diffusion of responsibility. Journal of Personality and Social Psychology, 8(4p1), 377-383. https://doi.org/10.1037/ H0025589

Dimberg, U., & Karlsson, B. (1997). Facial reactions to different emotionally relevant stimuli. Scandinavian Journal of Psychology, 38(4), 297-303. https://doi.org/10.1111/1467-9450.00039 Fridlund, A. J., & Cacioppo, J. T. (1986). Guidelines for human electromyographic research. Psychophysiology, 23(5), 567-589. https://doi.Org/10.1111/j.1469-8986.1986.tb00676.x Frosch, D. L„ Krueger, P. M„ Hornik, R. C., Cronholm, P. F., & Barg, F, K. (2007). Creating demand for prescription drugs: A content analysis of television direct-to-consumer advertising. The Annals of Family Medicine, 5(1), 6-13. https://doi.org/10.1370/afm.611 Garcia, S. M„ Weaver, K., Moskowitz, G. B„ & Darley, J. M. (2002). Crowded minds: The implicit bystander effect. Journal of Personality and Social Psychology. 83(4), 843-853. https://doi.org/ 10.1037/0022-3514.83.4.843 Genevsky, A., Vastfjall, D„ Slovic, P., & Knutson, B. (2013). Neural underpinnings of the identifiable victim effect: Affect shifts preferences for giving Journal of Neuroscience, 33(43), 17188-17196. https://doi.org/10.1523/JNEUROSCI.2348-13.2013 Giving USA Foundation (2019). Giving USA The Annual Report on Philanthropy for the Year 2018. Giving USA. https://store. givingusa.org

Gratton, P. C. (2018). Organization development and strategic planningfor non-profit organizations. Organization Development Journal, 36(2), 27-38 Haynes, M„ Thornton, J„ & Jones, S. C. (2004). An exploratory study on the effect of positive (warmth appeal) and negative (guilt appeal) print imagery on donation behaviour in animal welfare. Proceedings of the Marketing Accountabilities and Responsibilities: ANZMAC 2004 Conference, School of Marketing and International Business, Victoria University of Wellington, New Zealand.

Huang, Q„ Li, D„ Zhou, C„ Xu, Q„ Li, P., & Warren, C. M. (2021). Multivariate pattern analysis of electroencephalography data reveals information predictive of charitable giving. Neuroimage, 242, Article 118475. https://doi.Org/10.1016/j.neuroimage.2021. 118475

Kensinger, E. A., & Corkin, S. (2004). Two routes to emotional memory: Distinct neural processes for valence and arousal. Proceedings of the National Academy of Sciences, 101(9), 3310-3315. https://doi.org/10.1073/pnas.0306408101 Kosonogov, V. (2020). The effects of the order of picture presentation on the subjective emotional evaluation of pictures. Psicologio, 34(2), 171-178. https://doi.org/10.17575/psicologia. v34i2.1608

Lang, A., Dhillon, K., & Dong, Q. (1995). The effects of emotional arousal and valence on television viewers' cognitive capacity and memory. Journal of Broadcasting & Electronic Media, 39(3), 313-327. https://doi.org/10.1080/08838159509364309 Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1997). International Affective Picture System (IAPS): Technical Manual and Affective ratings. In NIMH Center for the Study of Emotion and Attention (pp. 39-58). https://doi.org/10.1027/0269-8803/ a000147

Lang, P. J., Greenwald, M. K., Bradley, M. M„ & Hamm, A 0. (1993). Looking at pictures: Affective, facial, visceral, and behavioral reactions. Psychophysiology. 30(3), 261-273. https://doi.org/10. 1111/j.1469-8986.1993.tb03352.x Lee, D., & Park, J. (2020). The relationship between a charity crowdfunding project's contents and donors' participation: An empirical study with deep learning methodologies. Computers in Human Behavior, 106, Article 106261. https://doi.Org/10.1016/j. chb.2020.106261

Lehmann, V., Huis, E. M„ & Vingerhoets, A J. (2013). The human and animal baby schema effect: Correlates of individual differences. Behavioural Processes, 94, 99-108. https://doi.Org/10.1016/j. beproc.2013.01.001 Liang, J., Chen, Z, & Lei, J. (2016). Inspire me to donate: The use of strength emotion in donation appeals. Journal of Consumer Psychology, 26(2), 283-288. https://doi.Org/10.1016/j.jcps.2015. 09.001

Mercadillo, R. E„ Barrios, F. A., & Diaz, J. L. (2007). Definition of compassion-evoking images in a Mexican sample. Perceptual and Motor Skills, 705(2), 661-676. https://doi.org/10.2466/pms. 105.2.661-676

Mohammed, A R., Kosonogov, V., & Lyusin, D. (2021). Expressive suppression versus cognitive reappraisal: Effects on self-report and peripheral psychophysiology. International Journal of Psychophysiology, 167, 30-37. https://doi.org/10.1016/jjjpsycho. 2021.06.007

Northey, G„ Dolan, R„ Etheridge, J., Septianto, F„ & Van Esch, P. (2020). LGBTQ imagery in advertising: How viewers' political ideology shapes their emotional response to gender and sexuality in advertisements. Journal of Advertising Research, 60(2), 222-236. https://doi.org/10.2501/JAR-2020-009 Parry, S„ Jones, R., Stern, P., & Robinson, M. (2013). 'Shockver-tising': An exploratory investigation into attitudinal variations and emotional reactions to shock advertising. Journal of Consumer Behaviour, 12(2), 112-121. https://doi.org/10.1002/cb.1430 Peirce, J., Gray, J. R., Simpson, S., MacAskill, M., Hochenberger, R„ Sogo, H„ Kastman, E., & Lindetev, J. K. (2019). PsychoPy2: Experiments in behavior made easy. Behavior Research Methods. 51(1), 195-203. https://doi.org/10.3758/s13428-018-01193-y

Piff, P. K„ Dietze, P., Feinberg, M„ Stancato, D. M„ & Keltner, D. (2015). Awe, the small self, and prosocial behavior. Journal of Personality and Social Psychology, 108(6), 883-899. https://doi. org/10.1037/pspi0000018 Poels, K„ & Dewitte, S. (2019). The role of emotions in advertising: A call to action. Journal of Advertising, 48(1), 81-90. https://doi. org/10.1080/00913367.2019.1579688 Pro Bono Economics (2020). Charities facing £ 10.1 billion funding gap over the next six months. Pro Bono Economics. https://www. probonoeconomics.com/news/pres-release-charities-facing-101-biltion-funding-gap-over-the-next-six-months Pu, G. J., Jawanda, J., Sital, M., Kryschuk, T„ & Liang, V. (2018). Arousal and willingness to contribute to sustainable efforts. UBC Social Ecological Economic Development Studies (SEEDS) Sustainability Program, https://open.library.ubc.ca/soa/ clRcle/collections/undergraduateresearch/18861/items/1. 0374281

Sánchez-Navarro, J. P., Martínez-Selva, J. M„ Torrente, G„ & Román, F. (2008). Psychophysiological, behavioral, and cognitive indexes of the emotional response: A factor-analytic study. Spanish Journal of Psychology, 11(1), 16-25. https://doi.org/10. 1017/S1138741600004078 Schindler, S., & Kissler, J. (2016). Selective visual attention to emotional words: Early parallel frontal and visual activations followed by interactive effects in visual cortex. Human Brain Mapping, 37(10), 3575-3587. https://doi.org/10.1002/hbm.23261

a

I

ji Ï. I

I

Sciulli, L. M„ Bebko, C. P., & Bhagat, P. (2017). How emotional arousal and attitudes influence ad response: Using eye tracking to gauge nonprofit print advertisement effectiveness. Journal of Marketing Management, 5(1), 1-11. https://doi.org/10.15640/ jmm.v5n1a1

Septianto, F., & Tjiptono, F. (2019). The interactive effect of emotional appeals and past performance of a charity on the effectiveness of charitable advertising. Journal of Retailing and Consumer Services, 50, 189-198. https://doi.Org/10.1016/j. jretconser.2019.05.013 Shepelenko, A. (2023, June, 29). How emotions induce charitable

giving [Data], https://osf.io/vam38 Small, D. A., & Verrochi, N. M. (2009). The face of need: Facial emotion expression on charity advertisements. Journal of Marketing Research, 46(6), 777-787. https://doi.org/10.1509/ jmkr.46.6.777

Stayman, D. M„ & Aaker, D. A. (1988). Are all the effects of ad-induced feelings mediated by Aad? Journal of Consumer Research, 15(3), 368-373. https://doi.org/10.1086/209173 Taylor, S. E. (1991). Asymmetrical effects of positive and negative events: The mobilization-minimization hypothesis. Psychological Bulletin, 770(1), 67-85. https://doi.Org/10.1037/0033-2909.110.1.67 Thornton, J„ & Rossiter, J. R. (2001). Advertising wearout of shock-value anti-speeding ads. Proceedings of the Australian and New Zealand Marketing Academy Conference, Albany. Trott, D. (2013). Predatory thinking: A Masterclass in out-thinking

the competition. Pan Macmillan. Vástfjáll, D„ Slovic, P., Mayorga, M„ & Peters, E. (2014). Compassion fade: Affect and charity are greatest for a single child in need. PLoS One, 9(6), Article e100115. https://doi.org/10.1371/journal. pone.0100115

Venables, P. H„ & Christie, M. J. (1980). Electrodermal activity. In I. Martin& P. H. Venables (Eds.), Techniques in psychophysiology, (pp. 3-67). John Wiley. Waller, D. S. (1999). Attitudes towards offensive advertising: An Australian study. Journal of Consumer Marketing, 16(3), 288-295. https://doi.org/10.1108/07363769910271513

Wiesenthal, D. L, Austrom, D„ & Silverman, I. (1983). Diffusion of responsibility in charitable donations. Basic and Applied Social Psychology, 4(1), 17-27. https://doi.org/10.1207/s15324834basp0401_2 Zielihski, P. (2016). An arousal effect of colors saturation: A study of self-reported ratings and electrodermal responses. Journal of Psychophysiology, 30(1), 9-16. https://doi.org/10.1027/0269-8803/ a000149

History

Received November 17, 2022 Revision received January 3, 2023 Accepted January 3, 2023 Published online September 13, 2023

Acknowledgments

This study was implemented in the framework of the Basic Research Program at the HSE University in 2023 and used the HSE Synchronous Eye-tracking, Brain Signal Recording, and Non-Invasive Brain Stimulation System.

Publication Ethics

All procedures were conducted in accordance with the Declaration of Helsinki and approved by the local ethics committee.

Authorship

All authors were involved in all parts of the research. Open Data

The data are available at https://osf.io/vam38/ (Shepelenko, 2023).

Anna Shepelenko

Institute for Cognitive Neuroscience

HSE University

20 Myasnitskaya Ulitsa

Moscow 101000

Russia

[e-mail]

Приложение 3. Статья «The relationship between charitable giving and emotional facial expressions: Results from affective computing»

Авторы: Шепеленко А.Ю., Шепеленко П.А., Обухова А.А., Косоногов В.В., Шестакова А.Н.

Аннотация: This study investigated the relationship between emotional states (valence, arousal, and six basic emotions) and donation size in pet charities, and it compared the effectiveness of affective computing and emotion self-report methods in assessing attractiveness. Using FaceReader soft-ware and self-report, we measured the emotional states of participants (N = 45) during the donation task. The results showed that sadness, happiness, and anger were significantly related to donation size. Sadness and anger increased donations, whereas happiness decreased them. Arousal was not significantly correlated with the willingness to donate. These results are supported by both methods, whereas the self-reported data regarding the association of surprise, fear, and disgust with donation size are inconclusive. Thus, unpleasant emotions increase donation size, and combining affective computing with self-reported data improves the prediction of the effectiveness of a charity appeal. This study contributes to the understanding of the relationship between emotions and charitable behavior toward pet charities and evaluates the effectiveness of marketing mix elements using affective computing. The limitations include the laboratory setting for this experiment and the lack of measurement of prolonged and repeated exposure to unpleasant charity appeals.

Contents lists available at ScienceDirect

Heliyon

journal homepage: www.cell.com/heliyon

H)

The relationship between charitable giving and emotional facial expressions: Results from affective computing

Anna Shepelenkoa'*, Pavel ShepelenkoLl, Anastasia Obukhova Vladimir Kosonogov", Anna Shestakova

a Institute for Cognitive Neuroscience, HSE University, Moscow, Russia

Independent Researcher, Moscow, Russia c Phystech School of Biological and Medical Physics, Moscow Institute of Physics and Technology, Moscow, Russia

ARTICLE INFO

ABSTRACT

Keywords: Donation Pet charities Basic emotions FaceReader Facial expression Valence Arousal

This study investigated the relationship between emotional states (valence, arousal, and six basic emotions) and donation size in pet charities, and it compared the effectiveness of affective computing and emotion self-report methods in assessing attractiveness. Using FaceReader software and self-report, we measured the emotional states of participants (N = 45) during the donation task. The results showed that sadness, happiness, and anger were significantly related to donation size. Sadness and anger increased donations, whereas happiness decreased them. Arousal was not significantly correlated with the willingness to donate. These results are supported by both methods, whereas the self-reported data regarding the association of surprise, fear, and disgust with donation size are inconclusive. Thus, unpleasant emotions increase donation size, and combining affective computing with self-reported data improves the prediction of the effectiveness of a charity appeal. This study contributes to the understanding of the relationship between emotions and charitable behavior toward pet charities and evaluates the effectiveness of marketing mix elements using affective computing. The limitations include the laboratory setting for this experiment and the lack of measurement of prolonged and repeated exposure to unpleasant charity appeals.

1. Introduction

Charitable organizations significantly impact socioeconomic processes in society by implementing programs aimed at improving various aspects of the population's life or preventing its deterioration [1]. As instruments for promoting community subjective well-being [2], charities are highly dependent on community support. For instance, in the United States, approximately 70 % of all contributions to charitable organizations were made by individuals as opposed to corporate bodies [3].

With economic instability and global challenges faced by people worldwide, the smoothing effect exerted by charities is crucial. However, raising funds to implement mandates is challenging for many charities [4].

Pet charities find themselves in an especially difficult situation because of the ambiguous attitude of society toward their beneficiaries. Stray dogs can be perceived as both victims and enemies [5], which reduces donations to pet charities. This attitude is caused by the fact that, left unattended, stray dogs can become a source of rabies [6] or other diseases [7] and behave aggressively [8]; this

Corresponding author.

E-mail address: [e-mail] (A. Shepelenko).

https://doi.oig/10.1016/j.heliyon.2023.e23728

Received 7 July 2022; Received in revised form 4 December 2023; Accepted 12 December 2023 Available online 15 December 2023

2405-8440/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.Org/licenses/by-nc-nd/4.0/).

leads to their cruel treatment [5,9], which makes it vitally important to solve the pr oblem of homeless pets before anything negative happens to either animals or humans. Supporting charitable organizations with individual donations contributes to a humane solution to the problems caused by homeless pets as it allows the use of modem methods such as sterilization, vaccination, and the pr©motion of responsible ownership [ 10,11 ].

Exploring the reasons behind the donation decision-making process can provide additional support to the nonprofit sector in general and pet charities in particular". However, prosocial behavior, that is, the willingness to sacrifice one's resources for the benefit of another [12], is a complex psychological phenomenon [13] that is difficult to study and sometimes leads to contradictory findings.

Emotional state is one of the key factors influencing charitable behavior [14-16]. Among the basic emotions that affect the willingness to donate are sadness, anger, fear, and happiness [16-19]. However, the effectiveness of specific discr ete emotions in the context of appeals for charity still remains under debate. Contr adictions in the results of research r elate not only to the type of emotions but also to their valence. In par ticular", some studies have shown that negative (unpleasant) emotions ar e more effective in attr acting charity [20,21], whereas others have shown the opposite [22,23]. At the same time, research on the relationship between arousal, caused by a call for charitable giving, and willingness to donate is lacking.

Overall, research on emotions and prosocial behavior has shown a significant bias toward intra-species versus inter-species assistance. In other words, the bulk of published research has examined appeals for charities that support people [16,20-24], while the impact of emotions on prosociality toward animals remains under studied.

Regar ding methodological appr oaches, it should be noted that the bulk of studies of charitable behavior and emotions use declarative methods, which has certain limitations since participants may not always be conscious of their" feelings or not be willing to disclose them. This is especially tr ue for controversial and taboo topics, such as sexual orientation and venereal disease, helping socially undesir able or outcast gr oups [25], and it may also be applicable for stray dogs as the attitude towar d them is often ambivalent [5].

A solution to overcome methodological limitations in assessing emotions by declarative methods could be automatic emotion r ecognition software based on facial expressions, such as FaceReader, iMotions, and other s, which is referred to as affective computing [26-29]. It is also possible to comprehensively and discretely assess one's actual emotional state owing to their ability to simultaneously and objectively assess all basic emotions as well as valence and ar ousal. Thus, it can be assumed that affective computing can be an effective method for resear ching the relationship between emotions and prosocial behavior.

Affective computing is currently used in several areas, including diagnostics of mental disorders [30], improvement of online learning [31,32], vir tual reality [33], client assessment satisfaction [34,35], and the evaluation of the effectiveness of advertising [25, 36]. However , academic research assessing the pr edictive ability of affective computing methods for evaluating advertising mater ials compared to traditional methods, such as self-reported emotion, is lacking [37]. Despite the potential benefits that automatic emotion recognition systems can bring to mar keting [37,38], an insufficient number of published studies have explored affective computing to investigate which emotions are triggered by a product, advertisement, or another element of the marketing mix. In addition, within the framework of non-commercial advertising, we have found no studies on the r elationship between emotions and the effectiveness of appeals for charities using automatic emotion recognition systems.

Therefore, the main aim of this study is to examine the r elationship between emotional state and donation size in the context of pet charities. Moreover, it compares the effectiveness of using affective computing and emotional self-reporting to predict donations for stray dogs. To achieve this, we obtained six basic self-reported emotions in addition to valence and arousal, as well as emotional expressions automatically recognized by FaceReader, while participants viewed pictures of dogs' in various conditions and made decisions about donations in response to them. Based on the above, the following r esear ch questions were formulated:

Q1: How do valence and ar ousal relate to donation size in the context of pet charities?

Q2: How do the six basic emotions relate to donation size in the context of pet charities?

Q3: How do the predictive abilities of FaceReader (as one of the affective computing methods) compare with the self-reported emotions method to study the relationship between donations and emotional state?

Thus, this study contributes to the research on how the six basic emotions—along with valence and ar ousal, which are provoked by charity appeals—ar e related to willingness to donate. In addition, this study expands the literature on human prosocial behavior to support other biological species. Moreover, it provides insight into the applicability of affective computing in evaluating the effectiveness of marketing communications, which is a gap in this research area, and it describes an exper imental paradigm that could be the basis for future investigations. This study is also the first to evaluate advertising materials asking for donations using affective computing, which may help other researcher's assess the feasibility of using such technologies to research charity appeals.

Given the vital role of individual donations for charity organizations, this study has practical implications as the results obtained may be useful for evaluating the effectiveness of their appeals befor e a public release. A substantial contribution for the nonpr ofit sector as a whole is to test the applicability of affective computing, based on FaceReader software, for assessing the effectiveness of charity appeals. This method uses objective measurements and allows for a complex assessment of the benefactors' emotions, even when they are subconscious or in cases of controversial and taboo topics or when the recipients are not ready to reveal their" true emotions—for instance, when under the pressure of some social norms.

The following is an overview of die research examining the relationship between emotions, charitable behavior, and decisionmaking. The literature on affective computing and its applicability in evaluating the effectiveness of promotional materials is also presented.

2. Literature review

2.1. Emotions and pro social behavior

Emotions are discrete automatic responses to different types of events, including universally shared, culture-specific, and individual-specific ones [39]. Numerous attempts have been made in the scientific literature to categorize emotions, separating them according to various criteria [40-44], which has created an extensive list of different emotions. In particular, one of the proposed models suggests the presence of 66 different emotions, 10 of which are primary and 56 secondary [31]. However, it is extremely difficult to estimate such a large number of emotions, particularly considering that similar emotional states may have overlapping parameters. To solve this problem, most researchers have used dimensional approaches [45] to assess emotional state, including valence (unpleasant-to-pleasant) and arousal (low-to-high), as well as six basic emotions (happiness, sadness, anger, surprise, fear, and disgust) [41], which are considered universal and have relatively accurate recognizable patterns [31,46].

Emotional state assessment is widely used to study various types of behavior and decision making. In the context of prosocial behavior, the willingness to donate has been associated with emotional state [14-16]. Donations and helping behavior have been found to be associated with emotions, conceptualized in both discrete [47-49] and dimensional approaches [50,51].

In particular, when anger aims to restore equity, it can act as an effective emotional appeal in charitable communications [19]. According to van Doom et al. [19], people who felt angiy after reading information about the needs of a charity were willing to donate significantly more to compensate for victims' suffering than those who did not feel angry.

Fear-raising charity appeals have also been shown to be effective in attracting donations. Comparing the effects of different types of messages, Hine and Gilford [17] found that exposure to a message about environmental pollution that caused fear contributed to financial donations more than exposure to a control, non-affective message [17]. Another study found that perceived fear is an important predictor of donation discrimination in charities [52] and showed that individuals donate money out of a general perceived fear of a condition experienced by a charity's potential beneficiaries. Another motive may be the opportunity to prevent the fear that they themselves will not be helped in such situations in the future.

Sadness is another negative emotion that provokes willingness to help and increases donation size [16]. At the same time, it is especially effective to use a combination of sadness with positive (pleasant) emotions such as strength [15] or hope. Appeals that combine hope and sadness produce enhanced levels of sympathy, inspiration, and donations [14].

Considering the effect of emotions on donation size in the context of dimensional approaches, most studies have demonstr ated the effectiveness of negative and positive emotions. For instance, Burt and Strongman [20] found that pictures showing negative emotions seem to increase the quantity of donations, including money, items, and time. It has also been shown that individuals are willing to donate more when they see pictures of sad dogs rather than happy ones [50]. Using different manipulations to induce moods, Sabato and Kogut [21] showed that participants' willingness to donate is stronger in a negative mood than in a positive mood [21]. In addition, an EEG study has shown that negative-valence advertising induces more attention to potentially threatening stimuli, as evidenced by the amplitude of the ERP component PI [53].

At the same time, the frequent use of negative emotions in appeals for charity has been criticized [54,55] and can lead to resentment toward charities [56]. Against this background, increasing evidence has shown that pleasant emotions can be effective in attracting donations [23,57]. Specifically, functional magnetic resonance imaging (fMRI) data showed that participants were more willing to donate when the presented images increased die activity of the nucleus accumbens, an area of the brain associated with positive arousal [23]. In the context of affective computing, donations have also been shown to be positively correlated with happiness and the valence experienced before making a donation decision [22].

Thus, special attention should be paid to the effect of arousal on donations. Eye-tracking studies have shown that, as arousal increases, interest in viewing charity ads increases, which, in turn, can increase the likelihood of a donation [58]. According to a survey conducted after a major fire in Australia by Amato [59], people who experienced a higher level of emotional arousal, such as ten or or shock, donated more often to relief charities and showed a willingness to help victims of the fire [59]. However, to the best of our knowledge, the relationship between the amount of donation and arousal has not been studied sufficiently.

Thus, results regarding the relationship between emotions and giving behavior are conflicting. In particular, some studies have shown the effectiveness of positive emotions in collecting donations, whereas others have shown the opposite. The role of individual discrete emotions as a whole also remains unclear as various studies have shown the effectiveness of a range of emotions; this may be due to differences in research paradigms and the conditions under which donation decisions were made. In addition, research examining the relationship between prosocial behavior and arousal is scarce. Hence, prerequisites allow for further studies on the relationship between emotions and donations.

2.2. Using affective computing to evaluate die effectiveness of promotion materials

The face and facial expressions are the most visible and commonly studied information signals for emotions [41,60]. According to Ekman [41], the six basic emotions (happiness, sadness, anger, surprise, fear, and disgust) can be differentiated based on different facial expressions [41]. Measuring facial muscle movements based on emotional patterning allowed Ekman and Fries en [46] to develop a Facial Action Coding System (FACS) for capturing and categorizing facial expressions [46].

The development of computer technology has allowed to automate the process of coding facial expressions, decomposed into action units, and the FACS has become the basis for such software as FaceReader [28], iMotions [27], and others. Some of these algorithms have undergone validation procedures. In particular, FaceReader recognized 88 % of the target emotional labels in the Warsaw Set of

Emotional Facial Expression Pictures and the Amsterdam Dynamic Facial Expression Set [61]. FaceReader uses a deep neural network-based facial modeling technique [62] that describes the locations of 468 keypoints on the face. Then, it analyzes 20 Action Units, described in the FACS [28], and classifies facial expressions (happiness, sadness, anger, surpr ise, fear , and disgust) using a deep artificial neur al network trained over 20,000 images that have been manually annotated. It also pr ovides data according to a dimensional approach to emotions: valence (unpleasant-to-pleasant) and ar ousal (low-to-high) [28].

Automatic emotion recognition systems are now used to analyze mater ials used in marketing campaigns [25,36]. However, the number of published academic studies is small, and some have called for the scientific community to continue resear ch on the predictive abilities of affective computing in the context of the marketing mix [37]. Such systems are relatively inexpensive and quick tools for predicting the effectiveness of an advertisement befor e its pr oduction. Given these pr operties, automated emotion r ecognition systems are likely to reduce the cost of marketing communications analysis for various organizations in both the commercial and nonprofit sectors.

Previously published works can serve as confirmation of the suitability of FaceReader for evaluating the effectiveness of promotional materials. In Hadinejad et al. [36], participants expressed low arousal and positive emotions while watching tourism advertisements [36]. Facial expression recognition analysis allowed researchers to conclude that the presented advertisement, except for some particular" moments, was not overly exciting. When assessing cross-cultural differences in the per ception of product packaging, Ploom et al. [63] showed that the level of positive emotions in Northeast Asians and Northern Eur opeans differs depending on the packaging color; however, they did not find any significant differences in product selection [63]. At the same time, Pichierri et al. [64] tested olive oil advertisements using FaceReader and discovered that perceived healthiness increased consumers' intentions where health claims elicited greater arousal [64].

In a study of shocking and social advertising, FaceReader, in comparison with declarative methods, captured differences in emotional reactions to advertising messages [25]. This allowed the researcher's to conclude that in the context of socially taboo topics, participants are not always ready to declar e their emotions openly, while these can be identified with an analysis of facial expr essions.

Pr eviously, FaceReader" has been used to study the effect of emotions on donation size, where it was shown that valence is positively correlated with donation size [22]. However, these authors did not use any material as stimuli for charity appeals, meaning that no practical applications are possible in the context of analyzing charity advertisements. In addition, Fiala and Noussair [22] indicated that self-reported valence data and the ones measured by FaceReader were not consistent, which also limits the possibilities for fmther application of the results of this study [22].

However", to our" knowledge, automated emotion recognition systems have not yet been applied to charity adver tising campaigns associated with donations. Nevertheless, the use of such systems could help nonpr ofit organizations improve the effectiveness of charitable communication, which is especially important when they are faced with a lade of donations [65,66].

Thus, published resear ch using affective computing covers only a small part of the resear ch questions related to evaluating the effectiveness of certain elements of the marketing mix. In particular, advertising r esearch applicable to the needs of the nonpr ofit sector is lacking. Consequently, pr erequisites allow for fur ther study on the capabilities of automatic emotion recognition systems in the context of evaluating the effectiveness of advertising materials, which served as an additional reason for conducting this study.

3. Methods

3.1. Measurements of emotional states

This study aimed to explore the relationship between emotional state and donation size in the context of stray animals. To obtain data on donations, we presented pictur es of dogs and invited the participants to make voluntary donations to support the animals. To change the emotional state of the participants, we showed them dogs in various states (sick/healthy, domestic/homeless, adult/puppy, with human/without human). The par ticipants' emotional states were measured using eight metrics: valence, arousal, and six basic emotions. These metrics were selected because they are universal and have clearly recognizable patterns. Participants* emotional state was recorded using Noldus' FaceReader softwar e and self-reported emotion. The affective computing method was chosen because it allows a real-time, objective recording of the six basic emotions and the valence and arousal of participants based on their facial expressions, which is a unique advantage over other common emotion measurement methods, such as self-report, skin conductance response, or electromyogr aphy. FaceReader was used because it has one of the highest recognition accuracy ratings among the eight most common automatic emotion recognition software [26]. As an additional measure, we used the emotional self-report to validate the FaceReader" data and assess its predictive ability compared to traditional self-reports. Furthermore, we surveyed the participants* financial well-being to assess the possible impact of this factor on char itable behavior [67].

3.2. Participants

The required sample size (N — 44) was determined based on a power analysis with G-power software [68], which was chosen for a medium effect of 0.40 and a power of 80 %. Therefore, 46 participants were recruited to par ticipate in the study owing to possible data loss. Due to the inability to recognize the FaceReader data, one of the participants was excluded. The final sample size was 45 healthy participants (58.8 % women, mean age = 23.9 year s, SD = 5.3, range — 19-40). The presence of glasses, a beard, and a mustache were a hindrance for participation as, during our pilot experiment, it transpired that these elements can inter fer e with FaceReader software; subsequently, all the par ticipants had short to no facial hair and were not wear ing glasses. Before participating in the experiment, all recruited participants reported that they had never" been injected with botulinum toxin or other dr ugs that affected facial muscle

activity. Participants were mainly students and people from the general public recruited via online advertisements. All procedures were conducted in accordance with the Declaration of Helsinki and were approved by The HSE Committee on Interuniversity Surveys and Ethical Assessment of Empirical Research. Prior to the experiment, the participants were given 320 monetary units (MU) of local cur rency—equivalent to 12 USD, with correction for purchasing power parities [69]. They were instructed that they would have the option to donate this money fully or par tially thr oughout the experiment and that they would be able to keep the remaining amount. Participants were instructed that all their donations were real and would be given to a charity whose name they would learn after participating in the experiment.

3.3. Material and design

Thirty-two pictures were selected from the Internet for the experiment, with two photos for each of the 16 categor ies. These 16 categories were selected by shuffling four" binary characteristics of dogs potentially significant for emotional state and donations: health status, presence of a home, age, and the presence of a human nearby.

3.4. Procedure

The experimental procedure comprised two steps. In the fir st part of the experiment, the participants observed 32 pictures of dogs (for 6 s). The pictures were presented on a 19-inch computer screen. The inter-trial interval was between 12 and 18 s. To reduce the sequential effect of affective valence, we used thr ee different orders of stimulus presentation [70]. To limit the presence of pictures in the same category, they were distributed quasi-randomly; in this way, pictures in the same category were not presented more than twice consecutively. The participants' emotional state was recorded via FaceReader software while they viewed pictures of dogs. After each picture, the par ticipants could make a voluntary donation by choosing one of the scale values from 0 to 10 monetary units. Thus, the participants had 320 monetary units for 32 pictures of dogs on the basis that they could donate no more than 10 monetary units per dog. On the same screen, participants could see the residual amount of money (from the 320 that they received before the experiment) and the counter of pictures (Fig. 1 A).

In the second part of the exper iment, participants were again presented with the same pictures (6 s), after which they rated the valence, arousal, and six basic emotions of each picture (from 1 to 9), without video recording, in a free-time setting (Fig. IB). Three quasi-randomized pr esentation orders were used for the presentation of pictur es in the second part of the exper iment, which differ ed from those used in the first part. A self-reported emotion assessment was performed in the second part of the exper iment to separate it from the donation task and to exclude the influence of conscious r eflection on the emotions that the pictures evoked when making the decision to donate.

To test for the influence of external factors, such as financial well-being, after- the experiment, participants assessed their- level of

EXPERIMENTAL PROCEDURE

A) Part 1 : Donation task

FIXATION SCREEN

current photo 3«! out of 32 CURRENT BALANCE: 305 MU DONATION SCALE (0-10 MU)

12- 18s.

6 s.

response (unconstrained time) time [s]

B) Part 2: Emotional self-report task

_ 8 emotional rating scales: valence,

arousal and 6 basic emotions

0 6 s. responses (unconstrained time) time [s]

Fig. 1. Experimental procedure. A) Port 1: Donation task; B) Pars 2: Emotional self-report task.

financial well-being, satisfaction with the current level of financial well-being, and the significance level of the reward for par ticipating in the study (320 MU) on a 7-point scale. In addition, par ticipants r epor ted their current salary in MU and the percentage of income they spent annually on char ity.

The donations attracted as part of this study were transfer red to a pet charity.

3.5. Data collection and reduction

PsychoPy softwar e [71 ] was used to per form the experiments. A Logitech Webcam Brio 960-001106 webcam with a resolution of 4096p, mounted on top of a computer monitor, recorded the par ticipants' facial expressions. FaceReader™ (version 8.0, Noldus Information Technology) was used for the automatic classification of facial expressions obtained from the video recordings at the sampling rate of 10 Hz and a resolution 720p (1280 x 720). FaceReader software scor ed two dimensions of emotions (valence and arousal), six basic emotions (happiness, sadness, anger, surprise, fear, and disgust), and the "neutr al" state. The values of valence vary from -1 (extremely negative emotions) to 1 (extremely positive emotions). The values of arousal, as well as the 6 basic emotions, He between 0 (inactive) and 1 (active). The par ticipant calibration pr ocedur e used the image or frame with the lowest model error and expressions other than neutral found in this image for calibration [28]. For each display of a picture, emotions measur ed by FaceReader were calculated by subtracting emotional state within 1 s before the picture onset from the value of the aver age emotional state measured during the 6 s of picture display, which neutralized the friction caused by changes in the participants' emotional state before stimuli pr esentation.

3.6. Data analysis

We used the Shapiro-Wilk test to measure the normality of the distributions for all variables and performed Spear man's correlation analysis to reveal the relationships between all dependent variables (with pictures as cases) that were ordinal or not normally distributed. Then, we conducted Pear son's correlation analysis to study the relationships between normally distributed measurement variables and performed quadratic regression to reveal r elationships between valence and ar ousal, measured by self-report as well as FaceReader (with pictures as cases). We considered p values less than 5 % to indicate significance. The false discovery rate [72] was used to correct multiple correlations. We performed a correlation analysis of the data using JASP version 0.11.1.0 and a regression analysis using IBM SPSS Statistics 26.

4. Results

4.1. Manipulation check

Table I presents descriptive statistics of the self-reported (SR) r esponses to die pictures. The donations' distribution was normal (W = 0.956, p = .21). The excellent internal consistency (a = 0.99) of the donation task suggests that all items measured the same construct.