Conceptualization and operationalization of empathy in design research

In this research note, we bring clarity to the concept of empathy in design research by discussing issues in its conceptualization and operationalization. We review literature to identify and clarify the core concepts of empathy and to showcase its potential operationalizations, borrowing from closely related ﬁelds of social psychology and neuroscience. We identify ﬁve core concepts: empathic understanding, empathic design research, empathic design action, empathic orientation

which is a common challenge in cross-disciplinary work (Aagaard-Hansen, 2007).We recognize that in order to combine knowledge and bring in theories and tools from one field to the other, we must ensure that they share a common goal, the context is appropriate, and the goals and operational objectives of the methods are understood (Brewer, 1999).We base the connections made in this note on over four years of collaboration between researchers in design, engineering design, social psychology and neuroscience.Working together helps bridge interdisciplinary knowledge when the fields have different traditions (Karlqvist, 1999).We do not claim to fully bridge the disciplinary gap, but rather start creating new understanding, specifically by clarifying the core concepts (Love, 2002), thus providing opportunities for debate (Jeffrey, 2003) about the science of design.This note has three sections.First, through a review of design literature we define the core concepts (Love, 2002) that make up what is understood by "empathy" in design.We focus on empathy in the sense of designers aiming to understand users and their context, which is a commonly shared aspect of various design process descriptions (Cross, 2001;Horv ath, 2004;Kannengiesser & Gero, 2015;Luo, 2015;McMahon, 2021).Next, we aggregate empirical indicators, i.e., operationalizations, of empathy from social psychology and neuroscience literature.Here, we focus on indicators developed to cognito-affectively understand human-to-human interactions, thus matching their research goals and context to designer-user understanding (Bayazit, 2004;Horv ath, 2004;Love, 2002).Lastly, we map the operationalizations to the core concepts, and point toward adaptations needed to make the core concepts and their relationships testable (Calhoun, 2002;Cohen, 1991;Davidsson, 2016).The last two steps involved iteration and discussion with a senior cognitive neuroscientist colleague, and aim to support theory building in design empathy.

Conceptualizations of empathy
Current conceptualizations of empathy range from "a value to aspire to, cultivate and reinforce" (Heylighen & Dong, 2019, p. 108) to "a professional state" and "a communication process" (Strobel et al., 2013, p. 154), including also related constructs, such as empathic design and its various tools to understand users and make design decisions (Koskinen, Mattelm€ aki, & Battarbee, 2003).This plurality of definitions would benefit from clarification, lest research efforts remain scattered and empathy research in design could reach the messiness of psychology, where "there are perhaps as many definitions as there are authors in the field" (Cuff, Brown, Taylor, & Howat, 2014, p. 144).Education.A few frequently-cited non-peer-reviewed publications were also included (e.g., Leonard & Rayport, 1997).We used keyword search and snowball sampling and included only articles that claimed some contribution toward empathy in design.The articles were qualitatively coded for conceptualizations of empathy, and the codes were grouped using a bottomup thematic analysis method (Braun & Clarke, 2006) with code groups and code assignments iterated on through the constant comparative method (Glaser, 1965).Based on our analysis, we propose five core concepts to depict empathy in design (Figure 1).Next, we present each concept in detail.
A pertinent factor of empathy is an empathic understanding of others, meaning a comprehensive understanding of others' experiences.Several indicators for this type of understanding have been hinted at, such as truthfulness (Walther, Miller & Sochacka, 2017), intimacy and depth (Smeenk, Sturm, Terken & Eggen, 2019b), as well as awareness of dilemmas in a given situation (Hess et al., 2016).Essentially, these articles conceptualize empathy as a specific type of knowledge.Since design is highly situated (McMahon, 2021;Sch€ on, 1983), we posit that empathic understanding, or empathic knowledge, is likewise situated and constructivist.Thus, with empathic understanding outlining the characteristics of empathic knowledge within an individual human, the remaining concepts are used to illustrate the context in which this knowledge is constructed.
Figure 1 The aggregation process of the core concepts of empathy in design

Empathy in design research
Empathic design research comprises the conscious methods used to understand others.This concept is often discussed in conjunction with the term "empathic design" (e.g., Koskinen, Mattelm€ aki, & Battarbee, 2003), with exemplary methods including the designer briefly simulating user experiences and contexts on themselves (Johnson et al., 2014;Kelley, 2015), in-depth user diaries (Mattelm€ aki, 2005), and various ethnography-like methods of inquiry (Wright & McCarthy, 2008).Even several practitioner-oriented reports have highlighted that these techniques, and thus "having access to the user's perspective" (Koskinen, Mattelm€ aki, & Battarbee, 2003, p. 49), are especially useful in the early stages of design processes (Both & Baggereor, 2009;Kelley, 2015).In a conceptual sense, these techniques provide the foundation for designers to understand others, and as such represent human-to-human and human-to-context interactions (Love, 2002).
Next, empathic design action can be conceptualized as the user-centered and generative activities that designers carry out in their work.While little research has explicitly focused on empathic design action, it is often mentioned that empathy would involve the design of valuable solutions from others' points of view (Fila & Hess, 2014;Hess, Strobel & Pan, 2016;Kouprie & Visser, 2009;Smeenk, Tomico, & Turnhout, 2016).Defending the necessity of considering decisions and actions as a part of empathy, Walther et al. (2017) argue that they are integral tools for design to accomplish its purpose, which is to facilitate and enable change and development in society.Together, empathic design research and empathic design action comprise the external aspects (as defined by Love, 2002, p. 350) related to empathic understanding.
Design research has also explored internal factors related to empathy.We conceptualize these factors as empathic orientation and empathic mental processes.The empathic orientation dimension depicts empathy as a conscious preference toward human-centered evidence.This manifested as designers' appreciation toward others' experiences and in designers aiming to understand others.Firstly, regarding appreciation, scholars mentioned the need for designers to avoid judgement when observing others (Kouprie & Visser, 2009), to respect others' responses to emotional stimuli (Heylighen & Dong, 2019), to value others' input into the design project (Postma, Zwartkruis-Pelgrim, Daemen & Du, 2012), and to be inclusive and transparent toward others (Smeenk, Sturm, Terken & Eggen, 2019b).Secondly, studies highlighting empathy as an aim to understand others emphasized the importance of uncovering "true needs and desires" (Walther et al., 2017) and "latent needs" (Leonard & Rayport, 1997).Together, these dimensions indicate a perceived value in learning from others, treating them respectfully, and in being motivated to understand some real and pertinent aspects of others.As such, empathic orientation is close to an epistemology that focuses on humancentered understanding as a powerful way of knowing and reasoning.A few authors take this even further, elevating empathy to a necessary social Design Studies Vol 78 No. C January 2022 competence in engineers and a necessary component in uncovering threats to sustainability, such as structural inequalities (Walther et al., 2017).
Research suggests that empathic understanding involves a plethora of empathic mental processes, both autonomous and controlled.While these mental processes are primarily studied in psychology, several design researchers have acknowledged their existence.These include Kouprie and Visser (2009) discussing whether designers should imagine themselves in users' situations or rather the user in the user's situation, and various papers (Chang-Arana et al., 2020;Hess & Fila, 2016;Smeenk, Sturm, & Eggen, 2019a;Walther et al., 2017) describing how empathy is an interplay between affective and cognitive dimensions.This affective-cognitive interplay highlights the importance of feeling others' emotions and purposefully trying to make sense of them (Kouprie & Visser, 2009) and, for example, using one's own relevant experiences to achieve understanding (Smeenk et al., 2016).This concept differs from empathic orientation in that it focuses on what happens inside one's mind when observing and trying to understand others' experiences, as opposed to the motivation and tendency for one to engage in such processes.Together, empathic orientation and empathic mental processes depict internal and individual (Love, 2002, p. 350) activities that support the development of information into empathic understanding.
In sum, we suggest that empathic understanding is a situated phenomenon effected by designers' external aspects, empathic design research and design action, as well as internal aspects, empathic orientation and mental processes (Figure 2).It is known that the development of empathic understanding involves other elements, such as personal and interpersonal factors, but we leave these factors outside this note.

Operationalization of empathy
While a conceptual framework alone can bring clarity, empirical evidence can bring a novel angle for theoretical claims, and is necessary if making predictions.To enable quantitative analyses, the concept needs to be Empathy in design research operationalized, i.e., translated into observable and testable empirical indicators, after which experimental research designs with proper sampling techniques can be applied (Cash, Isaksson, Maier, & Summers, 2022b).Keys to successful operationalization are reliability and validity of used measures (Calhoun, 2002).Here, reliability refers to the measure producing consistent and repeatable results, while validity includes assessing the degree to which a measure measures what it intends to, including face, convergent, predictive, and discriminant validity.In this research note, we focus on the initial step of translating the various concepts of empathy into empirical indicators, while also commenting on the validity of the proposed measures.In this process, we shift to a more positivistic view of empathy in design, and hence leave qualitative and constructivist measures as well as scenarios where the researcher would participate in the design processes under study (Van Oorschot et al., 2022) outside the scope of this research note.
Given the prior lack of a coherent conceptualization for empathy in design, it is natural that empathy has only rarely been measured or otherwise operationalized in design.The few attempts have used methods borrowed from other disciplines (Alsager Alzayed, McComb, Menold, Huff & Miller, 2020;Chang-Arana et al., 2020), with one exception having developed a survey for engineering designers to assess their perceptions of empathy (Hess et al., 2017).These are a good start, but more is needed to develop and validate design empathy measures.
Given that empathy has been operationalized in many ways in social psychology and neuroscience that share the goal of human-to-human understanding, we begin our search for potential operationalizations of empathy in design by discussing with colleagues and reviewing empathy indicators in these disciplines.We conducted a keyword search for empathy to identify an initial set of articles.We then performed a back and forward citation search for further relevant articles.We included articles that had a clear quantification for any aspect of empathy.The measures were then grouped thematically.From this we identify that current empathy measures can be categorized into six distinct categories (Table 1).The categories were discussed with an expert from social psychology and neuroscience to ensure correct representations of the metrics and categories.Next, we describe these measures and comment on their applicability in design research.
Empathic tendencies are measured via self-report surveys asking how a person would react or feel in given scenarios.For example, a common survey, the Interpersonal Reactivity Index (IRI) (Davis, 1980), asks a participant to rate on a 5 point Likert scale how well a set of statements describes them.Perspective taking is one of the four dimensions in IRI and it asks to rate statements such as "I sometimes try to understand my friends better by imagining how things look from their perspective".There are also other surveys for measuring Design Studies Vol 78 No. C January 2022

Measurement tool
Description of what is measured

Empathic tendency
Interpersonal reactivity index IRI (Davis, 1980) Self-reported tendency for perspective taking, empathic concern, fantasy and personal distress in everyday interpersonal situations Empathy scale (added to IRI) (Jordan et al., 2016) Self-reported tendency to feel what believes the other (person or animal) is feeling Empathy quotient EQ (Baron- Cohen & Wheelwright, 2004) Self-reported tendency to behave in or perception of social situations Toronto empathy questionnaire (Spreng et al., 2009) Self-reported tendency to feel or understand how the other is feeling in everyday situations Positive empathy scale (PES) (Light et al., 2019) Self-reported tendency for positive feelings and reactions in social or life situations Beliefs about empathy Jefferson scale of empathy (Hojat et al., 2001) Self-reported beliefs about how important it is for a doctor to understand patient and their feelings Emotion recognition Reading the mind in the eyes (Baron-Cohen, Wheelwright, Hill, Raste, & Plumb, 2001) Accuracy of inferring the mental state of a person from photographs of a person's eyes

Understanding mental contents
Ickes -continuous rating (Zaki et al., 2009) Continuous, real-time ratings of the valence and intensity of emotions being experienced by others Empathic accuracy standard stimulus paradigm (Marangoni et al., 1995) Accuracy of inferring thoughts and feelings of others when observing a conversation

Shared feeling
Facial synchrony (Light et al., 2015) Occurrence of positive and negative emotions from facial electromyography recorded from corrugator (frowning) and zygomatic (smiling) Motion synchrony (Varni, Camurri, Coletta & Volpe, 2009) Phase synchronization and reoccurrence quantification approach to measure nonverbal cues by motion capture Physiological Synchrony (Soto & Levenson, 2009) Emotional Empathy: "feeling" what the target individual is feeling using one or more of seven physiological signals: (1) cardiac inter-beat interval, (2) pulse transmission time to the finger, (3) finger pulse amplitude, (4) ear pulse transmission time, (5) skin conductance level, (6) finger temperature, (7) general somatic activity MRI (Fan, Duncan, Greck, & Northoff, 2011) fMRI -AI, ACC and MCC parts of the brain, activation of same brain areas for real and vicarious experience (pain) Prosocial responding Empathic responding coding scheme (Rieffe et al., 2021) Observed response to experimenter pretending to hurt themselves in 4 coding categories:" (1) pay attention to the event, (2) show emotional arousal, and (3) show a (verbal or non-verbal) prosocial, empathic reaction" Prosocial behavior (Van Baaren, Holland, Kawakami & Van Knippenberg, 2004) Performance on a simple helping task (pendrop) Prosocial behavior (Lim & DeSteno, 2016) Performance on a non-monetary altruistic decisionmaking task involving book selection and donation Prosocial behavior (Lim & DeSteno, 2016) Donation of up to 1USD in increments of 25c (scale 0e4) to test prosocial behavior Empathy in design research empathic tendencies (see Table 1).However, while the reliability of most survey instruments has been rigorously ensured, their validity in new contexts, such as design, needs attention.For example, IRI's perspective taking items hint that in a design context the scale would focus on empathy towards one's colleagues, with the use of scenarios, such as trying to understand friends and resolving a disagreement.Other disciplines have responded to the need for a more specific version of the instrument by altering the wordings of its items and performing rigorous validity and reliability testing (e.g., P eloquin & Lafontaine, 2010).
Beliefs about empathy are about how a person values empathy or related issues, such as understanding how the other person is feeling.In contrast to the tendency measures, these measures capture how important empathy is, and not how people would react in situations.The Jefferson Scale (Hojat et al., 2001) is a widely used measure for this, comprising Likert-scale reports of one's agreement with statements, such as "empathy is a therapeutic skill without which success in treatment is limited."It has been developed to measure doctor-patient empathy but has since been adapted to other areas of healthcare and to, e.g., law (Williams, Sifris & Lynch, 2016).Belief-focused surveys might be a better than tendency-focused surveys, since there is evidence of all surveys measuring people's attitudes better than tendencies, capabilities, and actual likelihoods of behavior (Krumpal, 2013).
Emotion recognition is about identifying the emotion(s) another person is demonstrating or experiencing.Valence, i.e., whether a feeling is positive or negative, is one measure used here (Ickes, 2001).Another common measure in this category is the Reading the Mind in the Eyes test (Baron-Cohen et al., 2001), where a person is asked to name the feeling a person is demonstrating based on only seeing the eyes of that person.When employing behavioral and performance-based instruments, such as the Eyes test, with rigorous experimental procedures, it is difficult to argue against construct validity: the test measures one's ability to recognize emotions from pictures of eyes, and not, for example, one's attitudes toward emotion-reading.If one's ability to read emotions from pictures of eyes is accepted as an indicator of one's general ability to read others' emotions, the key validity question remaining is that how is emotion-reading relevant in design.This requires a plausible conceptualization of empathy in design, and one that is more detailed in its individual constructs than Figure 2.
Understanding mental contents goes deeper in understanding an emotion by asking a person to guess the other's thoughts.The accuracy of these guesses can be evaluated by comparing them to the other's own recollections and reflections, for example with written reports (Ickes, 2001).These measures of being a good judge of others are statistically not connected to one's results in Design Studies Vol 78 No. C January 2022 other measures of empathy, such as the IRI survey, indicating that valuing empathy may not translate to understanding (Davis & Kraus, 1997).In addition to requiring a conceptualization of empathy in design, the potential validity of these performance-based indicators is largely defined by the experimental setting they are implemented in.Hence, detailed descriptions of the information one is provided about the other and their context is needed when arguing whether an understanding-focused instrument is measuring a construct relevant to design.
Shared feeling is the most common measure of empathy in neuroscience and related fields.A shared feeling is usually measured as similarity of some physiological signals, including facial expressions (Light et al., 2015), bodily motions (Varni et al., 2009), and internal physiological signals such as heart rate (e.g., Kleinbub et al., 2019;Soto & Levenson, 2009).There is no one agreed-upon means to measure the synchrony or similarity between two people, and accordingly many approaches have been developed, including different correlation analysis (Boker, Xu, Rotondo, & King, 2002;Soto & Levenson, 2009), and a 'shared component' in principal component analysis (Kleinbub et al., 2019).While the use of physiological measures is common and they are arguably free of biases, they have not been used in the context of design, except in Chang-Arana et al. (2020).The difficulty in employing these indicators for design research lies again in arguing about the specific constructs they measure and relevance of these constructs, with various other challenges pertaining to specific measurement techniques, such as the fMRI (Hay, Duffy, Gilbert, & Grealy, 2022).
Prosocial responding measures focus on an immediate reaction, desired outcome, or behavior that is believed to be an indicator of empathy.In social psychology, this is called prosocial behavior (Bernhardt & Singer, 2012).Example measures include watching how the participants react to the experimenter seemingly hurting themselves (Rieffe et al., 2021) or dropping several pens accidently (Van Baaren et al., 2004).These measures are experimental simplifications of situations that provide an opportunity for helping or otherwise prosocial behavior, which is thought to be something that an empathic person is more likely to engage in.However, no such scenarios have been developed for or tested in the context of design.The valid definition of prosocial responding is context-dependent, with general mishaps and inflicting pain having been used to assess empathic responding in general (Rieffe et al., 2021), but expressions of emotion having been more common in medicine (e.g., Bylund & Makoul, 2002).Further work is needed to develop design relevant measures for prosocial responding.
Overall, despite empathy instruments and indicators existing outside the field of design, ensuring their validity in the design context requires 1) theoretical and conceptual reasoning for the relevance of the measured construct in Empathy in design research design, and 2) possible adaptation of the instrument itself, potentially requiring a reassessment of reliability.

Mapping existing empathy measures to the conceptual framework of empathy
To support design researchers in developing appropriate operationalizations of empathy, we mapped the measure categories to the conceptualization of empathy in design (Figure 3).This work can be used to guide which areas of empathy to conceptualize further and where to search for benchmarks on their operationalization.Next, we will provide reasonings for our mapping along with practical recommendations for further work in each connection point.
Firstly, the behavioral and performance-based measures of emotion recognition and understanding mental contents map to empathic understanding.These measure if one is interpreting the user or their emotions correctly in a specific situation, leaving any activities that take place prior to or after the interaction outside of their scope.Emotion recognition has been used to assess individuals' reactions to products (Dahlgaard, Sch€ utte, Ayas, & Dahlgaard-Park, 2008;Desmet, 2003) and experimental tasks related to understanding mental contents have been used in studying designer-user understanding (Chang-Arana et al., 2020;Li, Surma-aho, Chang-Arana & H€ oltt€ a-Otto, 2021).We note that designers construct human-centered knowledge longitudinally from multiple sources (Oyg€ ur, 2018), whereas the behavioral and performance-based instruments reviewed here focus predominantly on objective similarities in interpretations.Thus, it would be beneficial to explore measures that emphasize other potential aspects of empathic understanding, such as the depth of understanding or designers' attention to dilemmas in the others' context, and to focus on the longitudinal development of empathic understanding.
We map the self-report measures for beliefs about empathy and empathic tendencies to empathic orientation.Despite common surveys, such as IRI, claiming to measure tendencies and capabilities, we highlight the inherent limitations of survey instruments, in that they are prone to social desirability bias and typically better suited for measuring attitudes (Krumpal, 2013).Indeed, IRI and few others have been used in design (e.g., Alsager Alzayed et al., 2020; Surma-aho, Bj€ orklund & H€ oltt€ a-Otto, 2018), with little acknowledgement to their limitations.Despite IRI's correlates having been validated in experimental studies, it could be argued that its items are not consistent enough for a design context.For example, the items do not ask about a specific target, such as a user or teammates.Hence, we recommend, in the short term, to use self-report instruments for measuring the perceived value of empathy in design, and, in the long term, to invest in developing more design-fit survey measures for empathic orientations, potentially benchmarking other humancentered fields, such as medicine (Hojat et al., 2001).
The other internal aspect measure, shared feeling, maps to empathic mental processes.The processes remain unstudied in design, but the experience in neuroscience and social psychology can help bring a novel angle how to possibly study them also in design.
The prosocial response measures are tentatively matched with the concept of empathic design action.However, it is not known if empathic design actions can be simplified for experimental purposes as done for prosocial action.The simplification assumes the immediate action would be indicative of also real life desired empathic or prosocial action.In design, the challenge is that the decision making in design projects can involve different lengths of incubation periods and the decisions involve multiple stakeholders, domain knowledge, and far-reaching consequences.Thus, we recommend developing new measures or experimental scenarios for empathic design action.
Lastly, we note that no measure in our review maps to empathic design research.Still, there are techniques to characterize interactions where one side is trying to understand the other.These include, for example, various quality indicators of doctor-patient interactions (Hojat, 2016), interactions between married couples (P eloquin & Lafontaine, 2010), and step-by-step instructions for many design research methods (e.g., Both & Baggereor, 2009).As our conceptualization of empathic design research is left broad, progress in this area would require increased detail and specificity in the conceptualization before searching for or developing empirical indicators.

Conclusion
The aim of this research note was to bring clarity to the concept of empathy.We identified five core concepts that together form the overall concept of empathy in design: understanding, action, research, orientation and mental processes.We identify six measure types and map them to the core concepts Empathy in design research of empathy.These two steps, conceptualization and operationalization provide a new perspective for transparent and explicit theory building and development around empathy and its sub-constructs.However, the existing measures for empathy are not developed for and have hardly been tested in the context of design, which presents a validity challenge.Hence, both the concept and the measures require adaptation, use, testing, and validation, to begin developing novel design-relevant operationalizations of empathy and start testing theories related to it.We hope this research note helps design researchers to continue the discourse on what empathy entails in the discipline, and ultimately develop scientific knowledge on the topic.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Thus, we reviewed existing conceptualizations of empathy in design journals, such as Design Studies and CoDesign and in engineering design and education journals, such as Journal of Mechanical Design and Journal of Engineering Design Studies Vol 78 No. C January 2022

Figure 2
Figure 2 Construction of empathic understanding in design

Figure 3
Figure 3 Core concepts and potential operationalizations for empathy in design

Table 1
Empathy measure categories identified in psychology and neuroscience