‘Doing gender and gender equality’ through emotional expressions during a research interview. Views of highly educated Swedish young adults

Abstract

The aim of this article is to contribute to the knowledge on how concepts of gender and gender equality are constructed within research interviews, deepening our understanding of the underlying gender system in society. We focus on emotions and emotional processes expressed during interviews on work and family when specific questions originating in the World Value Survey were asked. Our study is based on interviews with highly educated women and men, in two metropolitan areas of Sweden. In this article, we seek to shed more light on how incorporating emotional expressions and the evaluation of these emotions can grasp the construction of gender and gender equality. We highlight the range of emotional expressions that appear during the interviews, differences in their usage by women and men and the links to the construction of gender and gender equality. We explore how the specific situation of the interview influences ‘doing gender and gender equality’ through emotions. Our results reveal that men and women use similar but also different emotional expressions in conforming to the gender equality norm. Men and women, interviewers and interviewees agreed on this norm, but the ways they ’performed’ the norm are gender based.

Keywords:

• Emotions
• gender
• gender equality
• methodology
• feminism

Introduction

The aim of this article is to contribute to the knowledge on how concepts of gender and gender equality are constructed in a research interview situation on work and family, deepening our understanding of the underlying gender system in society (see Ridgeway & Correll, 2004). We argue that emotional reactions in addition to verbal expressions articulated during interviews should be taken into account in the analyses. This is a novel approach, since studies on the social constructions of gender and gender equality usually focus on what the interviewees say, without paying much attention to how they say it. The latter, however, reveals emotions expressed during a research interview. As emotional expressions are part of communication, the information they provide should not be disregarded. When we include in the analysis the emotional expressions appearing in research interviews, we can grasp the complexity of how gender and gender equality are constructed more fully than via survey responses in quantitative studies or via the verbal arguments alone, presented in interviews.

Our point of departure is that social norms embedded in prevailing discourses related to gender and gender equality influence both verbal articulations and emotional expressions in a specific situation, in this case a research interview. We address the following questions: Can incorporating emotional expressions provide a deeper insight into the construction of gender and gender equality? What kind of emotional expressions appear in interviews on work and family? Do women and men use emotional expressions differently and if so, can this be connected to the ‘doing’ of gender and gender equality? In which way does the specific situation of a research interview influence individuals in how they perform gender and gender equality through emotions?

In the article, we focus on Sweden given the particularly strong emphasis on gender equality in this society, and a hegemonic gender equality discourse (Borchorst, 2011; Holli, Magnusson, & Rönnblom, 2005). Considered as the most gender equal society in the world, Sweden has a long history of high female employment levels, comprehensive reconciliation policies promoting (more) equal share of earning and caring responsibilities among women and men, including fathers’ active engagement in childcare, and women and mothers realizing their aspirations beyond the private sphere of the family while fertility rates remain close to two children per woman on average unlike in most advanced societies (Oláh, 1998; Oláh & Bernhardt, 2008). Hence, Sweden is a particularly interesting context to explore how concepts of gender and gender equality are constructed.

We analyze interviews conducted with highly educated heterosexual Swedish young adults as part of an international research project about gender equity and fertility. A joint interview guide has been prepared incorporating among others certain questions from the World Value Survey (WVS) related to gender roles. It was in this part of the interview where emotions were most strongly expressed, which has intrigued us to include them into our analyses. We have taken into account that the WVS questions were constructed from the point of view of nuclear family based on the breadwinner/housewife model, in line with a (hetero)normative framework frequently used in studies addressing gender roles (Borchorst, 2011). They relate to the idea of a cultural system of femininity and masculinity, in which gender is seen as a sociocultural product appearing in individuals’ attitudes, ideas, believes and experiences of gender, rather than a relatively stable and fixed individual attribute (Luyt, 2015).

Previous research

In the 1970s, feminist researchers became increasingly interested in women’s experiences and the methods to grasp them, especially interviewing and participant observation. Applying these methods to different groups of women, the ‘rule’ of objectivity in social sciences has been increasingly questioned. Self-reflection has been used as a tool to reveal the subjectivity of research, taking into account that researcher’s age, class, ethnicity and sexuality also matter with respect to the research process and outcomes (see for example Clarke, 2005; DeVault & Gross, 2012; Kleinman, 2007). In addition, the ‘personal feeling’ of researchers were considered as part of self – reflexivity. The links between emotions and gender have thus been of interest for feminist research, both theoretically and methodologically (Blakely, 2007) since the late twentieth century.

Early studies focused on women’s experiences also highlighting differences between men and women’s emotional landscape (see for example Belenky, McVicker Clinchy, Goldberg, & Mattuck, 1997). Related to that, stereotypical views of women as emotional and men as rational have been reinforced even if unintendedly (Lewis & Simpson, 2007). A binary division of emotions into masculine and feminine, which also is a feature of gender essentialism, has been accompanied by a hierarchical view, in which attributes connected to masculinity are more highly valued than those linked to femininity. In contrast, post-structural feminists consider emotions as cultural resources which produce and reproduce gender (Butler, 1994; Sass, 2000). Specific emotions and emotional expressions are seen as aspects of gendered norms to which individuals relate when they perform gender.

Emotions are also connected to interactions and situations. It is in specific situations that individuals perform gendered norms through verbal and emotional expressions. Interactions as well as situations have their own ‘feeling rules’ in which emotional expressions are regulated by a ‘script’. This ‘script’ can be openly articulated through for example written work instructions. The ‘script’ is often hidden, however, as people are aware of expectations on how they should behave in a given interaction and/or situation displaying certain emotions but not others, or it is presented only verbally, for example by an employer to employees (Hochschild, 1983).

Recently much research addressed how emotions are regulated at different workplaces, in organizations and in relation to different professions (Coupland, Brown, Daniels, & Humphreys, 2008; Fineman, 2000; Lewis & Simpson, 2007), but the knowledge on how emotions are performed within a research interview is very limited, notwithstanding burgeoning interest related to this topic (Davies & Spencer, 2010; Flam & Kleres, 2015; Gilbert, 2000). A possible explanation for why emotions in research overall, and specifically in relation to research interview have been neglected is that there exists a ‘hidden agreement’ regarding how this interaction should be performed. According to the ‘script’ the interviewer should be as ‘neutral’ as possible, not showing any emotions and the interviewees should simply answer the questions (Fox Keller & Logino, 2006). According to the social norm surrounding the interview situation interviewees’ emotional reactions are permitted, but should not be taken into account in the analysis. It seems that there is an ideal in the academic world that emphasizes the spoken and written word at the expense of the body and emotional language (Hellum, 2002, 2010). Although emotions in the interview situation are expected to be disregarded, it is also advised to have sympathy, closeness and empathy with those interviewed (Blackman, 2007; Kleinman & Copp, 1993).

In previous research, two aspects are addressed: emotions of the researchers and those of the interviewees’ twentieth-century feminist scholars emphasized the importance of understanding both the researcher’s and the interviewee’s emotions in order to understand how gender and gender (in)equality are produced and reproduced (Behar, 1996; Kleinman, 2007; Wincup, 2001; Wolf, 1992). After the post-modern turn in the end of the 1980s many, especially feminist researchers emphasized the importance of emotions emerging in interviews and observations (e.g. Blakely, 2007; Foster, 1994; Lee, 1997; Lumsden, 2009; Skrinjar, 2003; Soilevou Grönnerud, 2004). Such studies, however, often failed in how to use emotions, which occurred in interviews and observations, analytically. Emotions mostly become ‘emotional footnotes’ (Hellum, 2010), accounted for in methodological chapters or in footnotes, but not actively used in the analysis (see also Blakely, 2007).

Lately more attention has been paid to how researchers’ emotional self-awareness contributes both to constructing and understanding data and to generate knowledge (see for example Holtan, Strandbu, & Eriksen, 2014; Kleinman, 2007; Wasserfall, 1993). Recent research provides examples on how emotions in interviews or observation situations can be used in an analytically fruitful way. Hellum (2002) and Bergman Blix (2010) have developed ways of how to use their own emotions as a methodological tool. As argued by Bergman Blix and Wettergren (2015), researchers do emotional labor, which include strategic emotion work, emotional reflexivity and their emotional work aimed to cope with their emotional dissonance. These studies focus on the researchers’ own emotions, emotional work and labor in the social interaction, but not on how the researcher and the respondent together create meaning. Soilevou Grönnerud (2004) in contrast, views the research interview as an interactive situation. In her article, she analyses how laughter creates certain interactive patterns, such as impression management, problem revelation and coping with stress in the interview situation. Only a few studies focus on the emotional process emerging as part of an interview, however. Also, it remains an open question, how this emotional process can be analyzed in a fruitful way (but see Marander-Eklund, 2000; Moran, Skeggs, Tyrer, & Corteen, 2002; Soilevou Grönnerud, 2004).

Theoretical frame

In order to understand how emotions are a part of ‘doing gender’ in the situation of a research interview, it is fruitful to incorporate the idea of the ‘performative gender’ (Butler, 1990) into the concept of ‘doing gender’ (West & Zimmerman, 1987, 2009). According to West and Zimmerman (2009), ‘doing gender’ refers to an ongoing process of recognizing and making gender accountable according to ‘current cultural conceptions of conduct becoming to -or compatible with the “essential natures” of- a woman or a man’ (p. 114). They view accountability as the basic characteristic for social relationships, and consider gender accountability to be both interactional and institutional. Hence, femininity and masculinity are moved from the concept of a natural biological ‘sex’ to that gender accountability creates in interactions, which is gender are ‘done’ when people make sense of ‘historically specific institutional and collective practices’ (West & Zimmerman, 2009, p. 115). Hence, individuals can act emotionally and thereby ‘do’ gender differently in different interactions and situations.

While West and Zimmerman emphasize the importance of the interaction in ‘doing’ gender, Butler’s (1990) main concern is how discourses define bodies as either female or male. She argues that it is on the body’s surface gender such as femininity and masculinity becomes visible through acting, gestures and other signs. Thus, Butler (with reference to Foucault’s work) views the body as ‘text’ in as much both communicate meaning, which is expressed in conversations by verbal articulations and the body’s acting and gestures together. Acting and gestures create a vision of an essential gender identity, both in the ‘self’ of the actor and in discourses which regulate and discipline the surface of the body, articulation and emotions, such as in the statement ‘women are more emotional than men’. This statement is part of a discourse, and according to Butler (1990, p. 173f) it is when this idea becomes a natural belief and/or a natural feeling as for example a person who position herself as a woman ‘feels’ that she is more emotional than a man, that the regulation and disciplinary practices truly have succeeded. When it comes to gendered emotions in a post-structural perspective such as Butlers’ term of the ‘performative gender’, accounting for above, emotional expressions are regulated through discourses, which again influence how individuals act emotionally. Some emotional expressions become connoted as ‘female’ and others as ‘male’ (within female and male discourses), and individuals act in accordance with how they position themselves as ‘male’ or ‘female’, and/or to how the interaction and situation are emotionally defined. Though West and Zimmerman (1987, p. 137) bring forward that enacting is a part of ‘doing gender’ they do not discuss how available discourses related to gender accountability both are ‘written’ into bodies surface and become embodied into individuals’ ‘self’. Butler on the other hand does not take the importance of the interaction and the situation into account.

According to Hochschild (1983) emotional work refers to individuals’ emotional expressions (surface acting) or to emotional experiences (deep acting) adjusted to culturally defined discourses and interactional or situational ‘feeling rules’. This distinction between surface and deep acting has been questioned by for example Bergman Blix (2010), because in reality they are interdependent. Sometimes there is concordance between the two, but not always. It is according to Butler (1990, 1994) a question on how regulating and disciplinary discourses influence both the surface acting on the body and the deep acting within the body and to how this is performed according to ‘feeling rules’ in interactions and situations.

Emotional labor on the other hand is connected to how professions and situations demand expressions of emotions as part of work descriptions, for example, how researchers are expected to act in an interview situation. It can be argued that if a behavior should be classified as emotional labor, it has, in addition to being paid for, to be explicitly scripted, either in professions and working-place instructions, or as verbal information. Sometimes it is quite difficult to recognize how the emotional labor is ‘scripted’, since it is ‘taken for granted’ in the profession or in work places, but it is always possible to observe how emotional labor is performed (Hochschild, 1983), for example how researchers/interviewers are acting during an interview. The interviewees’ emotional work defined as surface acting is not as regulated as the researchers’ emotional labor, though there are ‘feeling rules’, i.e. how the interviewer and the interviewees are expected to behave.

The research interview is an emotional setting constructed in line with Hochschild’s ‘feeling rules’, meaning that there are often not explicitly written or spoken expectations that emotions should not be displayed. It can be argued that the ‘script’ of a research interview is defined in relation to a specific ‘emotional regime’ (Barbalet, 2001), in this case the organization of ‘normative’ research defined as ‘rational’, ‘neutral’ and ‘objective’. Incorporated in the term ‘emotional regime’ are also organizations and settings where the norm is not to show visible emotions, such as the court and judiciary (Roach Anleu, Bergman Blix, & Mack, 2015) and research. Although the research interview is often defined as ‘neutral’, ‘objective’, with ‘focus on the articulated’, in order to get data as successfully as possible, a form of strategic emotional labor from the interviewer is recommended (Bergman Blix & Wettergren, 2015) to produce an environment of sympathy, closeness and empathy. In addition, the interview situation is also influenced by how women and men are supposed to behave emotionally. In other words, there are also gendered discourses connected to emotional behavior in research interview situation (Reinharz & Chase, 2001; Schwalbe & Wolkomir, 2001).

Method

This article relies on material from an interview study conducted in 2012 in Gothenburg and Stockholm, as part of an international study with highly educated heterosexual young professionals. The interview guide, translated to the native language for each country in the study, was rather strictly structured. In order to carry out 80 interviews in a short period, four assistants were employed (all heterosexuals), a man and a woman in Stockholm and two women in Gothenburg, all of them Master’s students. They conducted nearly all interviews, which were recorded, and transcribed. A post-doctoral researcher conducted a few interviews in Gothenburg. All interviewers were instructed to include the occurrences of emotional reactions, both those of the interviewees and the interviewers, into the transcriptions. Laughing, sighs, pause, irritation were some of the reactions observed from the interviewees. The interviewers were instructed to ask the questions as they were written in the interview guide and not to contribute with own comments, questions or in any other ways express their opinions or feelings. Thus, they were instructed according to the ‘script’, that they should be acting as ‘neutral’ as possible with respect to emotional labor (as explained by Hochschild, 1983). As the analysis will show, even though the interviewers tried to remain ‘neutral’, they could not.

The interviewees consist of 40 men and 40 women with post-secondary education, aged 24–35 years, born and/or brought up in Sweden. As for the interviewers, the two women from Gothenburg were in their early 30s. In Stockholm both interviewers were in their late twenties. Thus, the interviewers belonged to the same categories as the interviewees in terms of age and education.

We relied on the method of snowballing in our search for heterosexual men and women, self-identified as such, with the above specified characteristics to be interviewed. Snowballing is based on one interviewee suggesting further informants to include in the study. In order not to get a sample consisting of particular groups of friends or colleagues, a maximum of 2 interviewees could be suggested by one person, and at most 3 interviewees could work at the same work place. These rules ensured to get widest possible diversity among the interviewees.

The excerpts, used in this article are illustrations on the theme of the analysis. In line with rules of confidentiality, we use pseudonyms, which are known male and female names in Sweden, but not necessarily elsewhere in the world. Hence, we include an (f) or (m) after the name to distinguish between females and males. For confidentiality reason we do not make a distinction between the interviewers either in excerpts or in analyses of the interviewers. They will be referred to simply as ‘Interviewers’.

Techniques of analysis

In order to grasp the emotional work done related to the WVS section, we started to quantify the emotional reactions that occurred. We discovered quickly how problematical this was (see also Haakana, 1999, p. 72ff; Soilevou Grönnerud, 2004), as the interviewers were asked to simply transcribe emotions for example laugh (as laugh) and sigh (as sigh), but not to transcribe how the interviewer and the interviewees used emotional expressions which could vary widely (such as strong laugh, silly laugh, relieved sigh or relinquished sigh). Thus, the different ways of laughing and other emotional expressions could not be identified by just reading the transcripts. Therefore, we also listened to the audio files in order to find out how they laugh for example. It turned out that there were so many ways of laughing, using different voices, sighing and so on, that it was not possible to quantify all the diversity.

The lack of clarity regarding the content of the emotional expressions in the transcripts notwithstanding, we started the analysis by counting such expressions to see whether there was a difference in using emotional expressions by men and women. The next step was to go deeper into what kind of emotional expressions were used, in what way and by whom (interviewers, interviewees, male or female). Emotional expressions were then connected to words and sentences in order to detect how emotional expressions and words together created a meaning in relation to how the interviewees and interviewers (re)produced gender, and acted in line, or not, with the social norm of gender equality. In our coding process, we followed the logic of situational analysis (see Clarke, 2005), viewing the research interview as situated situation in which codes, and based on them categories are produced in an interrelationship between our experiences as researchers, the aim, research questions and the theoretical rationale of the project. Our analysis developed from a more inductive coding of the interview data, also recognizing how our research questions and theoretical background guided us in this process. As pointed out above, our interviewees and interviewers defined themselves as heterosexual men and women, as well as we as researchers define ourselves. This has probably influenced how we categorized and analysed the data, being thus active actors of (re)producing the meaning of gender in relation to the underlying gender system (see Clarke, 2005; DeVault & Gross, 2012; Kleinman, 2007). The inductive coding was done in NVivo, a software program designed for in-depth qualitative analyses, like ours.

Analysis

One way to see how gender and gender equality are ‘done’ is to take into account how people answer questionnaires. According to Luyt (2015), such information should not be seen as respondents reflecting on gender ideology but as situated, depending on situations. When respondents interpret questions in questionnaires, they rely on available (gendered) discourses as well as social norms. Indeed, based on how our informants reacted to the World Value Survey (WVS) questions, we found that they (tried to) position themselves in line with the pro-gender equality discourse. Despite small differences between interviewees’ responses to the WVS-statements (the response alternatives were: strongly agree, agree, disagree, strongly disagree, don’t know), there are many other ways to reveal their inherent opinions, both with respect to emotional reactions and the verbal expressions used. Strongest reactions were shown on the WVS questions with explicit pro-breadwinner ideals such as: men being better executives than women, men having priority when jobs are scarce, and statements related to the male breadwinner and housewife as specified below.

Constructing gender and gender equality through an emotional process

In order to show how concepts of gender and gender equality are constructed through emotional and verbal expressions between the interviewer and interviewee, we first present general observations from the interviews before analysing excerpts taken from the part of the interviews that focused on the WVS statements. This will be followed by a deeper analysis related to different meanings of emotional expressions.

Whenever the informants reacted to the WVS questions, they used many different emotional expressions, sometimes even in relation to the same question. The emotional process could start with an interviewee rather ‘calmly’ reflecting on the statement, possibly getting irritated after a few similar statements, asking for a break or starting to laugh. Some interviewees started with laughing, calmed down, became reflective, started ‘sighing’, or asked for the question to be repeated. Especially laugh, which was the main emotional expression, had various meanings, which we will get back to. Anger or irritation was rarely expressed in the research interviews, perhaps because the ‘feeling rules’ do not permit emotional expression considered to be negative (Barbalet, 2001; Hellum, 2010). Negatively connoted emotions, such as anger, fear and irritation, are not in line with the ‘script’ of sympathy, closeness and empathy related to a research interview either. Interestingly, only women expressed anger or irritation openly, none of the men. Gender equality is primarily a ‘women’s project’ (Inglehart & Norris, 2003), and the females in our material seem to have become more provoked by the gender-unequal statements than men.

In our material, men appear to try not to express any emotions, though there are reactions such as asking for repeating the questions and laughter. Men also seem to try to present themselves as more ‘rational’ compared to women who used more explicit emotional expressions. This behavior could be observed in men staying ‘calm’, giving ‘short answers’ or avoiding to show ‘any visible emotional reactions’ while answering the questions. This practice of short answers accompanied by some emotional expressions was more common among the male interviewees in Stockholm than in Gothenburg. In contrast in Gothenburg, men could give rather long ‘calm’ answers without showing any visible emotions. This difference may be partly related to the fact that in Stockholm a man conducted the interviews with male interviewees, while in Gothenburg a woman interviewed men (see Reinharz & Chase, 2001; Schwalbe & Wolkomir, 2001).

In the following excerpt, the interviewer and the interviewee together construct (hetero)normative gender equality using both emotional and verbal expressions in conjunction:

Interviewer:

(…) ‘A father cannot be fulfilled unless he is the primary breadwinner, meaning that he makes more money than his wife.’

Anna (f):

Four. [the number of the relevant response alternative]

Interviewer:

Maybe you could say the entire response with words?

Anna:

Strongly disagree! Oh how I shouted! (laugh)

Interviewer:

(laugh) Then it will be heard well [in the recordning]. Why not?

Anna:

No, it is an individual thing, some feels good of working, others do not feel good of working. Some fathers want to take parental leave.

Interviewer:

Then I have another statement. (…)

Anna:

Yes (laugh)

Interviewer:

Yes, then you should answer?

Anna:

Yea, yea, go on!

(…)

Interviewer:

It is not me who should speak!

Anna:

No! (laugh)

The excerpt reveals that fatherhood is a multifaceted concept in contemporary Sweden, as it offers the possibility to either prioritize paid work, or being a caring father. It also shows how the interview situation is influenced by ‘feeling rules’ connected to the ‘script’. Anna shouted, and excused herself for that. The interviewer tried to ‘neutralize’ the situation both by laughing and by saying that it was fine that she shouted as it will be easy to hear the answer on the tape. In Soilevou Grönnerud (2004) terminology this behavior from the interviewer would be named as ‘coping with stress’ laughter.

We instructed our interviewers, in line with the research interview ‘script’, that they should focus on the interviewers’ answers and hold on to the questions in the interview guide. The interviewer is conscious of the ‘script’ and it seems like she is not sure if Anna would follow the ‘hidden agreement’ of a research interview, and therefore got stressed. This may be the reason why the interviewer reminds Anna several times, that it is she who is expected to respond. The interview continues:

Interviewer:

(laugh) ‘A working mother can establish just as warm and secure a relationship with her children as a mother who does not work.’

Anna:

1, strongly agree.

Interviewer:

Why?

Anna:

Because it is not the question of how many hours you are away, but of quality time; it is not about quantity, but of quality, I think.

Interviewer:

Generally, do you think that fulltime working mothers are looked at differently by society compared to mothers who do not work? And you should not use the response card, just answer please.

Anna:

Oh! (sigh) Society? (…) I do not know, I cannot make a statement because I have not made up my mind about that myself.

Interviewer:

But is there anything you have noticed? How they are looked at? In some way?

Anna:

No.

(…)

Interviewer:

OK. Here comes another statement (coughing) ‘When children are age three or under, a married / cohabiting woman should be a housewife and focus on childcare.’

Anna:

When a child is three years old?

Interviewer:

Three years or younger.

Anna:

Or younger??

Interviewer:

Should I read it again?

Anna:

Mm.

(…)

Anna:

These questions are a little bit provocative! I have to say 3. Disagree. I think it has to be a little bit of collaboration. I cannot say ‘strongly disagree’, neither can I say ‘strongly agree’ (…). I think it should be (…) a little gender equality. Often it is the woman who takes more responsibility. For some reason (…). Yes.

There is no doubt that Anna finds these gender unequal statements provocative. In the emotional process-interplay between her and the interviewer, they distance themselves from the non-gender-equal discourse by using emotional expressions such as laughter, and Anna becoming almost angry. According to Borchorst (2011) and Holli et al. (2005), the hegemonic gender equality discourse is strong in Sweden, and it is this discourse that both Anna and the interviewer identify themselves with. It can be argued that the discourse is so dominant that it has been embedded into individuals’ ‘emotional selves’ using Butler’s (1994) terminology: When this ‘natural belief’ is challenged by, in this case, gender unequal statements, individuals react emotionally. However, both Anna and the interviewer seem to be concerned with how to use their emotions in the interview in line with the ‘feeling rules’.

Different ways of laughing

Many interviewees, both men and women laugh in relation to the statement: ‘On the whole, men make better business executives than women do.’ Indeed, ‘laughter’ is the dominant emotional expression in the interviews, but it was used in different ways. Anna in our previous example laughed in order to excuse that she shouted. The interviewer’s laugh can be interpreted as what Soilevou Grönnerud (2004) call ‘mutuality through laughter’. This kind of laugh is used to help the interviewee to go on with the story. In many interviews, we observed using laughter this way.

Camilla (f):

(Laugh) Yea, I do not think so! I strongly disagree.

Interviewer:

Why do you strongly disagree?

Camilla:

I think that women can be better, because we are better at the soft values. (…). I really think that women are better on that (to make the employees feel good at work), that the male domination we have today, a patriarchy, so I think … that women who are in leading positions in general use the male way of leading! That they have succeeded because they have been like men. But I think we can start to see a more female leadership, or what you may call it. More soft values, where both men and women are acting in line with the soft values. And then … the employees are happy and feel good. This is how I see it.

Camilla started with laughing and was certain that she strongly disagreed, but in the emotional process she calmed down and reasoned about her answer in a more reflexive way. In her explanation, she emphasized typically male and female qualities, such as women are better at ‘soft values’ and men represent ‘harder values’. Soft values can here be interpreted as ‘emotional’, and hard values as ‘rational and logical’. When it comes to leadership, she also points out that those women who are successful as leaders act like men, i.e. they become rational and logical.

Another way of using laughter is seen for Amanda (f). She agrees with the statement:

Amanda:

Yes, unfortunately! (laugh).

Interviewer:

Yes. (laugh) Why?

Amanda:

Experience … no, I think that women are more emotional and they are not so much … emotionally … they are controlled by emotions rather than the logical thinking. And if you lead you need more of the logic than the emotional, and men are generally better on that.

Compared to how Camilla reacted with laugh (‘because this is silly’), Amanda excuses her opinion with her laugh. Because of the dominance of the gender equality discourse in Sweden, it is not ‘politically correct’ to say that men actually are better executives than women, something which Amanda does, but she also excuses her statement both with the word ‘unfortunately’ and with an excuse laugh. As mentioned above, Anna also used laughter as an excuse, but in relation to her shouting during the interview.

As Amanda goes on with her argument, she also makes a distinction between men and women, viewing men as more logical and women as more emotional. When Camilla and Amanda ‘do’ gender, although they do not agree on the subject, they use a common discourse on how women and men ‘are’ (Belenky et al., 1997); that is, women are emotional and soft and men hard and logical. They also agree on how leadership is defined as logical, hence being a man’s world overall. But they differ in how they position themselves. Camilla says that women actually are better leaders because they represent ‘soft’ qualities, Amanda on the other hand points out that women are inferior leaders because they are emotional.

It is interesting to see how the interviewer (she is the same for both interviews) positions herself as pro-gender equal in these cases. Camilla is pro-gender-equal in her first answers and the interviewer seems to agree with that opinion, and continues without question or reflecting on the next question. But in Amanda’s case, who expresses herself as against the gender equality discourse, the interviewer laughs indicating ‘mutuality through laugh’. That laugh can also be interpreted as astonishment, a polite admiration and condemnation to the ‘politically-not-correct’ statement from Amanda (see Soilevou Grönnerud, 2004, p. 37 for similar interpretation in one of her interviews).

Irony

Another important emotion is irony. It can be indicated by laughter as in the following excerpts between Anders (m) and the interviewer (who is a man).

Interviewer:

Mm, ok. Eh next statement. ‘When jobs are scarce, men should have more right to a job than women.’

Anders:

(laugh) I would say I agree, because that would benefit me.

Interviewer:

[laugh]

Anders:

But I do not agree on that, not at all.

Interviewer:

Mm. And …

Anders:

But it would be, yea, just for my own personal benefit, but No.

Interviewer:

And what does that depend on, that you …

Anders:

Well it is that one sees ‘karlar’ and ‘fruntimmer’ (see explanation of the terms below) as equal.

Both Lee (1997), Skrinjar (2003) and Soilevou Grönnerud (2004) highlight the importance on how the organization of gender in the interview situation influences the interview. In the previous excerpts discussed above, the ‘doing’ of gender takes place between female interviewers and female interviewees. In this latter excerpt the conversation relates to two men and the interview situation seems to be influenced by both the dominant pro-gender equality discourse and of a ‘laddish tone’ of joke and irony expressed by Anders. He also uses the words ‘karlar’, which can be translated to English as ‘masculine men’ and ‘fruntimmer’ which originally had a positive connotation used for women who had the responsibility for the home, but nowadays has a rather degrading connotation. Hence, there is ambivalence in Anders’ answer, presenting himself as pro-gender equal but also using words which indicate a qualitative difference between women and men with men being more highly valued. In addition, this way of talking, that is using these words along with jokes and irony, is connected to a hierarchical non-gender-equal masculinity unlike the pro-gender-equal masculinity that the interviewer tries to maintain (see Connell, 2005 for discussion on changes in masculinities).

Irony can also be used in relation to the object of discussion instead of self-irony. Julia (f) even uses an artificial voice in order to show how ridiculous the statement is:

Julia:

Yes, but but yes I will … I (…) really hope in the future that that statement will change over time, that it will not be like that. That the man feels that he has to be the provider. And that they do not feel ‘satisfied’ (artificial voice) if they do not have the outermost provider … what did you say?

Interviewer:

yes, for … the outermost provider responsibility, that he earns more than the mother so to speak. But it is something then that … you agree with to a certain extent? (…) you know, that men …?

Julia:

that they feel like that, yes. Yes exactly.

Indeed irony is used in many ways. Irony can consist of certain elements of humor, when the ironic utterance is followed by laughter (see for example Partington, 2007; Soilevou Grönnerud, 2004). It can be argued that it is this kind of irony that becomes visible in Anders’ answer. Another way of irony, used both by Anders and Julia, is to say the opposite of what one really means. Usually ironic utterances are used in relation to someone or something (Burgers, van Mulken, & Schellens, 2011). In our case, the target of the irony in both Anders and Julia’s answers is the male breadwinner/housewife model, but the way they do it is gendered. Anders uses a hierarchical non-gender-equal masculinity in his irony to present himself as a reliable pro-gender-equal man. In contrast, Julia and the interviewer agree that men are still not really gender equal, and this agreement among the two women is demonstrated in their half-finished sentences.

‘Repeat the question’

A general pattern observed in the material is that ‘laughter’, ‘pause’ and asking to ‘repeat the question’ predominate among the practices that the interviewees use. However, a closer look shows that female interviewees use laughter more often than male interviewees do. An opposite pattern can be observed regarding the ‘repeat the question’ practice, as male interviewees tend to use it somewhat more often than women. In the next excerpt from the interview with Bengt (m), there are both pausing, focusing on something else/buying time and request to repeat the question.

Interviewer:

‘What do you think about the idea that in a couple the man should work outside home and the woman should focus on domestic work?’

Bengt:

Eh. It sounds very stereotypical, doesn’t it?

Interviewer:

Why?

Bengt:

It sounds old fashion I would claim. It feels like it is an old stereotype. That it is something from the past. Not like it is today. It feels like both [parents] are expected to share [caring for children] and to work.

Interviewer:

Yes exactly.

Bengt:

Can you wait for two seconds,..[ interviewer’s first name].

Interviewer:

Absolutely.

Bengt:

It sounds like somebody forgotten to turn off the tap in the bathroom.

Interviewer:

(laughing) Eh. I am going to read some statement for you. And for every statement you can say if you strongly agree, agree, disagree or strongly disagree.

(Pause from the interviewer)

Interviewer:

‘A father cannot be fulfilled unless he is the primary breadwinner, meaning that he makes more money than his wife.’ Would you tend to agree or disagree with this statement?

Bengt:

What did you say? He cannot feel fulfilled? (…)

Interviewer:

Eh! Cannot be fulfilled unless he is the primary breadwinner. Would you tend to agree or disagree with this statement?

Bengt:

Agree partly, isn’t it?

Interviewer:

Mm. (pause) Why?

(…)

Bengt:

It is the ideal in a way, it fits better the picture you grow up with, that is why you partly agree, sort of.

Interviewer:

Ok.

Bengt:

Isn’t it a little bit more masculine, isn’t it like this? It is more male sort of to own a lot and take care of your family: Isn’t it like that?

Interviewer:

Yes.

Bengt. Sort of be the man (karlen) in the house with the money sort of, doesn’t it?

Interviewer:

Yes exactly.

The interviewer herself writes in her reflections that the interview ‘felt’ calm and relaxed. But she adds: ‘another problem I discovered, is that it is difficult not to react or reveal my own values when the interviewees answer’.

She realized that she was struggling with the ‘feeling rules’ of the interview. At the same time she was doing emotional work. It is possible to see this struggle between emotional labor and emotional work in the way she replies to the male interviewee with short answers and ‘mumbling’. This type of acting from the interviewer was also used in other interviews, as seen in the interviews with Amanda and Anders above. It is used in situations when the interviewers are struggling, especially when the pro-gender equality discourse is challenged.

The interview with Bengt is different from the other interviews presented. Here, the interviewer is female and the interviewee is male. Schwalbe and Wolkomir (2001) describe how gender construction is performed in a research interview, considering also power relations. They point out that men use various strategies whenever they find it difficult to signify control, autonomy, rationality and sexual desirability, in order to maintain their masculine self. It is not uncommon, they argue, that in order to control the interview situation men use strategies such as testing, sexualizing and minimizing the interviewer. In the excerpt above, Bengt tests the interviewer by using ‘isn’t it?’, ‘doesn’t it?’ in relation to his answers which are not in line with the pro-gender equality discourse. He seems to seek confirmation from the interviewer that he is right even though his argumentation does not correspond to the dominant discourse of gender equality. The interviewer pretends to agree with him as to avoid tensions, although as she points out later in her reflective notes, she did not really agree, and felt it was difficult not to ‘react’ and reveal her true opinion At the same time, he also minimizes her, both by using the interviewer’s first name several times, and ‘buying time’ by going to the bathroom. It can be argued that Bengt tries to control the interview situation maintaining his masculine self. In this case the masculine self is built on a heteronormative idea of gender binary between masculinity and femininity, seen as fundamentally different from each other.

Conclusions

The aim of the article is to contribute to the knowledge on how gender and gender equality are constructed within research interviews by focusing on the emotional expressions occurring during the research interview. The process of ‘doing’ gender and gender equality is complex, as it involves discourses and emotional work, and also depends on how the situated situation that included emotional labor is ‘scripted’. It can be argued that in order to understand how gender and gender equality are constructed, a threefold analytical model should be used intertwining the three dimensions: (1) Discourses: here exemplified by the strong Swedish gender equality discourse, discourses connected to femininity and masculinity including gendered emotions, and discourse which defines how the research interview should be performed; (2) Emotional work: seen in the difference between how men and woman interviewees did their emotional work. This may be related to the discourse which describes ‘women as more emotional’ and ‘men as more rational’, but also to the discourse which defines how both informants and the interviewer are expected to behave emotionally in a research interview; (3) Situated situation, in terms of how the situation is defined and organized. The research interview is influenced by ‘feeling rules’ in Hochschild’s term, as it is (expected to be) ‘rational’, ‘neutral’ and ‘objective’ where no emotions should be expressed. It can be argued that the research interview is an ‘emotional regime’ (Barbalet, 2001; Fineman, 2003) which is male connoted, because the ‘feeling rules’ related to this specific situated situation is strongly connected to a specific form of masculinity, the logical, rational and objective. Our male informants seem to have acted more in accordance with this definition compared with our female informants. At the same time, our interviewers tried to stay loyal to the ‘script’ defining the situation, as they were instructed. The ‘script’ of the research interview used in our study is in line with ‘normative’ research, that is, an ‘emotional regime’ which defines research as ‘neutral’ and ‘objective’. Though our interviewers tried to act according to the ‘script’, using different emotional expressions they played an active role in maintaining the gender equality norm.

By incorporating emotions, most specifically the emotional expressions, we got another picture than by simply looking at the articulated (verbal) information. Most of our interviewees agreed with the pro-gender equality norm defined as a ‘dual earner-dual carer model’, but the way they emotionally performed this agreement turned out to be gender based. In other words, emotional work is gendered. We conclude that in order to understand how gender and gender equality are ‘done’ in the research interview, it is important to incorporate all three dimensions in the analysis: discourses, emotional work and the situated situation, as well as the links between them.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by the Riksbankens Jubileumsfond [grant number P11-1049:1].

Notes on contributors

Merete Hellum (Ph.D. 2002) is a senior lecture at Department of sociology and work science, Gothenburg University and has worked and published in different research topics, such as gender equality among Swedish sociologist, marriage between heterosexuals with different cultural background, young people using drugs and views of professionals working with children regarding cases of sexual abuse of children. Hellum has participated in international projects, such as the Family Change in Europe Today (EU project, 1997–1999). In 2005, she held a postdoctoral fellowship for female researchers financed by Selma Andersson’s foundation.

Livia Sz. Oláh (Ph.D. 2001) is an associate professor of Demography at the Department of Sociology, Stockholm University with expertise also in law and political science, comparative welfare state research and gender studies. Project coordinator of Families And Societies, large-scale collaborative project financed in the European Union Seventh Framework Programme (2013–2017), initiator and coordinator of the research network Gendering European Family Dynamics; a member of the Network of Excellence RECWOWE (2006–2011), editorial board member for several international journals, as well as of reference groups for Swedish government reports on family issues.

Acknowledgment

Financial support from Riksbankens Jubileumsfond [grant number P11-1049:1] for the data collection and analyses is gratefully acknowledged. We also thank Sanja Magdalenic for her work on preparing the data from Stockholm.