Competence and Warmth: How Gender Impacts Perceptions of Male and Female Speakers.

Competence and Warmth: How Gender Impacts Perceptions of Male and Female Speakers.

Jayne Summers

2017

Using the stereotype content model as a theoretical background, this study aimed to investigate the relationship between gender stereotypes and judgements of warmth and competence. Visual appearance has long been used to research these judgements while auditory cues have often been overlooked. This study therefore focused on judgements made about voice and subsequently did not influence participants with predetermined gender labels. 61 participants – aged 19 to 60 – listened to either 2 male or 2 female speakers talk about domestic violence and cancer research. Domestic violence is here defined as a women-centric topic, while cancer research is considered gender neutral. Participants completed person perception inventories of each speaker, rating them on 7-point Likert scales in terms of 10 competence and 10 warmth items. They also completed a sexism inventory to determine whether sexism predicted a more favourable attitude toward male speakers. A 2 between gender (male vs female) by 2 within topic (domestic violence vs cancer research) ANOVA was conducted, and female speakers were judged as more competent than males when speaking on domestic violence but not cancer research. They were considered warmer than men in both cases. This indicates that women are seen as competent when speaking on issues that directly affect them, suggesting that they should be taken more seriously when speaking out about their own rights. However, traditional warmth stereotypes regarding women were upheld. This, along with further implications, are discussed.

gender
stereotypes
competence, warmth
stereotype content model

Items. 10 competence items and 10 warmth items were selected to compile a 20-item list of characteristics for participants to judge speakers on. Of these items, 11 were taken from Rudman & Glick (1999) and the remaining 9 were considered in the original SCM. Items used in the competence and warmth scales were found to be reliable across speech topics, namely cancer
research (CR) and domestic violence (DV) (competenceCR α = .893, competenceDV α = .931). This indicates that the scales used were highly reliable. Similarly, for the warmth dimensions, Cronbach's Alpha was suitably high (WarmthCR α = .918, WarmthDV α = .944). The reliability for the sexism inventory was also acceptable, with an α value of .826. An example of several competence and warmth dimensions can be seen below, while a full list can be found in Appendix A. Competence: confident, ambitious, intelligent.. Warmth: trustworthy, likeable, supportive.
Speeches. Two speeches were recorded for the purpose of the experiment, one focused on domestic violence and the other on cancer research. The speeches were written to closely match each other in terms of wording and the information being presented. For instance, the opening and closing sentences of each speech were similarly structured, as seen below.
Table 1. Examples of speech text.
Domestic Violence Cancer Research
Opening sentence
Domestic Violence. A topic that is often glossed over as something that effects other people - not me; not you.
Cancer. A topic we don't often like to think about – something that effects other people, but not me: not you.
Closing sentence
By going to our website www.dvrefuges.co.uk you can find out more information about the great work women's refuges around the country do, and help them continue to change women's lives by donating to our cause.
By going to our website www.ukcancer.co.uk you can find out more information about the great work that we do, and by donating to our cause, help us continue to help people diagnosed with cancer live a normal life.
The details of the speeches differed, and the content was varied enough so as not to be obviously the same to participants, but the speeches were largely similar, as can be seen in Appendix B.
Four speakers were responsible for recording the two speeches, a male and female speaker for each topic. This allowed participants to hear both speeches either spoken by two male or two female speakers. All four speakers were from the same region and had northern accents, however, two speakers' accents differed slightly from the remaining two, which may have been particularly noticeable to northern participants. To account for this, one speaker with each accent was assigned to each topic condition and so any accent effects were counterbalanced and can be assumed to not have influenced judgements.
Speeches were recorded using an iPhone 6 microphone and edited using Audacity in order to eliminate background noise and static. Recordings were then given a plain video image of a black background with text reading either 'Recording One' or 'Recording Two' respectively. Due to the fact that recordings were counterbalanced across conditions, all four recordings were presented either as first or second in at least one condition, so in total 8 versions of the recordings were made and embedded into Qualtrics, where the body of the survey was hosted. Participants listened to recordings using Sony headphones during the experiment.
Procedure
Participants were assigned to one of four conditions. In each condition they were asked to listen to the first speech, either domestic violence or cancer research, spoken by either a male or female speaker. After listening to the speech, they proceeded to the next online page and completed the speaker evaluation, rating the speaker on the 20 warmth and competence
dimensions. This was indicated by how well they believed each item fit the speaker by choosing a point on a 7-point Likert scale (1 = completely disagree, 7 = completely agree). Following this they listened to the second speech spoken by a different speaker of the same gender. They then completed the same speaker evaluation for the second speaker. Finally, they completed the sexism inventory (The Ambivalent Sexism Inventory, Glick & Fiske, 1996) which measured the participants' explicit sexist attitudes on a 5-point Likert scale (1 = strongly agree, 5 = strongly disagree). A copy of the items in this inventory can be found in Appendix C. As this was a 2 (gender: female vs. male) x 2 (topic: domestic violence vs. cancer research) experimental design with repeated measures on the second factor, the difference between each condition was purely the order in which the speeches were presented (domestic violence first or second) and the gender of speaker that each participant heard (male or female) for the purpose of counterbalancing. So as not to influence participants to respond in a set way, the experiment was presented as regarding the evaluation of speakers and not as explicitly about gender.
Following the main section of the experiment, participants were asked a number of questions regarding how they experienced the recording, the first of which was answered on a 5-point scale (1 = strongly agree, 5 = strongly disagree). The question was: 'how likely are you to visit the website mentioned in this speech.' This was relevant in order to measure whether the competence of the speaker affected the likelihood of the participant to engage with the issue. Importantly, participants were also asked whether they considered each topic to be masculine or feminine, again measured on a 7-point scale (1 = feminine, 4 = neither feminine nor masculine, 7 = masculine). This was included in order to provide validity to the assumption that the domestic violence topic would indeed be judged as more women-centric, and the cancer research topic would be neutral. It is therefore of note that over 50% of participants considered domestic
violence to be a feminine topic, others considered it gender neutral, but very few considered it a masculine topic. The majority of participants judged cancer research as gender neutral, as was intended.
Finally, participants were asked whether or not they had any experience of the topic at hand, either personally or from a friend or family member, as this may have caused them to make more favourable judgements towards the topic they were more invested in. Participants also gave their gender, nationality and age. Gender and nationality were exploratory variables of particular interest due to the belief that other women may be more likely than men to evaluate women as competent. Nationality was of interest due to the fact that people from other cultures, particularly Eastern cultures, have different gender roles than we do in the UK, and so their responses during the experiment may have reflected this. Once the experiment was complete participants were fully debriefed and had the chance to enter a competition to win a prize in return for their participation.

Lancaster University

data/SPSS.sav

Sumners2017

John Towse

Open

English

Data

LA1 4YF

Tamara Rakic

MSc

Social Psychology

61 participants (14 male, 41 female, and 6 non-binary people) with an age range from 19 to 60 (M= 24.95, SD =9.63), were recruited through opportunity and snowball sampling

ANOVA