Digital Data 2019, Week 11

Week 11: But Her Emails

Much of our focus in this class has been on collecting data publicly accessible via the web, and while social media platforms are an important data source, so are other less trendy digital platforms. For this week’s post, I reviewed two articles that made use of email as a data source for studying gender. In the first, Ma and Seate (2017) used email to study gendered communication patterns, asking whether people would use more tentative language when discussing a topic on which their gender group was not considered to have expertise. To do this, the researchers first recruited a group of communication students to identify which topics from a prepared set were considered masculine, feminine, and gender neutral. A second group of participants were then presented with a woman’s autobiography (coded by the initial pilot group as either ‘typical’ or ‘atypical’ of a woman) and asked to respond to her via email on a discussion question coded as masculine, feminine, or gender neutral. Interestingly, in this case email was not important as a platform for study, except that it served as a mechanism for distancing subjects from each other, and obscuring the researchers’ methods. Christina, the woman to whom all participants were asked to respond, was a fictitious person whose ‘typical’ and ‘atypical’ biographies were created for the sole purpose of the experiment.

The researchers found their hypotheses supported; men and women both use more tentative language when discussing subjects in which their gender group is not generally regarded as having expertise. Given that the pilot study participants, the main study participants, and the researchers later coding the results are all part of the same community, it makes sense that those groups would share common ideas of who can be an expert in what topics, and what is considered appropriate behavior for their gender group around gendered discussion topics. That said, the bounds around who counts as a “typical” woman were very narrowly constructed, and I’d be interested to see whether these results hold up in a less confined and less problematic context. For example, compare these portions of the biographies presented to participants:

Here’s how I want to spend my perfect day. […] We will go into [location of university] and go shopping for clothes and shoes. Not that I have any more room for clothes in my closet — I need more space. (“Typical” woman version)

Here’s how I want to spend my perfect day. […] We will go into [location of university] and go to a basketball game. I want to get a new jersey. Not that I have any more room for jerseys in my closet — I need more space.” (“Atypical” woman version)

Apparently, both typical and atypical women are obsessed with clothes, but what makes the second woman atypical is that she’s obsessed with basketball jerseys specifically instead of clothes and shoes more generally. Read generously, this could unintentionally play on stereotypes of both race and sexual orientation; read less generously, these biographies are dog whistling that the ‘atypical’ version of Christina is Black, a lesbian, or both. Given the very elementary understanding of gender embedded in this study, it’s hard to take the results seriously.

The second article I reviewed does in fact use email as a meaningful platform; Sappleton and Lourenço (2016) wondered whether the subject line of emails could be used more strategically in to increase the rate of reaction from those solicited. Specifically, this study investigated whether emailing a survey solicitation with a blank subject line would play upon the receiver’s curiosity and lead to their taking action at great rates than a provocative subject line, regardless of whether that action would be completing the survey, or actively refusing to do so. Considering notoriously low email survey response rates, the researchers first identified a population of likely respondents — “business owners examining gender segregation in business sectors” (p. 615) — and targeted sectors knows for being male dominated (constructions and audio recording), female dominated (childcare), and integrated (publishing). Half of the owners in these sectors were sent email solicitations with a provocative subject line, while the other half received email solicitations with a blank subject line.

This study found that a blank email subject did in fact increase the rate of reaction of those solicited; while the rate of survey completion was slightly higher, this was not a significant difference (39.7% with the blank subject line participated, versus 38.7% with the provocative subject line). There was a significant increase in the number of business owners who actively refused to take the survey: 8.6% of those with a blank subject line refused, versus 0.9% with a provocative subject line.

(Sappleton and Lourenço 2016:616)

A second study, targeted towards academics but using similar methodology otherwise, found similar results. Faculty members receiving a solicitation with a blank subject line gave a slightly higher response rate, but it was statistically insignificant; the active refusal rate for those receiving the blank subject line was twice that of those receiving an informative subject line.

(Sappleton and Lourenço 2016:618)

The main takeaways from this article for me are the impressive and thoughtful decisions that demonstrate the researchers had a great deal of knowledge about the behavior of their study populations in advance, and were able to use that knowledge to more carefully target their audience. They admit in the conclusion that much information about the specific demography of the participants is missing; they know generally what business sector or what academic discipline their study populations work within, and made inferences about gender based on names (which is problematic), but otherwise cannot estimate information about race, socioeconomic or marital status, age, or other variables which may be related to email survey response rates. However, the knowledge that blank subject lines do seem to increase the rate at which people open the solicitation emails and to react in some way is a lesson that future research can build upon.

Ma, Rong and Anita Atwell Seate. 2017. “Reexamining the Use of Tentative Language in Emails: The Effects of Gender Salience and Gender Typicality.” Journal of Language and Social Psychology; Thousand Oaks 36(6):694–714.

Sappleton, Natalie and Fernando Lourenço. 2016. “Email Subject Lines and Response Rates to Invitations to Participate in a Web Survey and a Face-to-Face Interview: The Sound of Silence.” International Journal of Social Research Methodology; Abingdon 19(5):611–22.