Start Research Project. Fix. Then Actually Start.
If you were living in a new country, would you know how to enroll your child in school, get access to health insurance, or find affordable legal assistance? And if you didn’t, how would you deal?
As Maggie, Kristen, and I are starting to interview immigrant women living in the US and the organizations that provide support to them, we are trying to understand how they deal – particularly, how they seek social support when they face stress.
This post gives a bit of an orientation to our project and helps us document our research process.
We’ve developed two semi-structured questionnaires to guide our interviews: one for immigrants and one for staff at service providers (organizations that provide immigrants with legal aid, job training, access to resources for navigating life in the US, and otherwise support their entry and integration). We are seeking to learn about women immigrants who have been in the US between 1-7 years. All interviews are conducted in English by one of the team members. For this reason, we are striving for a high degree of consistency in our interview process because we will each be conducting one-on-one interviews separately from each other.
As we have begun interviewing, we’ve realized a couple of things:
- Balancing specificity and ambiguity is difficult, but necessary. We need questions that are agile enough to be adapted to each respondent, but answers that are not too divergent to compare once we do analysis.
- The first version of our interview questions needed to be sharpened to better enable us to learn more about the particular technologies respondents use for specific types of relationships and with respect to different modes of social support. Without this refinement, it may have been difficult to complete the analysis as we had intended.
We have sought mentorship and advice from several mentors, and that has helped us arrive at our current balance of trade-offs. We have found support in Gina Neff and Nancy Rivenburgh, Maggie and Kristen’s PhD advisors, as well as in Betsy Cooper, the new director of the Center for Long-term Cybersecurity.
When we met with Betsy, she took about 30 seconds to assess the work we had done so far and begin to offer suggestions for dramatic improvements. She challenged us to reexamine our instruments, asking, “Do the answers you will get by asking these questions actually get at your big research questions?” The other advisors have challenged us similarly.
This helpful scrutiny has pushed us to complete a full revision of our migrant interview questions. Betsy also recommended we further narrow our pool of immigrants to limit to one to three geographical regions of origin, so that we might be able to either compare between groups, or make qualified claims about at least one subset of interviewees. As a result, we would like to interview at least 5 people within particular subsets based on geography or other demographics. We are making an effort to interview in clusters that are geographical. For example, we are angling to interview individuals originally from East Africa, Central America, and/or Mexico, given each group’s large presence at both sites.
However, we anticipate that there may be greater similarities among immigrants who are similarly positioned in terms of socioeconomic status and potentially reason for immigrating than those that are from similar geographies. We’ll be considering as many points of similarity between interviewees as possibility.
We’re finding that these three scholars, Nancy, Gina, and Betsy, fall on different points on the spectrum of preferring structured interviews to eschewing rigidity and allowing the interview to take the direction that the researcher finds most fruitful during conversation. Our compromise has been to improve our interview process by sharpening the questions (ambiguity removed or reduced as much as possible) we will step through with interviewees, leaning on clarifying notes and subquestions we are asking when the themes we’re looking for don’t organically emerge from the answer to the broader question.
Honing in on what’s really important is both imperative and iterative. In addition to our questionnaire, this includes definitions, objectives, and projected outcomes. This may sound like a banality, but doing so has been quite challenging. For example, what specifically do we mean by ‘migrant’, ‘social support’, ‘problem’, or ‘stress’? Reaching group clarity about this is essential. We also must remain as flexible as possible, because in the spirit of our work, we recognize we have a lot to learn, so artificially rigid definitions may not position us as well to learn from those we are interviewing (or even to find the right people to interview).
As we seek clarity in our definitions, we’ve looked to existing models in the great scientific tradition of standing on the shoulders of others. Social support defies easy definition, but one helpful distinction in an article by Cutrona & Suhr splits social support into two types, nurturant (attempts to care for person rather than the problem) and action-facilitating (attempts to mitigate the problem causing the stress). We’ve found this distinction helpful as both a guide and a foil for revising our questionnaire instrument to clarify what we’re looking for. We may find another classification from the literature that better matches what we are finding in our interviews, so we’ll stay open to this possibility.
Stay tuned. We’re excited to share what we learn as we get deeper into the interviews and begin analysis in a couple of weeks.
Cross-posted over at the Migrant Women & Technology Project blog.