Download Inclusive Demographic Surveys: Best Practices and Coding for Data Analysis and more Study notes Logic in PDF only on Docsity!
Demographics
Inclusive and Functional Demographic Questions
Survey instruments often include questions that allow respondents to self-decribe various aspects of their personal and academic identities. Such questions and response items should be:
- Purposefully included based on an intentional rationale (e.g., monitor trends over time, disaggregate responses based on identity, distribute information) and;
- Designed in a way that acknowledges the scope, complexities, and contextualized nature of social identities (ACPA, 2013).
Demographic questions which exclude identity options where students are unable to select or are unable to select more than one category may result in unintended negative outcomes for respondents (Sanchez, 2010; Townsend, Markus, & Bergsieker, 2009). Rankin and Garvey (2015) discuss this as a balancing act of demographic survey questions which supports intersectional theory, inclusiveness, and quantitative methods. The following statements and survey questions are examples of how one might intentionally support this balance and honor students’ fluid and intersecting identities.
Checklist: Select Best Practices for Inclusive Survey Design
Whenever possible, allow questions 'select all that apply’ (noted in questions below) Whenever possible, do not require a response, or provide an option for respondents to mark ’prefer not to answer’ or ’don’t know’ (depending on the question)
Avoid ’othering’ language when providing a write-in option. For example, choose ’self-identify’ or ’write-in’ instead of ’other’
When using an online format (such as Qualtrics or Campus Labs Baseline), utilize ’display/skip logic’ to ask more in-depth inclusive questions only of those who select a particular response (noted in questions below)
Demographics
Demographic Questions
The following questions about your identity and background will be kept private and secure. Responses will be used to better understand and serve all students at the University of Arizona.
Note: The above text should be displayed within the survey before asking for demographic questions.
What is your academic class standing?
- First-year
- Sophomore
- Junior
- Senior
- Graduate or Professional
What is your enrollment status? Display/Skip logic: If “First-year, Sophomore, Junior, or Senior”
- Part-time (Fewer than 12 credits)
- Full-time (12 or more credits)
What is your enrollment status? Display/Skip logic: If “Graduate or Professional”
- Part-time (Fewer than 9 credits)
- Full-time (9 or more credits)
Are you a transfer student?
In which academic college is your major at UA? (Select all that apply if more than one major)
- College of Agriculture and Life Science
- College of Architecture, Planning and Landscape Architecture
- College of Education
- College of Engineering
- College of Fine Arts
- College of Humanities
- Eller College of Management
- James E. Rogers College of Law
- College of Letters, Arts and Science
- College of Medicine
- College of Nursing
- College of Optical Sciences
- College of Pharmacy
- Mel & Enid Zuckerman College of Public Health
Demographics
Do you consider yourself to be: (Select all that apply)
- Heterosexual or straight
- Gay or lesbian
- Bisexual
- Asexual
- Queer
- Questioning or unsure
- Not listed. I consider myself (specify) ________________
- Decline to state
Which best describes where you currently live?
- Residence hall
- Fraternity housing
- Sorority housing
- Off-campus housing (within 5 miles of campus)
- Off-campus housing (farther than 5 miles from campus)
- Living at home with family
- No stable residence
Do you have a child or children under age 18 for whom you are a primary caretaker?
Please check any of the following experiences you have had, or are currently experiencing. (Select all that apply)
- I have been in foster care (at any point in your life).
- I have been homeless or have had to stay with others that are not my family to have a safe place to live.
- I am a young adult that is age 24 or younger that does not have contact or support from parents or guardians.
Have you been diagnosed with any disability or impairment?
Do you identify as a veteran or active member of the U.S. Armed Forces, Reserves, or National Guard?
- No
- Yes, veteran
- Yes, active duty
Demographics
Have either of your parent(s) or guardians earned a bachelor’s degree or higher?
Which social class group do you identify with?
- Poor
- Working Class
- Middle Class
- Upper Class
Which of the following best describes the area you lived in before attending University of Arizona?
Are you employed?
- Yes, more than 40 hours each week
- Yes, 20 to 11 hours each week
- Yes, 10 hours or less each week
- No
Is your place of employment on-campus? Display/Skip logic: If “Yes”
Which of the following most accurately describes your background?
- My parents/legal guardians and I were born in the U.S.
- I was born in the U.S.; one parent/guardian was not
- I was born in the U.S.; both of my parents/guardians were not
- Foreign-born naturalized citizen
- Permanent legal resident
- Foreign born on student visa
- Deferred Action for Childhood Arrivals (DACA) recipient
- Refugee status
- Prefer not to answer
What is your preferred religious identification? Open-ended text box_____________
Demographics
Coding Demographic Questions
The process of coding survey questions is part of the “Collect & Analyze Data” step of the Assessment Cycle. Coding survey questions involves making thoughtful decisions about what numeric value is assigned with the question response items. Survey software (e.g. Campus Labs) automatically code responses in an ascending numerical order (e.g. 1, 2, 3…6). However, steps need to be taken to determine if variable responses need to be recoded into different or new categories to better support the analysis of data. Once question response items are coded with appropriate numeric values, software (i.e. Excel or Stata) allows for analysis of the data to produce assessment findings and insights. Demographic questions are commonly recoded to support statistical analysis based on these four reasons:
- Open-ended responses : “An identity not listed, self-identify” response option produces write-in responses that need to be coded
- Select all that apply questions : More than one response is selected in multiple choice questions indicated by “Select all that apply” survey prompt
- Less than five respondents : Demographic identity response items have less than five respondents
- Collapsing demographic responses : Data analysis and reporting context supports the need to collapse demographic responses
Open-ended responses
Responses require recoding when a respondent enters a response within the opened-ended text box (i.e. “An identity not listed, self-identify”). This option provides a respondent the opportunity to self-identify instead of being limited by inadequate preset identity options. Open-ended responses are coded using a three step process:
Write-in responses which are already listed as preset response items should be recoded to the same corresponding numeric value. For example, a respondent who writes in “man” when asked about their gender identity should be coded to the same numeric value for “Man” in the survey question.
Similar write-in demographic responses should be collapsed together creating a new category. For example, five or more students who write in a religious or spiritual identity which is not listed should be coded as a new response item.
Remaining responses (i.e. identity groups with less than five respondents) should be collapsed into “Another identity” category and coded with a numeric value.
Demographics
Select all that apply questions
Demographic questions which entail the option of selecting multiple response items may need to be recoded depending on the context of the question and need for data analysis. Example : Respondents who mark two or more race/ethnicities may be recoded as multiracial, which may support data analysis with having one demographic variable response item per respondent. This may not always be the case if your purpose for asking a student’s racial or ethnic identity is to separately understand and communicate the two or more unique identities selected.
Less than five respondents
To honor confidentiality, it is common practice in survey methodology to not report on demographic response identities if the number of respondents is less than five. This practice reduces the likelihood of respondents being identified and therefore breaking confidentiality and anonymity. Survey items with less than five respondents should be recoded and collapsed into other categories or into a new one. Example : “Transgender”, “non-binary”, and “Genderqueer or gender nonconforming” may be recoded into a new category labeled “Trans*/Another gender identity” to uphold confidentiality.
Collapsing demographic responses
Assessment of outcomes and corresponding communication of findings may require an aggregated category of student response items. Example : Academic class standing may be better suited for data analysis and reporting by collapsing “First-year”, “Sophomore”, “Junior”, and “Senior” into a new coded category labeled “Undergraduate Students”.
Contact Assessment, Research, & Grant Development for additional recommendations on coding guidelines for survey questions at AssessmentResearch@email.arizona.edu or (520) 626-4889.
