Applied Research Methods in Behavioral Science

Assignment Analysis Methods

Grade: 10% of the total grade

Length: 1-2 pages (a little longer is ok) single spaced, 11 pt font, 1-inch margins
Submission: via Canvas by 11:55pm the day the assignment is due, hard copy in class

Assignment: Submit and discuss draft analysis plan. Please refer to the sample analysis plan attached. It
is very detailed and probably has more details than you would want to write. As you can see there are
several sections in the analysis plan:
 Sample: Inclusion criteria and the final number of observations in the analytic sample, purpose
of the study
 Procedures: Discuss how the data was collected
 Measures:
o The dependent variables, possible responses, and how the variables were collapsed
o The independent variables, possible responses, and how the variables were collapse
 Analytic Plan:
o Univariate analysis you will conduct
o Bivariate analysis you will conduct
o Multivarite analysis you will conduct (if any)

The example included in this assignment has more information than you will need to include since this
project relies on secondary data analysis.

Remember: This assignment should be completed alone, without consultation with other students. No
late papers will be accepted.

Other Info: Do not plagiarize. All assignments will be checked with Turnitin. As a general rule of thumb, if
you are unsure if you need to cite, you should err on the side of caution and cite. Please use APA format
for citations.

1
Applied Research Methods in Behavioral Science

Example:
Methods

Study Design

This study was conducted using the 2018 BRFSS Core Module data. The methods used by BRFSS
have been explained in detail elsewhere (Prevention., 2018a). Briefly, these data were collected by the
individual states in collaboration with the CDC. The purpose of this survey is to collect data from
residents of the United States who are over the age of 18 on health-related risk factors and preventative
behaviors. Surveys were administered over the phone (both landline and cellular).

Sample

The sample for this study is comprised of all participants who met the inclusion criteria.
Specifically, the participants had to be between the ages of 18 and 44. For statistical purposes the
sample was further restricted to only participants who had provided an answer to all of the following
items: gender identity, sexual orientation, race/ethnicity, income, educational level, employment status,
living in a metropolitan area, insurance status, if they have a PCP, and if they have received a HPV
vaccination. The final analytic sample consists of 7330 participants.

Measures

Gender identity and sexual orientation. Data on gender identity were collected for all
participants by asking if they considered themselves to be transgender. Respondents could select one of
the following responses: yes (transgender male-to-female), yes (transgender female-to-male), yes
(transgender, gender non-conforming), and no. A new dichotomous variable was created from the
aforementioned transgender item that collapsed the three yes responses into a single transgender
category, while the no responses were categorized as cisgender. In addition to gender identity, all
participants were asked about their sex via the following prompt: “What is your sex or what is your sex
at birth?” Participants were given two options: male and female. Based on the gender identity and sex
variables, new variables were created that identified cisgender males and cisgender females.

Data on sexual orientation were collected by sex. Separately, male and female participants were
asked the following question related to their sexual orientation: “Which of the following best represents
how you think of yourself?” Response choices were: gay/lesbian, straight (i.e., not gay), and bisexual.
Using these two variables, new variables were created that identified heterosexual, homosexual (lesbian
or gay), and bisexual participants, regardless of sex or gender identity.

Predictor variables. Participants were asked to report their age and BRFSS reported it in five-
year intervals. For this analysis, participants between the ages of 18 – 44 were included in this analysis.
These five age groups were then dichotomized into those between age 18-34 and those between age
35-44, as these two age groups most closely reflect the previous HPV vaccination recommendations by
the CDC (Control & Prevention, 2010a, 2010b, 2011; Meites, 2016; Meites et al., 2019; Petrosky et al.,
2015; Prevention., 2018b; Release et al., 2007).

Race and ethnicity were collected and collapsed by BRFSS staff into five categories: white only
(non-Hispanic), Black only (non-Hispanic), other race (non-Hispanic), multi-racial (non-Hispanic), and

2
Applied Research Methods in Behavioral Science

Hispanic. These categories were collapsed into two groups: white non-Hispanic participants and
participants of color.

Participants reported their income using the following responses: less than $10,000, less than
$15,000, less than $20,000, less than $25,000, less than $35,000, less than $50,000, less than $75,000,
$75,000 or more. Next, income information was dichotomized into two groups: those making less than
$15,000 per year and those making over $15,000. This dichotomization was based on the 2018 Federal
Poverty Level ($12,140) for the United States (Health & Services, 2018).

Education level was collected using the following response choices: did not graduate high
school, graduated high school, attended college/technical school, graduated from college or technical
school. These response choices were dichotomized in a new variable and reported as either completed
some or all of high school or completed some or all of college/technical school.

Employment status was assessed by asking participants if they were currently employed for
wages, self-employed, out of work for 1 year or more, out of work for less than 1 year, a homemaker, a
student, retired or unable to work. For the purposes of this analysis, a new dichotomous variable was
created that collapsed those who were employed or self-employed and those who were out of work,
homemaker, students, retired or unable to work.

Metropolitan status was reported dichotomously in the BRFSS data set with participants
categorized into either residing in a metropolitan county or residing in a non-metropolitan county.
Similarly, insurance status was collected dichotomously via the following yes/no question: “Do you have
any kind of health care coverage, including health insurance, prepaid plans such as HMOs, or
government plans such as Medicare, or Indian Health Service?” Finally, participants were asked if they
have one person that they think of as their personal doctor or health care provider. Responses were
collected using the following options: yes only one, more than one, and no. Those indicating that they
had one or more providers they thought of as their personal doctor were collapsed into a yes response
in order to dichotomize the PCP variable.

Outcome Variable. Receiving the HPV vaccine was assessed by asking participants the following
question “A vaccine to prevent the human papilloma virus or H.P.V. infection is available and is called
the cervical cancer or genital warts vaccine, H.P.V. shot. Have you ever had the H.P.V. vaccination?”
Participants were given three response options: yes, no, doctor refused when asked. For the purposes
of this analysis, the options for no and doctor refused when asked were collapsed into a single category
to create a dichotomous variable.

Analytic Plan

Descriptive statistics were computed for participant sociodemographic characteristics (gender

identity, sexual orientation, age, race and ethnicity, income, educational level, employment status, living
in a metropolitan area, insurance status, and if they have a primary care provider) and the outcome
variable (if they have received an HPV vaccination). Prior to the bivariate analysis, all variables were
dichotomized as previously described. For the remaining analysis, the sample was stratified based on
gender identity (cis men, cis women, and transgender individuals) and sexual orientation (heterosexuals,
gay men/lesbian women, and bisexual individuals). The number of transgender participants was too
small (n=20) to be included as a subgroup in the statistical analysis and are not reported. Bivariate
analysis included chi-square tests for categorical variables to assess associations at p>0.05. Pearson chi-

3
Applied Research Methods in Behavioral Science

square tests were used to test associations between dependent and independent variables.
Multivariable analysis used multivariable logistic regression to test the relationship between the
sociodemographic characteristics and receiving the HPV vaccination. All analyses were conducted with
SPSS version 23 (IBM Corporation, Armonk, New York).