Social Work Research

Beginning in 1972, the General Social Survey (GSS) is a social research initiative conducted by the National Opinion Research Center (NORC). This program is aimed with the purpose of obtaining extensive quantitative data on a variety of social indicators from the general population of the US (Smith, Marsden, & Hout, 2016). By utilizing interviews with a national sample, standard questionnaire procedures are employed to measure diverse data including demographic information, as well as to identify trends in attitudes on social issues. As part of the National Data Program for the Social Sciences, the GSS serves to assist in fulfilling the goals of, “conducting basic scientific research on the structure and development of American society, and distributing up-to-date, important, high-quality data to social scientists, students, policy makers, and others” (NORC, 2016). The GSS was administered annually from the years 1972 to 1994, and biannually from 1996 to 2014 (Smith, Marsden, & Hout, 2016). For most of the history of the GSS, a cross-sectional survey design was utilized, however, from 2008 onward, rotating panels of the same respondents were incorporated for assessing changes in social attitudes and opinions over the course of time (Smith, Marsden, & Hout, 2016). 
This paper examines data taken from the latest GSS conducted in 2014. This survey included a total of 2538 participants. Ten variables, which represent specific characteristics of the study participants, were analyzed. These 10 variables are described in the sections delineated below.

Household Size and Composition
In reporting the population of the household, participants were asked to identify the total number of persons living in their home, excluding individuals who are living away at college, stationed away in the Armed Forces, or living in an institution. According to survey respondents, the mean number of persons living in the household was two, and 92% (n=2338) of participants reported a household size of 4 or less individuals. It was interesting to discover that 67% (n=1692), or about two-thirds of participants, had a home population of between one and two people, and 39% (n=997) had a total household size of two people. In addition, 15% (n=385) of participants had a household size of 3, 10% (n=261) had a household of four, and only one percent (n=21) had a household size of seven or more.
The mean household population was only slightly surprising to me. Although I expected the average size to be closer to three or four people, I recognize that likely due to cultural factors and poor economic circumstances over the last decade or so, household composition has changed. Due to my knowledge of declining marriage rates and the tendency for many individuals to marry at a later age and pursue higher levels of education to gain competitiveness in the challenging labor force, I am under the assumption that birth rates have been decreasing. Additionally, I am aware that the percentage of people getting divorced, as well as the existence of single-parent households, has increased. Increasing rates of divorce may explain the division of households into smaller compositions. I believe that these declines in both birth and marriage rates, as well as an increase in divorce rates, would account for the overall low household composition reported in the 2014 GSS.
Race/Ethnicity

In terms of race and ethnicity, participants were asked to identify their race out of a total of 16 possible categories. The race that had the highest representation in this sample was White, which constituted 75% (n=1878), or three-quarters of participants’ racial identity. The second most represented race was Black or African American, which comprised 15% (n=385) of participants, followed by American Indian or Alaska Native, which made up two percent (n=42) of participants’ race. The remaining 13 racial categories were identified as representing only eight percent (n=211) of participants.

The overrepresentation of White participants in this sample may be due to either the sample selection process or rates of participation in the survey. People with higher levels of privilege or socioeconomic status may have responded to the survey at higher rates than members of disadvantaged communities or minority racial and ethnic groups. In addition, the way in which the racial category of White is defined is broader than many of the other racial categories such as Chinese, Japanese, Korean, Vietnamese, and others, which comprise single countries of origin. According to the the Office of Management and Budget (1997), the category of White is defined as “a person having origins in any of the original peoples or Europe, the Middle East, or North Africa” (p. 58789). This clumping together of distinct regions of origin, and the absence of alternative categories for those of Middle Eastern and North African descent, may partially explain the higher level of representation for the White racial category.
Sex
With regards to participants’ sex, females had the highest representation, at 55% (n=1397) of the sample, while males made up 45% (n=1141) of participants. This 10% discrepancy in sex representation may be caused by differences in behaviors between females and males. Although it is an assumption or stereotype of the sexes, I tend to believe that women are more willing or inclined to have a desire to participate in a research study, as they may be more concerned with social matters. From my own experience, women seem to be more cooperative when asked to take the time to do something which requires information gathering, and appear to be more cooperative or accommodating in general. Men, on the other hand, seem to have an easier time turning down tasks which they do not wish to engage in. These behavioral differences may be rooted in the social conditioning of gender roles, inherent biological tendencies of the sexes, or in the differences in societal power and privilege between men and women that make men able to exert greater degrees of choice in non-participation or defiance.
Labor Force Status
Categories for labor force status included “working full-time, working part-time, temporarily not working, unemployed or laid off, retired, in school, keeping house, and other.” Of the total number of participants, 49% (n=1230), or almost half, reported working full time, while 11% (n=273) reported working part time. A total of 6% (n=144) of participants reported either temporarily not working or being unemployed or laid off. In addition, 18% (n=460) of participants were retired, 4% (n=90) were in school, 10% (n=263) were keeping house, and 3% (n=76) reported “other.” In anticipating how these labor force data from 2014 may compare to that collected in 2016, newer data would be dependent upon changing conditions of the US economy over the two-year span. Based on my own limited knowledge, personal experience, and observation of the US political and economic circumstances since 2014, it seems as if the job market has become increasingly restricted and unemployment rates did not appear to improve over this timeframe, as the US debt continued to climb and no tangible development of the economy seemed to be achieved.
Number of Children

My reaction to the mean number of children of the survey participants, similarly to the data on household composition, was not significant. According to responses, participants had an average number of 2 children. Based on my general bias, this is not surprising, as it seems as if birth rates have been declining as the US economy has continued to demonstrate poor performance. Additionally, many individuals are opting out to marriage or getting married later in life, which has a negative impact on birth rates. It seems as if throughout the history of the US prior to the economic collapse, family size was considerably larger. While I was not surprised by the average number of children, I did not expect there to be so many participants to report having no children. A total of 28% (n=704) people reported having zero children, and it was surprising to discover that over one-quarter of individuals were childless. I expected this percentage to be substantially lower, as I was not aware of how many people do not have any children.
Education
The education level of respondents was determined based on the highest year of school completed. Participants had a range of 20 years of school completed. The average number of years of school completed was 14, while 42% of respondents completed up to 12 years of school, or finished high school. In addition, 8% (n=193) completed at least one year of college, 18% (n=452) attended 4 years of college, and 15% (n=369) completed 5 to 8 years of college.
In order to more accurately capture the education level of respondents, education could be categorized in alternative methods. Instead of asking participants about the highest year of school completed, educational categories could have been divided into three parts: A. less than high school completed, B. high school completed, and C. attended college. Alternatively, higher education could be designated into separate categories based on degree acquired, such as: A. associate degree, B. bachelors degree, C. masters degree, and D. doctorate degree.
Marital Status
In terms of marital status, 46% (n=1158) participants are married, while 8% (n=209) are widowed, 16% (n=411) are divorced, 3% (n=81) are separated, and 27% (n=675), or over one quarter, were never married. This data shows that over half of participants were not married at the time the survey was taken. As those in the married category constitute the highest representation, the age distribution of participants may have influenced the marital status reported.
Religious Preference

Regarding religious preference, 13 categories were included to represent participants’ religious identity. Twenty one percent (n=522) of participants reported their religious preference as “none,” while the remaining participants selected a religious preference. Protestant was the highest represented religion, making up 45% (n=1125) of participants’ religious preference, while the next highest represented religion was Catholic, composing 24% (n=606) of participants’ religion. Differences in religious representation may be due to the geographic location, as well as the race and ethnicity of the sample. The high representation of White participants may account for the elevated percentage of Protestant and Catholic preferences. 
The categories constituting religious preference are neither exhaustive nor mutually exclusive. A wide range of additional religions exist; however, these religions were excluded from the category list. Alternatively, a category of “other” was included. Furthermore, participants were unable to select more than one religious category. Particularly as interfaith marriage has been growing, it is likely possible that a substantial number of participants identify as possessing multiple religious identities, however, this is not reflected in the data.

Age
As both the mean and median age of participants is 49 years, and the most frequently occurring ages are those in the 50’s, middle-aged participants made up the highest representation of the sample. Ages reported by respondents were in a range of 71 years, showing a wide age distribution.
The advantages of collecting data measured at the interval and/or ratio level compared with collecting data measured at the nominal/ordinal level are that data are more specific and quantifiably comparable. In terms of age, collecting respondent’s exact age at the ratio level allows the researcher to examine this variable at a higher level, analyzing differences and the distances between values in a more precise manner. By collapsing the data into categories, interpretation and analysis of findings would be easier, as there would be a vast reduction in values. If age was collected at a lower level of measurement, in condensed ranges, groups of participants may be grouped together, and these groups may aide in comparative analysis.
Political Party Affiliation
Participants’ political party affiliation was distributed in a considerably evenly dispersed manner, showing a reasonable degree of diversity. The most represented political party was independent, characterizing 20% (n=502) of participants’ affiliation. The participants’ political party affiliation leaned more heavily towards the democratic side of the spectrum, as a total of 46%, or almost half of all participants (n=1162) identified as either “strong democrat, not strong democrat, or independent near democrat,” while 31% (n=786) of participants chose from the categories of “strong republican, not strong republican, or independent near republican.” As a result of these findings, and the higher levels of democratic participants, I would make the assumption that perhaps individuals identifying towards a democratic affiliation are more interested in social science research, and in turn, more likely willing to participate in the study.
If these findings were to be compared to the MSW student population at the SSSW, I believe that there would be a sizeable discrepancy in the percentage of republican-leaning respondents. As MSW students tend to be characteristically liberal and progressive in terms of their political ideology, rejecting many conservative political positions and policies as oppressive and antithetical to social work values, I would expect to see a much higher percentage of students with democratic and independent affiliations, and a significantly low number identifying on the republican side of the spectrum.
Conclusion/Reflection
The analysis of these data from the GSS has helped me to gain an understanding of the demographics and social attitudes of the participants involved in the survey, revealing general information about the backgrounds, beliefs, and socioeconomic factors of respondents. The picture the data presents will be helpful in the final research report, as knowledge on characteristics of the sample will help to guide the interpretation of the research question and problem area. Through learning how to analyze these variables, I will be better able to understand my agency’s data on client population, as well as make sense of statistical information included in community reports and studies.