Study of Instructional Improvement

Family Structure & SES

The Parent Survey (available here) is a cross-sectional, rather than longitudinal, designed survey consisting of interviews with parents (or guardians) of children who were active participants of the study. Most of the interviews were conducted by telephone, but where necessary, field staff arranged face-to face interview sessions. The bulk of parent interviews were conducted in the spring of the 2001-2002 academic school year; however, additional interviews were conducted during the third and fourth years of data collection to capture information for new students moving into a sample school, and to obtain data from parents not previously reached. The survey includes questions about each child’s home environment, home activities, the child’s experiences with school, reports of services received, basic demographic information and questions about the family’s access to basic needs. The data are available for download here.

The Parent Survey data were central in the development of demographic control measures used in our analytic work and these measures frequently appear in published articles by SII researchers. For this reason, we provide readers the opportunity to examine our basic methods and procedures for developing a composite socioeconomic status measure, assignment of parent occupational codes, and determining family structure from the household roster.

To construct a composite socioeconomic status (SES) measure, SII researchers replicated the procedures commonly used in the development of education databases sponsored by the National Center for Education Statistics (NCES). It is important for readers to note, however, that the SES measure developed by SII is not norm-referenced to a national school population. Instead, a standardized coefficient of SES represents a student’s status compared only to other students in the SII population. Readers interested in generalized inference may consider propensity stratification score methods such as those described here.

SII researchers developed the SES measure using the exact items reported for the SES measure available in the National Educational Longitudinal Study ’88 (NELS: 88) and the Schools and Staffing Survey (SASS). The five-item composite measure includes the highest education levels reported for the (1) mother and (2) father, (3) reported total family income level, and the occupational prestige scores of the (4) mother and (5) father. Occupational prestige scores or Total Socioeconomic Index (TSEI) were adopted from the work of Robert Hauser. Succinctly explained, SII researchers standardized the mean average of the five items to produce the SES measure. Readers familiar with SPSS syntax can view the steps of the SES measure’s development here, but note that most preliminary steps shown are not necessary as SII provides a refined parent file.

As in all survey data collection, there are limitations and sources of potential measurement error. Parent information was successfully collected for about 75% of the SII sample, and it is known that the missing information is slightly disproportional in the direction of lower income families. However, we remind the reader that face-to-face interviews were arranged to mediate the effects of this common occurrence in survey data collection. Users of SII data will also notice a high rate of missing information for father’s educational background and occupation. The interview protocol called for information to be gathered only for parental figures physically residing in a household where a child participant lived. At the conception of the study, SII researchers did not anticipate that the rate of single parent female households would be as striking (42%). As a consequence, the amount of missing information for male parental figures is high and this limits the number of items to be averaged for the SES calculation.

Although the study was conducted during academic years occurring between 2000 through 2004, the census tracts used in sampling were based on 1990 Census information. Similarly, occupation codes were drawn from occupational titles associated with the 1990 Census. Therefore, it was also necessary to match these occupational codes to 1990 Total Socioeconomic Index (TSEI) scores, also referred to as occupational prestige scores. Readers interested in scrolling through the range of occupational codes may do so here, but we also recommend referring to the Parent File codebook to observe the frequency of general job categories held by parents of children active in the SII study (mothers – here; fathers - here).

The household roster section (here) of the Parent Survey is critical in developing several family background variables. First, it helps determine the type of family structure (e.g., mother and father present, single parent home), especially if the marital status information is missing. It also helps determine the relationships and age ranges of individuals reported occupying a residence. Additionally, SII researchers used this information to determine a mother’s age at the time of first born child, and sorting informant identity to assign the education levels and occupations of a mother and/or father (or other parent figure). For interested readers, we provide some of the coding syntax used for this work here.