The Home Observation Measurement of the Environment-Short Form (HOME-SF) is the primary measure of the quality of a child's home environment included in the NLSY79 child survey. It has been extensively used as both an input in helping to explain other child characteristics or behaviors and as an outcome in its own right for researchers whose objective is to explain associations between the quality of a child's home environment and earlier familial and maternal traits and behaviors.

The HOME-SF is a modification of the HOME inventory (Caldwell and Bradley 1984, 1992), a unique observational measure of the quality of the cognitive stimulation and emotional support provided by a child's family. The HOME-SF is about half as long as the original HOME Inventory, an adaptation necessitated by survey time and cost constraints. More than half of the HOME-SF's items are multi-response maternal reports that were reworded, with the assistance of the instrument's designers, from the original HOME Inventory's dichotomous observer ratings.

In addition to the overall HOME-SF score, the Child file includes two subscores: a cognitive stimulation and emotional support score. These are also described below.

Description of the HOME

A complete listing of the HOME-SF items appears in Appendix A in the NLSY79 Child/YA Codebook Supplement. From 1986-1998, all mother-report HOME items were located in Section 1 of the Mother Supplement. In 2000 the HOME items designed for children under age four were moved to the Child Supplement, but returned to the Mother Supplement the following survey year. All the mother report HOME items were consolidated in the Mother Supplement starting in 2002, the year the Mother's questionnaire was converted to CAPI. The HOME items based on interviewer observation appear in the Child Supplement for all survey years.

The HOME-SF is divided into four parts based on child age: (1) children under age three; (2) children ages 3-5; (3) children ages 6-9; and (4) starting with the 1988 survey round, children ages 10 and over. As outlined more fully below, the set of questions included in the HOME sections of the survey was expanded at several survey points. Note: None of the added items affect the overall HOME score or subscores in any way.

Bettye Caldwell designed the Infant version of the original HOME Inventory and, with Robert Bradley, developed the Preschool and School Age versions. Bradley and Caldwell reviewed and approved the final draft of the Infant, Preschool, and Elementary HOME-SF versions used in the 1986 Mother and Child Supplements of the NLSY79, and Bradley was involved in a 1988 review of what has now become known as the HOME-SF to distinguish it from the original.

Changes in the HOME

The following changes have been made to the HOME sections since they were included in the first child survey in 1986:

• In 1988 a version of the mother-report HOME was added for children 10 and older (in 1986 one set of items was used for all children ages 6 and older). 
• Beginning with the 1992 Mother Supplement, code categories were added to the questions on the relationship of the child's father/father-figure to the mother, and a category was added to the parent efficacy question for children ages xx and older. 
• Beginning in 1994, one additional sequence was added to the HOME assessment to measure child-parent closeness: Mothers are asked how close each of her children feels to her, to the child's biological father, or to his or her stepparents (e.g., see Q16a for children under the age of three).
• In 1996, check questions asking if the child ever sees his or her father were dropped.
• In 2000 only, the HOME questions addressed to the mother for children under the age of 4 years were moved from the paper Mother Supplement to the Child Supplement. In 2002 all the mother report HOME items were returned to the Mother Supplement. While the question format was slightly altered from the Mother Supplement paper questionnaire for CAPI administration starting in 2002, the content of these items remains the same.

Scoring the HOME-SF and its subscores

The total raw score for the HOME-SF is a simple summation of the recoded individual item scores and it varies by age group, as the number of individual items varies according to the age of the child. The procedure used to recode non-dichotomous responses into a 1,0 form is documented in Appendix A. The data transformation process used in all survey years was essentially the same.

The total HOME-SF score and the two subscores have one implied decimal place. For example, a score of 30 in the public data file is really 3, and so on. In addition, total scores were imputed for children where one or more of the component items had inadvertently been left unanswered. The imputation procedure assigns an average value, derived from all those items that had been completed, to each of the unanswered items. Proration flag variables specify the number of items that require imputation for the different age groups; a score of zero on this proration flag variable means that all individual component items were answered. Since the maximum number of missing items in the early paper questionnaire survey rounds can range up to a total of 26, users may wish to check the flags that indicate the total number of missing items (HOMEFA-Dyyyy) for the extent of imputation applied in the total score. For the two subscores specified below, a more stringent proration rule was followed: scores were derived only for cases where no more than one item was missing.

Recoding of HOME items

Prior to constructing an overall score and the two subscores for the HOME-SF, all of the individual items were translated into dichotomous zero-one variables and then summed. The precise recoding used in computing the HOME scores can be found in Appendix A. The dichotomous HOME items, available for on the public data file, are assigned respectively to the following areas of interest:

• ASSESSMENT 1996 (C17600.01-C17600.89)
• ASSESSMENT 1998 (C18996.00-C19084.00)
• ASSESSMENT 2000 (C25061.00-C25149.00)
• ASSESSMENT 2002 (C25344.00-C25432.00)
• ASSESSMENT 2004 (C28047.00-C28135.00)
• ASSESSMENT 2006 (C31140.00-C31220.00)
• ASSESSMENT 2008 (C36182.00-C36262.00) 
• ASSESSMENT 2010 (C39970.00-C40050.00) 
• ASSESSMENT 2012 (C55408.00-C55488.00) 
• ASSESSMENT 2014 (C58161.00-C58231.00)

Cognitive stimulation and emotional support subscales

In addition to the overall HOME-SF score, the Child file includes two subscores: a cognitive stimulation and emotional support score. The (questionnaire item) components of the total scale, as well as cognitive stimulation and emotional support subscales are specified in Appendix A. Because there are no appropriate national norms available for the overall HOME-SF score or its components, internally normed standard and percentile scores are provided for the overall HOME-SF scores and for the cognitive stimulation and emotional support subscores.

The HOME discipline items

Several of the HOME-SF items required review and coding of verbatim comments from the mother in order to fully utilize the responses originally coded as "other." The HOME-SF Part B (for children 3-5) contains items concerning the mother's response to the child hitting her (question HOME-B08B starting in 2002). Part C (for children 6-9) and Part D (for children 10-14) of the HOME-SF contain items concerning the mother's response to the child saying "I hate you" or swearing at her in a tantrum (question series HOME-C22B and HOME-D21B respectively, starting in 2002). These items are recoded "1" if the parent's response is moderate, defined as without harsh reprisal. Note: For question names prior to 2002, users should search on "spank" using Word in Title in NLS Investigator.

The discipline item for children ages 3-5 contributes to the HOME-SF scale scores only when certain alternatives ("send to room," "talk," "ignore," and "give a chore") are selected and the "other" alternative is without harsh reprisal--that is, if a mild reaction is the first response. The discipline item is scored zero if any of the following are selected: "hit," "spank," or the "other" alternative is harsh. Harshness is defined as either extensive or excessive deprivation (time-out longer than two hours; deprivation longer than two days) or physical punishment (firmly grasping the child, spanking then talking, or talking then spanking).

The HOME discipline item for children 6 and older was scored in a similar manner. Yelling back and withdrawal of love were assigned a value of "1" if there was no indication of a severe or physical response. The item is scored zero if "spanking" is selected or if the "other" alternative is excessive (longer than three hours of time-out; longer than three days of deprivation) or if physical means ("eat soap") are the first types of punishment selected. Examples of verbatim scores as harsh are "break him up," "spank and ground for two weeks," and "spank then explain why." If the length of time was not specified ("send to room"), then it was assumed to be a moderate amount of time, so scoring the item was mild. Other examples of verbatims scored non-harsh are "never happens," "depends on the situation," "stand in corner until apologizes." Another example of a mild response (conveying no harsh discipline) was "give him something to eat." A few other verbatims should be noted. Some mothers selected the "hit" category and commented, "Then say I'm sorry," while others who checked "hit" said, "But not like I'd hit an adult." The latter were scored as harsh and assigned a value of "0."

Three primary coders were typically used in this process. Inter-coder reliability on the coding of the HOME discipline items has been evaluated using Cohen's Kappa and also by computing percent disagreement. In 2002 and 2004 there was 100 percent agreement on the recoding decisions for the discipline items.

Age eligibility for the HOME

All children under the age of 15 living with their mothers are eligible for the HOME assessment (until 1994 all children, regardless of age, had a HOME supplement completed by their mother). Thus, children born by the 1986 survey date (and still below the age of 15 in 2002) may have eight rounds of HOME scores available. Children born between 1986 and 1988 may have seven HOME scores and children born since the 1988 survey can have up to six HOME scores--assuming of course that their mothers completed a HOME assessment for them at the relevant survey points. The HOME was not administered after the 2014 survey round.

Norms for the HOME

Internal norms for the NLSY79 HOME were developed using standard normal curve assumptions. Children were normed on a single year of age basis with each (weighted) single year age of age group being assigned a standard score mean of 100 and standard deviation of 15. Percentile scores were derived from the standard scores using an inverse normal routine. To the extent that the single year of age data deviate from normality, this procedure produces less than optimal results. An alternate percentile score can be generated using the empirical cumulative distribution function by age that is computed using the sampling weights. That frequency distribution could be used to crosswalk from raw score to percentile score.

Completion rates for the HOME

Table 6 in the Child Assessments—Introduction section indicates the number and types of children at different ages for which a HOME assessment was completed in the current survey round. Overall, 88 percent of children (under age 15) interviewed or assessed this round have a valid HOME score. This level of completion for the HOME assessment is slightly lower than the level that has been maintained over recent survey rounds. Completion rates are lower for the youngest children, because the interviewer observation items at the end of the Child Supplement are less likely to be completed for this age group. In recent survey rounds, since there are no assessments directly addressed to children under the age of four, interviewers are less likely to see the mother in direct contact with the child, and thus may be unable to answer items that require direct observation of mother-child contact. (This lack of data is a more significant issue with respect to the emotional support subscore than the overall HOME-SF score because the conditions permitting proration of subscores are more stringent, as noted above.)

Validity and reliability for the HOME

The HOME assessment is one of the most used child assessments. It is widely employed both as an input, predicting many other child outcomes, and as an outcome in its own right. For example, analyses based on the NLSY79 child data indicate that the HOME-SF is closely related to several different indices of family poverty, and, that the measure is sensitive to increments in family income, particularly when looking at children born into poverty. The overall HOME-SF scale and the cognitive stimulation and emotional support subscales are used frequently as outcomes of interest predicted by various family circumstances, such as mothers' and fathers' employment. These measures of the home environment are often used as predictors of children's cognitive and/or behavioral outcomes using PPVT, PIAT, and BPI scores. Studies that utilize existing or previously constructed measures from the HOME-SF typically cite one or more of the following sources for reliability and validity information: Baker and Mott, 1989; Menaghan and Parcel, 1989 & 1991; Luster and Dubow, 1990 & 1992; Parcel and Menaghan, 1990. A comprehensive review evaluates the strengths and limitations of the NLSY79 HOME-SF scales as well their utility in research across a variety of disciplines (Mott, 2004).

Researchers have used variations on the overall HOME-SF scale to isolate a facet of the home environment for theoretical reasons. For example, father presence has been isolated in some studies so that its unique effects can be examined (Mott, 1993; Baydar, 1995). Although not as frequently utilized as the overall scale and the two main subscales, individual items and researcher-constructed subscales representing a variety of specific concepts are also studied as predictors and as outcomes. For example, some studies use measures of discipline/punitiveness constructed from one or more HOME-SF items to predict later child behavioral outcomes (Dornfeld and Kruttschnitt, 1992; McLeod and Shanahan, 1993; McLeod et al., 1994; Rodgers et al., 1994; Parcel and Geschwender, 1995; Straus et al., 1997; McLoyd and Smith, 2002).

Some researchers with concerns about the reliability of some of the subscales have opted to rely on the overall HOME-SF (Ketterlinus et al., 1992). Some analysts, who have conducted confirmatory factor analysis on factor structures for the HOME-SF for younger children, recommend that the HOME-SF should not be thought of as measuring a single concept (Ferron et al., 1994). Researchers with concerns about the validity of the HOME for Hispanic children should be aware that, during the survey years in which there were significant numbers of Spanish-speaking mothers and children, the instrument was available in Spanish.  An extensive discussion of the unique aspects and numerous applications of the NLSY79 HOME scales can be found in Mott, 2004.

An extensive body of research has developed in which the NLSY79 HOME scales have been used. The NLSY Child Handbook: 1986-1990 describes this research in detail through 1990, emphasizing earlier work that examined the reliability and validity of the HOME. The NLSY Children-1992 provides further evidence regarding linkages between the HOME scale and subscales, and a variety of family and maternal antecedents. Finally, The NLSY79 Child Assessments: Selected Tables provides detailed distributions by age and race/ethnicity for the overall HOME scores and the two subscores. All of these documents are available on the Research/Technical Reports page. Data users are encouraged to use the online, searchable NLS Annotated Bibliography to identify the extensive publications using the HOME items.

HOME scores in the database

As indicated above, the items that mothers complete depend on the age of their child: children under age 3 years, 3-5, 6-9, and 10 and over follow different question sequences. The HOME items and the recoding instructions can be found in Appendix A. Question names for the raw and normed HOME scores for the final survey round are listed in Table 1 in the Child Assessments—Introduction section.