## Codes and Frequencies

## Description

For sample children 4-17 who had difficulties with emotions, concentration, behavior, or being able to get along with others and who had these difficulties for one month or more, DIFSCORE reports the impact score based on responses to five variables:

- Difficulties upset or distressed the child (DIFUPSET)

- Child's difficulties interfere with home life (DIFHOME)

- Child's difficulties interfere with friendships (DIFFRIEND)

- Child's difficulties interfere with classroom learning (DIFLEARN)

- Child's difficulties interfere with leisure activities (DIFLEISURE)

NHIS began offering DIFSCORE in 2019, and IPUMS NHIS staff constructed the variable for 2001, 2003, and 2004. Beginning in 2019, those with more than one of the five variables missing are assigned a code of 98, "Not ascertained-more than one response is missing."

Calculating a Score for the SDQ-EX Impact Supplement

The SDQ Scoring Guide provides guidance on how responses for DIFUPSET, DIFHOME, DIFFRIEND, DIFLEARN, and DIFLEISURE "can be summed to generate an impact score that ranges from 0 to 10 for the parent-completed version." Under these guidelines, the responses "not at all" and "a little" are given a score of 0; a response of "a medium amount" is given a score of 1; and a response of "a great deal" is given a score of 2. For each of these variables, a score of 0 implies the least impact from the child's difficulty, and a score of 2 implies the greatest impact from the child's difficulty. If the scores for these five variables are summed, the total ranges from 0 to 10, with 0 implying the least impact from the child's difficulties and 10 implying the greatest impact from the child's difficulties.

IPUMS NHIS staff constructed DIFSCORE for 2001, 2003, and 2004 by summing the scores across DIFUPSET, DIFHOME, DIFFRIEND, DIFLEARN, and DIFLEISURE. These five variables receive codes in IPUMS NHIS that facilitate this scoring process. Responses that should receive a score of 0 in the scoring process--namely, "not at all" or "a little"--receive a code of 0 in the first digit. In addition, sample children with no reported difficulties (in EMODIFF) or with difficulties lasting "less than one month" (code 1 in DIFMOS) for 2001, 2003, and 2004 also receive a code of 0 in the first digit in these five variables. (The three different categories, "No difficulty or no difficulty for greater than or equal to 1 month," "not at all," and "a little" can still be distinguished from each other by the respective values of 0, 1, and 2 in the second digit.) The response "a medium amount," which should receive a score of 1, receives a code of 1 in the first digit (and a code of 0 in the second digit). The response "a great deal," which should receive a score of 2, receives a code of 2 in the first digit (and a code of 0 in the second digit). Cases which should be excluded from this summing of scores--that is, not in universe cases consisting of persons other than sample children age 4-17, and cases with a response of "unknown"--all receive a code of 9 in the first digit (with the second digit distinguishing between those not in the variable universe and unknowns).

Put succinctly, to calculate an "impact score" for the child's difficulties, researchers should 1) exclude cases beginning with a code of 9; 2) group together other categories that share a common first digit; 3) sum the scores across the five variables, using only the first digit of the codes for each variable. The result will range from 0 to 10.

Prior to 2019, IPUMS NHIS staff constructed DIFSCORE by following the above guidelines. Beginning in 2019, NHIS began offering the impact score as a variable constructed by following the guidelines in the SDQ Scoring Guide.

According to the SDQ Scoring Guide, the impact score can be used as a continuous variable, but it is sometimes convenient to classify it into broader categories for easier interpretation of results. Under the original three-band categorization, total impact scores can be identified as normal, borderline, or abnormal: a total impact score of 2 or more is abnormal; a score of 1 is borderline; and a score of 0 is normal. Under the newer four-band categorization, total impact scores can be identified as close to average, slightly raised, high, or very high: a total impact score of 0 is close to average; a score of 1 is slightly raised, a score of 2 is high, and a score of 3 or higher is very high.

## Comparability

DIFSCORE is constructed from responses to DIFUPSET, DIFHOME, DIFFRIEND, DIFLEARN, and DIFLEISURE, which are part of the extended version of the Strengths and Difficulties Questionnaire (SDQ-EX) fielded for sample children in the NHIS in 2001, 2003, 2004, and 2019 forward. It is not an element of the abbreviated version of SDQ, which was fielded for sample children in 2002, 2005-2007, and 2010-2018.

The NHIS questionnaire was substantially redesigned in 2019 to introduce a different data collection structure and new content. For more information on changes in terminology, universes, and data collection methods beginning in 2019, please see the user note.

## Universe

- 2001; 2003; 2004: Sample children age 4 to 17 who had a known response for whether they had difficulties with emotions, concentration, or behavior and had a known response for how many months their difficulties were present.
- 2019: Sample children age 4 to 17 who have had minor, definite, or severe difficulties with emotions, concentration, behavior, or being able to get along with other people (EMODIFF).

## Availability

- 2001, 2003-2004, 2019

## Weights

- 2001, 2003-2004, 2019 : SAMPWEIGHT