Attachment 6 - New Questions and Lines of Inquiry

Attachment 6—New Questions and Lines of Inquiry

As mentioned earlier in this document, BLS has undertaken a continuing redesign effort to examine the current content of the NLSY79 and provide direction for changes that may be appropriate as the respondents enter middle age. Based on the 1998 redesign conference and subsequent discussions, as well as our experiences in 2000–2006, the 2008 instrument reflects a number of changes recommended by experts in various social science fields and by our own internal review of the survey’s content. The major changes are described in this attachment. Additions to the questionnaire have been balanced by deletions of previous questions so that the overall time required to complete the survey should remain about the same.

Additions/Modifications

Assets. After round 19, we determined that it was not necessary to ask an extended series of assets questions in every survey round. The assets module was excluded for rounds 20 and 22 and included in round 21. It will again be rotated into the survey in round 23, using the same questions as round 21. These questions, which appear after income in the survey, are numbered SC.1-SC.12b, FA.1-FA.11a, Q13.116-Q13.116A, NFA.0-NFA.2c, Q13.127A-Q13.132D, NFA.4-NFA.7d, DBT.1-DBT.4, Q13.141-Q13.142, and PS.1-PS.6.

Retirement Expectations. As our respondents move closer to their retirement, we must start to gather information about how they are planning for retirement and, indeed, even how they define retirement. As we examined the Health and Retirement Survey and other sources of information on retirement it became apparent that this is a complex issue. It is no longer sufficient to ask what a respondent will do upon reaching age 65 or 67; rather, we must seek to accommodate retirement plans ranging from a reduction in hours worked to a change in occupation (to a less demanding job) to a complete cessation of employment. In round 22, we included a retirement expectations experiment, with questions addressed to approximately 1000 respondents, to help us develop a module for use with the entire sample.

Most questions in the experiment worked well, and respondents were able to answer the questions without experiencing confusion. Therefore, most of the questions will be included in round 23 for the whole sample without significant changes from round 22. The questions which remain the same are RETIRE.EXP.P2.2-RETIRE.EXP.P2.6. The only change in this series of questions is an expansion of the response categories in question RETIRE.EXP.P2.2, based on responses in round 22 that were not able to be coded in the original list of categories.

The experiment indicated that we needed to include a couple of questions investigating the employment status of respondents who are not currently employed before entering the retirement module. Although the cohort is still fairly young, a few respondents are already retired (particularly those in the military), and some respondents are out of the labor market and do not expect to work for pay in the future (e.g., homemakers and respondents who are disabled). These initial screening questions will permit us to ask these respondents only the questions that make sense for their situation.

Finally, the retirement expectations experiment included an open-ended question, RETIRE.EXP.P2.1, intended to help us determine the response categories necessary for RETIRE.EXP.P2.2. This question is no longer necessary since the experiment is over and will be dropped.

Spouse/Partner Race and Ethnicity. Although we have collected a great deal of information about respondents’ spouses and partners, we have never asked the respondent to report the race or Hispanic ethnicity of spouses or partners. This is considered to be vital information by sociologists, demographers, and many other social scientists, and race/ethnicity is frequently a key control variable in a wide variety of analyses. Beginning in round 23, we plan to add this information to the data collected about spouses and partners. Race and Hispanic ethnicity will be collected for all spouses and partners reported at each interview using the standard OMB format in SPARRACE.1-2. In round 23 only, we will also ask retrospective questions about spouses and partners reported by the respondent in past rounds, to fill in this important gap in our demographic data.

Highest Degree Received. In each round, we update the highest degree received information for any respondents who report attending school since the last interview. However, it has been many rounds since we independently recorded this information for the entire sample. Asking a sample-wide highest degree received question provides a valuable double-check on the accuracy of this vital piece of information, at very low cost in terms of time or respondent burden. The questions for highest degree ever received, and date of degree, are Q3.10D and Q3.10E.

Health. The most extensive changes in the survey occurred in the health section. In the past few rounds, the survey has included a “40+ health module,” a series of questions about health status and behaviors addressed to respondents in the first interview after they turned 40. All respondents have now aged out of this module, so it will be dropped. In its place, we have developed a “50+ health module,” with the goal of updating data from the 40+ module and incorporating new questions representing additional health concerns of aging. This module was developed after extensive consultation with medical and public health experts about the types of data needed for health research, especially as it relates to the longitudinal nature of the survey and the rich store of background information on respondents collected over the past few decades.

Many of the questions in the 50+ module are simple repetitions of the 40+ module. These include the CES-D (depression) scale, health status of the respondent’s biological parents, general state of health and health problems in the past 4 weeks, whether the respondent suffers from a variety of specific health conditions (for example, high blood pressure, cancer, lung disease). These repeated questions are the Q11.H50CESD series, the Q11.H50BPAR series, the Q11.H50SF12 series, and all questions in the Q11.H50CHRC.1-Q11.H50CHRC.9B range not specified below as new.

Based on feedback from our panel of experts, we also incorporated new questions in response to new areas of research in public health and the aging of the respondents. New questions about specific health conditions include skin cancer (Q11.H50CHRC.2B), asthma (Q11.H50CHRC.3E), depression (Q11.H50CHRC.7B-7D), and osteoporosis (Q11.H50CHRC.9C). The 50+ health module also includes new questions on functional limitations in performing various activities (Q11.H50FL.1-2). Q11.H50FL.1 is taken from the National Health Interview Survey (NHIS), and the FL.2 series is based on questions in the Health and Retirement Survey (HRS) and the NLSW. These questions are frequently used in public health research and will permit comparison of the NLSY79 with other large national samples. Another new series (Q11.H50SLP.1-5)asks about respondents’ usual sleep patterns and frequency of sleep problems; these questions are drawn from the National Survey of Midlife Development in the United States (MIDUS II) and the HRS. Again, these questions address an area which recent research suggests may have a significant impact on health, and they will permit comparison with other national surveys. Finally, we have added an open-ended question (Q11.H50OPEN.1) which gives the respondent the opportunity to tell us anything else about their health that they feel is important.

On the advice of the experts, we also deleted some questions from the 40+ module. These include whether the respondent suffered from a broken bone or head injury in the past 10 years, the respondent’s use of corrective lenses and hearing aids, and various items from a lengthy list of minor health concerns (Q11.H40CHRC.10a-ii in the round 22 survey).

As our panel of experts collaborated on the 50+ health module, they also made recommendations for revising the general health questions addressed to all respondents. In general, the changes to this section were made with the intention of addressing emerging concerns in public health research while removing unnecessary questions to reduce time burden. The new questions include:

• Q11.CARE.1-4B, a series of questions on whether the respondent regularly cares for a household or family member due to their health concerns. These questions are important because care giving can directly impact labor supply, and as NLSY79 respondents age they are more likely to spend time caring for aging parents. This series was taken from the NSLW questionnaire, so cross-cohort comparisons will be feasible.

• Q11.GENHLTH.PRV1-2, which ask whether the respondent has a regular heath care provider. These questions will provide important information about the accessibility and type of health care available to the respondent.

• Q11.GENHLTH.4C.M through Q11.GENHLTH.4E.F, a series of questions focused on preventive health care. Respondents are asked to report whether they have screening tests such as cholesterol, diabetes, colonoscopy, blood pressure, PAP smear, etc. Female respondents also report whether they are taking estrogen or other hormone replacement medications. These preventive health measures are widely recognized as an important aspect of an individual’s long-term health.

• Q11.GENHLTH.5A.1-3, which asks about the respondent’s dental health. Recent research indicates that dental health and preventive care can have a significant impact on other aspects of an individual’s health. These questions are drawn from the British National Diet and Nutrition Survey.

• Q11.GENHLTH.7C.1-7F.2, a series of questions about the respondent’s eating habits. These questions are drawn from the National Longitudinal Survey of Adolescent Health (wave III).

The series of questions on the source of the respondent’s health insurance was modified, in an attempt to simplify these questions so that respondents can answer them more quickly and easily. The new questions are based on the NHIS 2005 Family Health Module, with some additional wording from the PSID and MIDUS. These questions collect the same basic information about health insurance coverage as the previous series, but we anticipate that they will be perceived as easier by respondents. Further, the new questions will permit easier comparison of the NLSY79 with other national surveys. These questions are Q11.HLTHPLN.INTRO through Q11.80B, Q11.83-84B, and Q11.87-88B.

Deleted health questions include whether the respondent has missed any days of work due to a child’s asthma (Q11.ASTHMA.10-10B in round 22) and health care topics discussed at the respondent’s last general exam (Q11.GENHLTH.4B in round 22).

Deletions

Current Population Survey. Round 22 included questions on current labor force status based on the Current Population Survey. After the 1998 redesign conference, it was determined that they could be repeated every third or fourth round rather than every round. Round 22 represented their regular rotation, so they will be excluded in round 23. “CPS” questions were those prefixed with “Q5” (Q5.1-Q5.93).

Attitudes. The round 22 survey included a series of questions adapted from the Rosenberg Self-Esteem Scale. This scale is included in the survey only periodically (1980, 1987, and 2006), so it is not necessary to re-ask the scale in 2008.

Consumption, Impatience, and Risk. With regard to consumption, questions were added in round 22 that gather information on monthly amounts spent on groceries, non-food items, eating out, telephone service, internet service, and utilities. The questions were CONSUMPTN_1 through CONSUMPTN_8A. These questions are intended to permit researchers to explore important aspects of wealth accrual in middle age. However, we feel that it is not feasible nor necessary to ask these questions in every survey round; we may consider rotating them back into the survey at some point in the future.

Round 22 also included a series on risk and impatience. Numbered RISK_1 through RISK_4_SR000001 and IMPATIENCE_1 through IMPATIENCE_2, these questions were intended to help researchers to identify attitudes that may affect a variety of choices respondents may make with respect to investing, job changes, retirement planning, and the like. Existing research on risk and impatience indicates that these attitudes change very slowly, so it is not necessary to ask these questions in every round.

Finally, several recent surveys including round 22 contained a set of questions on job risk. While added to the survey before the risk and impatience series mentioned above, these questions complemented that series. As the job risk questions (JOB_RISK-1 through JOB_RISK-3) similarly record attitudes that change slowly over time, it is not necessary to ask these in every round and they are being rotated out of the survey.

Volunteerism. Round 22 included a series of questions about volunteerism, such as whether respondents had any unpaid volunteer work, the number of weeks, number of hours worked per week, types of organizations, and the types of organizations through which the respondents volunteered the most. The questions were ACP_16A through ACP_20A. Although these questions are important and provide an opportunity for cross-generational analysis with other NLS cohorts, we do not view them as core questions and so we intend to ask them only on a periodic basis.