Retention & Reasons for Noninterview

National Longitudinal Survey of Youth - 1979 Cohort

Retention & Reasons for Noninterview

Sample Representativeness & Attrition

This section reviews the number of respondents by race, sex, and NLSY79 sample type who have continued to be interviewed during all surveys. It also takes a brief look at the racial composition of the cohort at the initial and latest survey points. Table 4 shows the number of respondents, excluding dropped respondents, who were interviewed at all survey points. This table exhibits the high degree of NLSY79 retention. From 1979 to 2014 the survey has been administered 26 times.

Table 4. Percentage of NLSY79 Respondents, Excluding Dropped Respondents, Who Answered Every Survey: 1979-2014

Year Percent Number   Year Percent Number
1979 100% 9964 1992 74.2% 7396
1980 96.0% 9571 1993 73.2% 7291
1981 94.3% 9395 1994 71.8% 7153
1982 92.7% 9234 1996 69.6% 6935
1983 91.6% 9125 1998 66.9% 6664
1984 89.7% 8942 2000 63.3% 6310
1985 87.5% 8721 2002 60.3% 6004
1986 85.0% 8472 2004 57.6% 5736
1987 82.3% 8203 2006 55.6% 5538
1988 79.9% 7957 2008 54.3% 5407
1989 78.5% 7819 2010 52.3% 5208
1990 76.7% 7642 2012 50.4%  5013
1991 75.5% 7521 2014 48.1% 4788

Table 5 shows the distribution of the number of interviews completed by respondents, broken down by sample type. The "# who completed" column shows how many respondents completed exactly that number of surveys. These numbers refer to any surveys completed since the NLSY79 began, not necessarily consecutive surveys completed or surveys completed in particular years. The cumulative percent column shows a cumulative total percent of those completing at least a given number of surveys rather than a percentage of those completing an exact number of surveys. Readers should note the attrition suggested in Table 3 greatly overstates the amount of lost information. The NLSY79 asks detailed questions about work history, education, training, marital status, and fertility since the date of the respondent's last interview. These retrospective questions capture information lost due to missed interviews. Hence, a perfect response record is not needed for researchers to understand how the respondent's life changes over time, unless he or she leaves the survey forever.

Table 5. Number of Interviews Respondents Completed out of 25 Surveys by Sample Type: NLSY79 1979-2014

  Total Sample Cross Sectional Sample Supplemental Sample Military Sample
# of Surveys1 # who Completed Cumul. Percent # who Completed Cumul. Percent # who Completed Cumul. Percent # who Completed Cumul. Percent
1 40  0.4 26 0.4 14 0.4 0 -
2 38  0.4 22 0.4  15  0.4  1 0. 5
3 28 0.3 18 0.3 10 0.3 0 -
4 40 0.4 19 0.3 19 0.5 2 1.0
5 52 0.5 39 0.6 12 0.3 1 0.5
6 62 0.6 36 0.6 25 0.7 1 0.5
7 75 0.8 53 0.9 21 0.6 1 0.5
8 70 0.7 44 0.7 25 0.7 1 0.5
9 88 0.9 53 0.9 35 1.0 0 -
10 72 0.7 45 0.7 26 0.7 1 0.5
11 75 0.8 40 0.7 35 1.0 0 -
12 74 0.7 54 0.9 18 0.5 2 1.0
13 107 1.1 66 1.1 41 1.1 0 -
14 114 1.1 78 1.3 35 1.0 1 0.5
15 116 1.2 72 1.2 42 1.2 2 1.0
16 175 1.8 107 1.8 62 1.7 6 3.0
17 180 1.8 130 2.1 47 1.3 3 1.5
18 251 2.5 162 2.7 78 2.1 11 5.5
19 254 2.6 156 2.6 92 2.5 6 3.0
20 285 2.9 165 2.7 114 3.1 6 3.0
21 285 2.9 167 2.7 106 2.9 12 6.0
22 371 3.7 203 3.3 159 4.4 9 4.5
23 478 4.8 262 4.3 199 5.5 17 8.5
24 672 6.7 361 5.9 295 8.1 16 8.0
25 1174 11.8 663 10.9 486 13.3 25 12.4
26 4788 48.1 3070 50.2 1641 44.9 77 38,3
Total 9964 100% 6111 100% 3652 100% 201 100%
 
Note: Universe excludes the 1,079 members of the military subsample and the 1,643 members of the economically disadvantaged, non-black/non-Hispanic oversample dropped from interviewing; it includes the remaining 9,964 eligible members.
1 Surveys completed in any year, not necessarily consecutive survey years.