Survey Methodology #2006-02 - Use Of Dependent Interviewing Procedures To Improve Data Quality In The Measurement Of Change - U.s. Census Bureau Page 23

ADVERTISEMENT

PART 2: CHARACTERISTICS WITH THE SAME DEPENDENT INTERVIEWING PROCEDURES IN BOTH PANELS
n = 51,950
n = 48,508
n = 47,809
#
2001
67.9%
0.15
0.05
0.71
14.2
0
0.86
ª = 706
ª = 429
ª = 404
Medicare
n = 75,713
n = 70,152
n = 68,067
#
2004
69.5%
0.12
0.04
0.59
14.8
0.02
0.71
ª = 868
ª = 483
ª = 445
n = 40,684
n = 29,533
n = 36,562
@
2001
36.2%
5.4
4.0
13.7
3.4
8.8
33.6
ª = 15,410
ª = 11,811
ª = 13,140
employment at the
3/
same job
n = 52,799
n = 50,862
n = 48,873
@
2004
37.7%
5.7
4.1
15.3
3.7
9.4
44.8
ª = 17,510
ª = 22,052
ª = 20,879
Table 1 Notes
Table entries represent the simple mean of three estimates derived from individual analyses of each of three pairs of SIPP waves – waves 1 and 2, 2 and 3, and 3
and 4. Each individual analysis included only “adults” (people age 15+) for whom a completed interview was obtained in both waves, and excluded cases with
missing data in either seam month. Additional analysis restrictions are as follows:
1/ Analysis restricted to people aged 20+.
2/ Analysis restricted to people who reported at least one month of receipt of this income type in either of the two successive survey waves.
3/ The unit of analysis is the job, rather than the person; the analysis includes all jobs held in at least one month in either of the two successive survey
waves. Note that an error in the 2001 wave 3 rotation 1 instrument resulted in the inability to link wave 3 jobs with wave 2 jobs. The error only affected
rotation group 1 – the instrument operated correctly for rotation groups 2, 3, and 4. Therefore, the data from rotation 1 are excluded from the 2001 W2-
W3 analysis.
“AVERAGE % OF ALL CHANGES THAT WERE SEAM CHANGES” – statistical analysis notes:
***
[and estimates in bold font] According to a simple t-test of the difference between two proportions, the 2001-2004 difference was statistically significant
in each of the three individual analyses.
**
According to a simple t-test of the difference between two proportions, the 2001-2004 difference was significant in two of the three individual analyses,
and in the same direction (but non-significant) in the third.
#
All three individual 2001-2004 differences were non-significant.
@ All three individual 2001-2004 differences were statistically significant, but the sign of the difference was inconsistent across the three comparisons.

ADVERTISEMENT

00 votes

Related Articles

Related forms

Related Categories

Parent category: Legal