APPENDIX C

EXAMPLES IN SAS

In SAS, the NCVS sample design must be appropriately specified for all survey
procedures using the varmethod, strata, and cluster statements. The varmethod=taylor statement
is included within the proc statement of each analysis. The following statements must be
included to provide the stratification and PSU variables.

strata yr_grp v2117;

cluster v2118;
Weight statements are also required, but vary depending on the type of estimate (as
shown in Table 3-1). Because SAS does not support the specification of formulas within the
analysis procedures, for the calculation of victimization rates analysts must first create a new
variable equal to the product of the victimization count and the adjustment factor (ADJINC_WT),
multiplied by 1,000 (as outlined in the examples below).

The domain statement is used to limit the analysis to the appropriate set of cases 1. While
all combinations of variables included in the domain statement will be included in the results,
only the results meeting all subpopulation conditions are relevant. For this reason, prior to each
analysis a “sub” variable is created based on all exclusions (e.g. victimizations occurring outside
of the US and outside the year(s) of interest). Including this single recoded variable on the
domain statement simplifies examination of the results.

Examples 1 and 2 demonstrate how to estimate the total number of victimizations.

Examples 3 and 4 demonstrate how to calculate the proportion of victimizations with given
characteristics. Examples 5 and 6 demonstrate the calculation of personal and property
victimization rates for victimization characteristics included on the modified person and

1
In Section 3, the term analysis domain is used to define subsets of the population based on victims or
victimizations. In SAS, the domain statement is used to identify records to include in analyses. To avoid
confusion between the two uses, the term “analysis class” will replace the term “analysis domain” in
Appendix C.

1
household files, and Example 7 demonstrates the calculation of victimization rates for
victimization characteristics not on the modified files. Examples are included for both single-
and pooled-year estimates.

C.1 Victimization Totals

The SAS surveymeans procedure is used to estimate the total number of victimizations
from the modified incident-level file. The specific crime type of interest is specified in the var
statement, and the analysis class(es) of interest (i.e. subsets of the population based on
characteristics of the victimization or victim) are specified on the domain statement. Examples
1A-1C demonstrate the calculation of victimization totals for single years, and Examples 2A-2C
demonstrate pooled year estimates.

Example 1: Number of victimizations, single year

Examples 1A and 1B below are estimates of personal crimes, while Example 1C is an
estimate of property crimes. Examples 1A and 1C provide overall crime estimates, while
Example 1B provides estimated totals for person-level characteristics of interest.

Example 1A - Total number of violent victimizations, 2011

Year(s): 2011
Crime Type: violent victimizations (vcrime)
Analysis Classes: n/a
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2011 (year=2011)

Code:
data anset;
set adjincident;
sub=(exclude_outUS=0 and dummy=0 and year=2011);
run;

proc surveymeans data=anset varmethod=taylor sum; *a;

strata yr_grp v2117; *c;
cluster v2118; *d;
weight newwgt; *e;
domain sub; *b;
var vcrime; *f;
run;

2
Code Comment(s):
a) Specifies input dataset and sample design
b) Subpopulation indicator
c) Specifies stratification
d) Specifies primary sampling units
e) Specifies analysis weight
f) Outcome of interest

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted rows).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 37853
Sum of Weights 115824878

Statistics

Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
vcrime 29789137 1061238
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub

sub Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 vcrime 23976614 927768
1 vcrime 5812523 357912
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

3
Example 1B - Total number of violent victimizations by sex, age category, and race/ethnicity,
2011
Year(s): 2011
Crime Type: violent victimizations (vcrime)
Analysis Classes: sex (sex), age category (agecat), and race/ethnicity (race_eth)
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2011 (year=2011)

Code:
data anset;
set adjincident;
sub=(exclude_outUS=0 and dummy=0 and year=2011);
run;

proc surveymeans data=anset varmethod=taylor sum;

domain sub*sex sub*agecat sub*race_eth; *a;
strata yr_grp v2117;
cluster v2118;
weight newwgt;
var vcrime;
run;

Code Comment(s):
a) Analysis class variables (victim characteristics), crossed with
subpopulation indicator

Output:
NOTE: Output is provided overall and by subpopulation. Only the output
under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted rows).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 37853
Sum of Weights 115824878

Statistics

Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
vcrime 29789137 1061238
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

4
 Domain Analysis: sub*sex

sub sex Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Male vcrime 12443058 568391
Female vcrime 11533557 551251
1 Male vcrime 3209725 236961
Female vcrime 2602798 252786
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*agecat

sub agecat Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 12 to 14 vcrime 2350414 190557
15 to 17 vcrime 2351591 231698
18 to 20 vcrime 2120993 169061
21 to 24 vcrime 2671424 218252
25 to 34 vcrime 5052311 372358
35 to 49 vcrime 5846191 336304
50 to 64 vcrime 3086188 201164
65 or older vcrime 497502 61068
1 12 to 14 vcrime 488383 77368
15 to 17 vcrime 428880 64692
18 to 20 vcrime 843753 220104
21 to 24 vcrime 617720 69926
25 to 34 vcrime 1114834 118262
35 to 49 vcrime 1365450 141011
50 to 64 vcrime 776857 90922
65 or older vcrime 176646 46321
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*race_eth

sub race_eth Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Non-Hispanic White vcrime 16135969 781567
Non-Hispanic Black vcrime 3535332 327426
Hispanic vcrime 2877571 187303
Non-Hispanic Other vcrime 807284 131642
Non-Hispanic More than One Race vcrime 620459 102505
1 Non-Hispanic White vcrime 3719729 242039
Non-Hispanic Black vcrime 812173 181536
Hispanic vcrime 895434 130311
Non-Hispanic Other vcrime 197082 46736
Non-Hispanic More than One Race vcrime 188105 43090
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

5
Example 1C - Total number of property victimizations, 2011
Year(s): 2011
Crime Type: property victimizations (pcrime)
Analysis Classes: n/a
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2011 (year=2011)

Code:

data anset;
set adjincident;
sub=(exclude_outUS=0 and dummy=0 and year=2011);
run;

proc surveymeans data=anset varmethod=taylor sum;

domain sub;
strata yr_grp v2117;
cluster v2118;
weight newwgt;
var pcrime;
run;

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 37853
Sum of Weights 115824878

Statistics

Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
pcrime 85126878 1577101
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub

sub Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

6
 0 pcrime 68063730 1324122
1 pcrime 17063148 543579
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

7
Example 2: Number of crimes, aggregated years
Examples 2A and 2B are estimates of personal crimes, while Example 2C is an estimate
of a property crime. Examples 2A and 2C provide overall crime estimates and Example 2B
provides estimated totals for a victimization-level characteristic of interest. The code provided
calculates the estimated number of victimization across the pooled years. To calculate the
average number of victimizations per year, estimated totals and standard errors must be divided
by the number of pooled years, as shown in the following examples.

Example 2A - Total and Average number of aggravated assaults, 2009-2011

Year(s): 2009-2011
Crime Type: aggravated assaults (aast)
Analysis Classes: n/a
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2009-2011 ((2009 <= year) and (year <=2011))
Code:

data anset;
set adjincident;
sub=(exclude_outUS=0 and dummy=0 and (2009 <= year) and
(year <=2011)); *a;
run;

proc surveymeans data=anset varmethod=taylor sum;

domain sub;
strata yr_grp v2117;
cluster v2118;
weight newwgt;
var aast;
run;

Code Comment(s):
a) Subpopulation indicator incorporates the range of years to be
analyzed

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

8
The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 37853
Sum of Weights 115824878

Statistics

Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
AAST 5189174 257111
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub

sub Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 AAST 2248758 206953
1 AAST 2940416 160814
Ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

NOTE: The estimate above represents the total number of aggravated assaults from 2009-
2011. To obtain the average number of aggravated assaults, both the estimate and the standard
error must be divided by the number of pooled years (3), as follows:

𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎

𝑎𝑎𝑎𝑎𝑎𝑎 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 =
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦

2940416
=
3

= 980139

𝑠𝑠𝑠𝑠(𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎)

𝑆𝑆𝑆𝑆(𝑎𝑎𝑎𝑎𝑎𝑎 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎) =
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦

160814
=
3

= 53605

9
Example 2B - Total and average number of aggravated assaults involving firearm, 2009–2011
Year(s): 2009-2011
Crime Type: aggravated assaults (aast)
Analysis Classes: weapon category (weapcat)
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2009-2011 ((2009 <= year) and (year <=2011))

Code:

data anset;
set adjincident;
sub=(exclude_outUS=0 and dummy=0 and (2009 <= year) and
(year <=2011));
run;

proc surveymeans data=anset varmethod=taylor sum;

domain sub*weapcat;
strata yr_grp v2117;
cluster v2118;
weight newwgt;
var aast;
run;

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 37853
Sum of Weights 115824878

Statistics

Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
AAST 5189174 257111
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

10
 Domain Analysis: sub*weapcat

sub weapcat Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 No Weapon AAST 110713 21766
Firearm AAST 647127 176781
Knife or sharp object AAST 601554 74541
Other type weapon AAST 706901 98434
Type weapon unknown AAST 158175 40728
Do Not Know if off had weapon AAST 24289 6567.738419
1 No Weapon AAST 318542 53740
Firearm AAST 778706 81552
Knife or sharp object AAST 642022 65521
Other type weapon AAST 939927 96628
Type weapon unknown AAST 228755 52288
Do Not Know if off had weapon AAST 32463 11506
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

NOTE: The estimate above represents the total number of aggravated assaults involving a
firearm from 2009-2011. To obtain the average number of aggravated assaults involving a
firearm, both the estimate and the standard error must be divided by the number of pooled years
(3), as follows:

𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑒𝑒𝑒𝑒 𝑜𝑜𝑜𝑜 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑤𝑤 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓

𝑎𝑎𝑎𝑎𝑎𝑎 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑤𝑤 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 =
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦

778706
=
3

= 259569

𝑠𝑠𝑠𝑠(𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑡𝑡𝑠𝑠 𝑤𝑤 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓)

𝑆𝑆𝑆𝑆(𝑎𝑎𝑎𝑎𝑎𝑎 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑤𝑤 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓) =
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦

81552
=
3

= 27184

11
Example 2C - Total and average number of household burglaries, 2009-2011
Year(s): 2009-2011
Crime Type: household burglary (hburg)
Analysis Classes: n/a
Weight: newwgt
Subpopulation: Within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2009-2011 ((2009 <= year) and (year <=2011))

Code:

data anset;
set adjincident;
sub=(exclude_outUS=0 and dummy=0 and (2009 <= year) and
(year <=2011));
run;
proc surveymeans data=anset varmethod=taylor sum;
domain sub;
strata yr_grp v2117;
cluster v2118;
weight newwgt;
var hburg;
run;

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 37853
Sum of Weights 115824878

Statistics

Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
HBURG 17350128 480495
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub

sub Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 HBURG 7148811 254130

12
 1 HBURG 10201317 336711
Ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

NOTE: The estimate above represents the total number of household burglaries from
2009-2011. To obtain the average number of household burglaries, both the estimate and the
standard error must be divided by the number of pooled years (3), as follows:

𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑟𝑟 𝑜𝑜𝑜𝑜 ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜ℎ𝑜𝑜𝑜𝑜𝑜𝑜 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏

𝑎𝑎𝑎𝑎𝑎𝑎 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜ℎ𝑜𝑜𝑜𝑜𝑜𝑜 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 =
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦

10201317
=
3

= 3400439

𝑠𝑠𝑠𝑠(𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜ℎ𝑜𝑜𝑜𝑜𝑜𝑜 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏)

𝑆𝑆𝑆𝑆(𝑎𝑎𝑎𝑎𝑎𝑎 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜ℎ𝑜𝑜𝑜𝑜𝑜𝑜 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏) =
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦

336711
=
3

= 112237

13
C.2 Victimization Proportions
The SAS surveyfreq procedure is used to estimate the percent distribution of
victimizations in a specific analysis class across covariates from the modified incident-level file.
Subpopulation and covariates of interest are specified on the table statement.

Example 3: Percent distribution of victims across characteristics, single year

Example 3A is an estimate of personal crimes, while Example 3B is an estimate of
property crimes. In Example 3A, both the analysis classes and covariates are specified based on
person- and incident-level characteristics. In Example 3B, the analysis class is specified based on
an incident-level characteristic, while the covariates are household-level characteristics.

Example 3A - Distribution of female violent crime victims by age category, race/ethnicity, and
victim-offender relationship, 2011
Year(s): 2011
Analysis Classes: female (sex=2); violent victimizations (vcrime=1)
Covariate(s): age category (agecat); race/ethnicity (race_eth); victim-offender relationship
(direl)
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2011 (year=2011); female (sex=2); violent victimizations (vcrime=1)

Code:

data anset;
set adjincident;
sub=(exclude_outUS=0 and dummy=0 and year=2011 and sex=2 and
vcrime=1);
run;

proc surveyfreq data=anset varmethod=taylor; *a;

strata yr_grp v2117; *b;
cluster v2118; *c;
weight newwgt; *d;
table /*e*/sub*(/*f*/agecat race_eth direl) / row nocellpercent
nofreq nowt;
run;

Code Comment(s):
a) Specifies input dataset and sample design
b) Specifies stratification
c) Specifies primary sampling units
d) Specifies analysis weight
e) Subpopulation indicator
f) Covariates of interest

14
Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYFREQ Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 37853
Sum of Weights 115824878

Table of sub by agecat

Row Std Err of

sub agecat Percent Row Percent
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 12 to 14 4.4827 0.2112
15 to 17 5.2166 0.2419
18 to 20 6.6596 0.2631
21 to 24 9.0951 0.3107
25 to 34 20.6693 0.4669
35 to 49 27.2827 0.4232
50 to 64 19.1136 0.4089
65 or older 7.4803 0.2600

Total 100.000
----------------------------------------------
1 12 to 14 7.6428 1.8588
15 to 17 6.2620 1.3309
18 to 20 15.5020 5.3677
21 to 24 9.3116 1.6367
25 to 34 19.3023 2.3616
35 to 49 24.5841 3.0542
50 to 64 12.9129 1.9929
65 or older 4.4822 1.7220
Total 100.000
----------------------------------------------
Total 12 to 14
15 to 17
18 to 20
21 to 24
25 to 34
35 to 49
50 to 64
65 or older

Total
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
Frequency Missing = 1600

Table of sub by race_eth

Row Std Err of

sub race_eth Percent Row Percent
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Non-Hispanic White 65.4165 0.7871
Non-Hispanic Black 14.2469 0.6862
Hispanic 14.4004 0.5881
Non-Hispanic Other 3.6912 0.2494
Non-Hispanic More than One Race 2.2450 0.1751

15
 Total 100.000
------------------------------------------------------------------
1 Non-Hispanic White 59.8656 4.4568
Non-Hispanic Black 19.4289 5.1812
Hispanic 11.6743 1.9826
Non-Hispanic Other 4.4212 1.5083
Non-Hispanic More than One Race 4.6101 1.2872

Total 100.000
------------------------------------------------------------------
Total Non-Hispanic White
Non-Hispanic Black
Hispanic
Non-Hispanic Other
Non-Hispanic More than One Race

Total
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
Frequency Missing = 1600

Table of sub by direl

Row Std Err of

sub direl Percent Row Percent
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 intimates 10.6971 0.6071
other relatives 6.8193 0.5450
well known/casual acquaintances 26.4125 0.7750
strangers 33.1309 0.8289
do not know relationship 6.6340 0.2820
do not know number of offenders 16.3063 0.5064

Total 100.000
------------------------------------------------------------------
1 intimates 23.2306 3.5018
other relatives 9.9822 1.6418
well known/casual acquaintances 35.2831 4.5594
strangers 25.3356 3.1567
do not know relationship 2.7056 0.6099
do not know number of offenders 3.4630 0.9380

Total 100.000

16
------------------------------------------------------------------
Total intimates
other relatives
well known/casual acquaintances
strangers
do not know relationship
do not know number of offenders

Total
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
Frequency Missing = 26335

17
Example 3B - Distribution of property crime victims by household income, MSA status, and
region, 2011
Year(s): 2011
Analysis Classes: property victimizations (pcrime=1)
Covariate(s): household income (hincome); MSA status (msa); region (region)
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2011 (year=2011); property victimizations (pcrime=1)

Code:

data anset;
set adjincident;
sub=(exclude_outUS=0 and dummy=0 and year=2011 and pcrime=1);
run;

proc surveyfreq data=anset varmethod=taylor;

strata yr_grp v2117;
cluster v2118;
weight newwgt;
table sub*(hincome msa region) / row nocellpercent nofreq nowt;
run;

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYFREQ Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 37853
Sum of Weights 115824878

Table of sub by hincome

Row Std Err of

sub hincome Percent Row Percent

18
-----------------------------------------------------
0 Less than $7,500 5.8068 0.3420
$7,500 to $14,999 7.9607 0.3342
$15,000 to $24,999 9.7039 0.3290
$25,000 to $34,999 9.2893 0.3713
$35,000 to $49,999 11.4065 0.3506
$50,000 to $74,999 11.6421 0.3815
$75,000 or more 17.9292 0.4836
Unknown 26.2615 0.7121

Total 100.000
-----------------------------------------------------
1 Less than $7,500 6.4502 0.6755
$7,500 to $14,999 8.6763 0.6591
$15,000 to $24,999 10.2487 0.6577
$25,000 to $34,999 8.8855 0.6006
$35,000 to $49,999 11.7194 0.6256
$50,000 to $74,999 11.4977 0.6959
$75,000 or more 16.4669 0.7300
Unknown 26.0553 1.1284

Total 100.000
-----------------------------------------------------

Total Less than $7,500

$7,500 to $14,999
$15,000 to $24,999
$25,000 to $34,999
$35,000 to $49,999
$50,000 to $74,999
$75,000 or more
Unknown

Total
-----------------------------------------------------
Frequency Missing = 1600

Table of sub by msa

Row Std Err of

sub msa Percent Row Percent
--------------------------------------------
0 Urban 39.8233 1.3159
Suburban 44.9932 1.3979
Rural 15.1835 1.5780

Total 100.000
-------------------------------------------
1 Urban 40.5967 1.5111
Suburban 46.6167 1.6160
Rural 12.7866 1.6281

Total 100.000

19
-------------------------------------------
Total Urban
Suburban
Rural

Total
--------------------------------------------
Frequency Missing = 1600

Table of sub by region

Row Std Err of

sub region Percent Row Percent
--------------------------------------------
0 Northeast 14.2339 0.5989
Midwest 24.0198 1.1795
South 35.9385 1.0096
West 25.8078 0.7835

Total 100.000
--------------------------------------------

1 Northeast 13.4005 1.0813

Midwest 21.4105 1.4478
South 34.8072 1.5195
West 30.3818 1.3962

Total 100.000
--------------------------------------------
Total Northeast
Midwest
South
West

Total
--------------------------------------------
Frequency Missing = 1600

20
Example 4: Percent distribution of victims across characteristics, aggregated years
Example 4A is an estimate of personal crimes, while Example 4B is an estimate of
property crimes. Both examples specify multiple covariates based on incident-level
characteristics.

Example 4A - Percent of violent victimizations reported and not reported to police by type of
crime, 2009-2011
Year(s): 2009-2011
Analysis Classes: violent victimizations (vcrime=1)
Covariate(s): report status (notify); type of crime (newoff)
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2009-2011 ((2009 <= year) and (year <=2011)); violent victimizations (vcrime=1)

Code:
data anset;
set adjincident;
sub=(exclude_outUS=0 and dummy=0 and (2009 <= year) and (year
<=2011) and vcrime=1);
run;

proc surveyfreq data=anset varmethod=taylor;

strata yr_grp v2117;
cluster v2118;
weight newwgt;
table sub*newoff*notify / row nocellpercent nofreq nowt;*a;
run;

Code Comment(s):
a) The goal is to obtain the distribution of victimizations by
report status for each type of violent crime. Therefore,
newoff*notify is specified on the table statement and row
percents are requested. If the percent distribution of crimes
among those reported or not reported to police were desired,
column percents would be requested.
Output:
NOTE: Output is provided overall and by subpopulation. Only the output
under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

21
The SURVEYFREQ Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 37853
Sum of Weights 115824878

The SURVEYFREQ Procedure

Table of newoff by notify

Controlling for sub=0

Row Std Err of

newoff notify Percent Row Percent
-------------------------------------------------------------------------
Rape & Sexual Assault reported to police 41.1259 6.2743
not reported to police 58.8741 6.2743
do not know . .

Total 100.000
-------------------------------------------------------------------------
Robbery reported to police 64.8907 3.5035
not reported to police 35.1093 3.5035
do not know . .

Total 100.000
-------------------------------------------------------------------------
Aggravated Assault reported to police 54.4203 5.0909
not reported to police 43.9368 5.0355
do not know 1.6429 0.6685

Total 100.000
-------------------------------------------------------------------------

22
Example 4B - Percent of property victimizations reported and not reported to police by type of
crime, 2009-2011
Year(s): 2009-2011
Analysis Classes: property victimizations (pcrime=1)
Covariate(s): report status (notify); type of crime (newoff)
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2009-2011 ((2009 <= year) and (year <=2011)); property victimizations (pcrime=1)

Code:

data anset;
set adjincident;
sub=(exclude_outUS=0 and dummy=0 and (2009 <= year) and (year
<=2011) and pcrime=1);
run;

proc surveyfreq data=anset varmethod=taylor;

strata yr_grp v2117;
cluster v2118;
weight newwgt;
table sub*newoff*notify / row nocellpercent nofreq nowt;
run;

Output:
NOTE: Output is provided overall and by subpopulation. Only the output
under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYFREQ Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 37853
Sum of Weights 115824878

Table of newoff by notify

Controlling for sub=0
Row Std Err of
newoff notify Percent Row Percent

23
-------------------------------------------------------------------------
Rape & Sexual Assault reported to police 36.1639 4.6125
not reported to police 63.8361 4.6125
do not know . .

Total 100.000
-------------------------------------------------------------------------
Robbery reported to police 63.1051 2.4092
not reported to police 36.4855 2.4511
do not know 0.4093 0.2529

Total 100.000
-------------------------------------------------------------------------
Aggravated Assault reported to police 57.0243 2.6465
not reported to police 41.4185 2.6741
do not know 1.5572 0.6167

Total 100.000
-------------------------------------------------------------------------
Simple Assault reported to police 42.5540 1.3444
not reported to police 56.1714 1.3307
do not know 1.2746 0.2690

Total 100.000
-------------------------------------------------------------------------
Personal Theft reported to police 53.6418 4.1041
not reported to police 44.8037 4.0528
do not know 1.5545 0.9273

Total 100.000
-------------------------------------------------------------------------
Household Burglary reported to police 52.8492 1.3038
not reported to police 45.6522 1.2619
do not know 1.4986 0.5373

Total 100.000
-------------------------------------------------------------------------
Motor Vehicle Theft reported to police 82.4474 1.8638
not reported to police 17.1571 1.8509
do not know 0.3955 0.2299

Total 100.000
-------------------------------------------------------------------------
Theft reported to police 31.8270 0.7299
not reported to police 66.9627 0.7458
do not know 1.2103 0.1756
Total 100.000
-------------------------------------------------------------------------
Total reported to police
not reported to police
do not know
Total
-------------------------------------------------------------------------

24
 Table of newoff by notify

Controlling for sub=1

Row Std Err of
newoff notify Percent Row Percent
-------------------------------------------------------------------------
Rape & Sexual Assault reported to police . .
not reported to police . .
do not know . .
Total .
-------------------------------------------------------------------------
Robbery reported to police . .
not reported to police . .
do not know . .

Total .
-------------------------------------------------------------------------
Aggravated Assault reported to police . .
not reported to police . .
do not know . .

Total .
-------------------------------------------------------------------------
Simple Assault reported to police . .
not reported to police . .
do not know . .

Total .
-------------------------------------------------------------------------
Personal Theft reported to police . .
not reported to police . .
do not know . .

Total .
-------------------------------------------------------------------------
Household Burglary reported to police 54.5535 1.5082
not reported to police 44.6500 1.4979
do not know 0.7964 0.4070

Total 100.000
-------------------------------------------------------------------------
Motor Vehicle Theft reported to police 83.7377 1.4000
not reported to police 15.9900 1.3898
do not know 0.2723 0.1922

Total 100.000
-------------------------------------------------------------------------
Theft reported to police 31.1678 0.4962
not reported to police 68.1654 0.4969
do not know 0.6668 0.0674
Total 100.000
-------------------------------------------------------------------------
Total reported to police
not reported to police
do not know
Total
-------------------------------------------------------------------------

25
C.3 Victimization Rates
Victimization rates are calculated from the modified person-level file (for personal
crimes) or the modified household-level file (for property crimes) using the SAS surveymeans
procedure. Since the victimization rate is a function of a constant and two variables, the analysis
variable is created as a recode in a data step preceding the procedure. The victimization count for
the victimization type of interest is multiplied by the victimization adjustment factor
(ADJINC_WT), and this product is multiplied by 1,000. This new variable is used as the analysis
variable in surveymeans’ var statement, as demonstrated in the examples below.

To calculate victimization rates for person- or household-level characteristics, these

characteristics are included on the domain statement (in addition to the derived “sub” variable
that limits the analysis to records of interest). However, if victimization rates are desired based
on incident-level characteristics, separate variables are defined for each level of the incident
characteristic of interest and are included on the var statement. Exclusions based on the incident
characteristics must be made when calculating victimization summaries, as outlined in
Section 3.2. For example, the modified person- and household-level files exclude victimizations
occurring outside of the United States. Because there are no dummy records on the modified
person and household files, no exclusions are needed to remove dummy records from the
analysis.

The modified person and household files contain the victimization counts needed to
calculate victimization rates for the most common victimization characteristics analyzed using
NCVS data. Examples 5A-5C demonstrate the calculation of personal and property
victimizations rates that can be calculated directly from the modified person and household level
files provided for a single year and Examples 6A-6B demonstrate these calculations for pooled
year estimates. If an analyst wants to calculate a victimization rate for an incident-level
characteristic that is not included on the modified files, preprocessing steps are needed to
calculate victimization summaries from the incident-level file and move these summaries to the
person file (for personal crimes) or the household file (for property crimes). Section 3.2
documents these steps. Examples 7A-7B demonstrate the calculation of personal and property
victimization rates for incident characteristics not included on the modified person and
household files.

26
Example 5: Rate of crime, single year
Example 5A is an estimate of an overall personal victimization rate. Example 5B is an
estimate of a property victimization rate with the inclusion of household-level characteristics.
Example 5C is an estimate of a personal victimization rate, where the characteristics of interest
are incident-level characteristics. All three estimates are computed with variables available on
the modified person and household files.

Example 5A - Rate of simple assaults, 2011

Year(s): 2011
Crime Type(s): simple assault (sast)
Analysis Classes: n/a
Weight: wgtpercy
Subpopulation: 2011 (year=2011)
Calculated Directly from Adjusted Files?: yes

Code:

data anset;
set adjper;
sub=(year=2011);
vrsast=adjinc_wt*sast*1000; *a;
run;

proc surveymeans data=anset varmethod=taylor mean; *b;

domain sub; *c;
strata yr_grp v2117; *d;
cluster v2118; *e;
weight wgtpercy; *f;
var vrsast;
run;

Code Comment(s):
a) Specifies victimization recode
b) Specifies input dataset and sample design
c) Subpopulation indicator
d) Specifies stratification
e) Specifies primary sampling units
f) Specifies analysis weight

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

27
The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 814680
Number of Observations Used 708493
Number of Obs with Nonpositive Weights 106187
Sum of Weights 1270197175

Statistics

Std Error
Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
vrsast 15.638820 0.610515
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub

Std Error
sub Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 vrsast 15.707917 0.669826
1 vrsast 15.367133 1.176853
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

28
Example 5B - Rate of household burglary by MSA status, household income, and family
structure, 2011
Year(s): 2011
Crime Type(s): household burglary (hburg)
Analysis Classes: MSA status (msa); household income (hincome); family structure
(fam_structure2)
Weight: wgthhcy
Subpopulation: 2011 (year=2011)
Calculated Directly from Adjusted Files?: yes

Code:

data anset;
set adjhh;
sub=(year=2011);
vrburg=adjinc_wt*hburg*1000;
run;

proc surveymeans data=anset varmethod=taylor mean;

domain sub*msa sub*hincome sub*fam_structure2;
strata yr_grp v2117;
cluster v2118;
weight wgthhcy;
var vrburg;
run;

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 527673
Number of Observations Used 398238
Number of Obs with Nonpositive Weights 129435
Sum of Weights 608895975

Statistics

Std Error
Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
vrburg 28.319646 0.771345
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
Domain Analysis: sub*HOUSEHOLD LOCALE

HOUSEHOLD Std Error

sub LOCALE Variable Mean of Mean

29
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Urban vrburg 33.388469 1.179397
Suburban vrburg 22.284276 1.161647
Rural vrburg 35.072801 2.724182
1 Urban vrburg 33.487009 2.515425
Suburban vrburg 25.445345 2.008611
Rural vrburg 33.019059 4.534446
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*hincome

Std Error
sub hincome Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Less than $7,500 vrburg 59.651675 5.784041
$7,500 to $14,999 vrburg 56.887617 4.760163
$15,000 to $24,999 vrburg 36.988937 3.625684
$25,000 to $34,999 vrburg 32.461893 2.317476
$35,000 to $49,999 vrburg 28.371919 1.752191
$50,000 to $74,999 vrburg 21.164290 1.171311
$75,000 or more vrburg 16.811149 0.868502
Unknown vrburg 25.476500 1.281015
1 Less than $7,500 vrburg 67.164972 14.006749
$7,500 to $14,999 vrburg 58.845099 9.944788
$15,000 to $24,999 vrburg 42.776246 5.556782
$25,000 to $34,999 vrburg 33.340142 5.830933
$35,000 to $49,999 vrburg 26.301166 2.682444
$50,000 to $74,999 vrburg 21.506342 3.006216
$75,000 or more vrburg 12.288703 1.235878
Unknown vrburg 29.549576 2.702723
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*Family Structure

Std Error
sub Family Structure Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Two or more adults - W/O Children vrburg 21.290559 0.930596
Two or more adults - With Children vrburg 28.441133 1.059043
One Male Adult - W/O Children vrburg 36.163618 3.603260
One Male Adult - With Children vrburg 52.526591 9.167820
One Female Adult - W/O Children vrburg 25.750941 1.588754
One Female Adult - With Children vrburg 64.961684 4.899564
1 Two or more adults - W/O Children vrburg 24.208885 2.304288
Two or more adults - With Children vrburg 29.610024 2.381918
One Male Adult - W/O Children vrburg 30.997012 2.922347
One Male Adult - With Children vrburg 46.669715 14.319366
One Female Adult - W/O Children vrburg 30.105671 4.797870
One Female Adult - With Children vrburg 61.962753 8.315094
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

30
Example 5C - Rate of violent crime by weapon involvement, injury, and victim-offender
relationship, 2011
Year(s): 2011
Crime Type(s): violent victimizations (rsa+rob+aast+sast)
Analysis Classes: weapon involvement (wpnuse); injury (inj), victim-offender relationship (rel)
Weight: wgtpercy
Subpopulation: 2011 (year=2011)
Calculated Directly from Adjusted Files?: yes

Code:

data anset;
set adjper;

sub=(year=2011);

*aggregate violent crimes into a single measure for each outcome of interest.
*a;
viol_weap1=(rsa_wpnuse1+rob_wpnuse1+aast_wpnuse1+sast_wpnuse1)*adjinc_wt*1000;
*b;
viol_weap2=(rsa_wpnuse2+rob_wpnuse2+aast_wpnuse2+sast_wpnuse2)*adjinc_wt*1000;
*c;
viol_weap3=(rsa_wpnuse3+rob_wpnuse3+aast_wpnuse3+sast_wpnuse3)*adjinc_wt*1000;

*d;
viol_inj1=(rsa_inj1+rob_inj1+aast_inj1+sast_inj1)*adjinc_wt*1000;
*e;
viol_inj2=(rsa_inj2+rob_inj2+aast_inj2+sast_inj2)*adjinc_wt*1000;
*f;
viol_inj3=(rsa_inj3+rob_inj3+aast_inj3+sast_inj3)*adjinc_wt*1000;
*g;
viol_inj4=(rsa_inj4+rob_inj4+aast_inj4+sast_inj4)*adjinc_wt*1000;

*h;
viol_rel1=(rsa_rel1+rob_rel1+aast_rel1+sast_rel1)*adjinc_wt*1000;
*i;
viol_rel2=(rsa_rel2+rob_rel2+aast_rel2+sast_rel2)*adjinc_wt*1000;
*j;
viol_rel3=(rsa_rel3+rob_rel3+aast_rel3+sast_rel3)*adjinc_wt*1000;
*k;
viol_rel4=(rsa_rel4+rob_rel4+aast_rel4+sast_rel4)*adjinc_wt*1000;
*l;
viol_rel5=(rsa_rel5+rob_rel5+aast_rel5+sast_rel5)*adjinc_wt*1000;
*m;
viol_rel6=(rsa_rel6+rob_rel6+aast_rel6+sast_rel6)*adjinc_wt*1000;

run;

proc surveymeans data=anset varmethod=taylor mean;

domain sub;
strata yr_grp v2117;
cluster v2118;
weight wgtpercy;
var viol_weap1-viol_weap3 viol_inj1-viol_inj4 viol_rel1-viol_rel6;
run;

31
Code Comment(s):
a) Yes, offender had a weapon
b) No, offender did not have a weapon
c) Do not know if offender had a weapon
d) No injury
e) Serious injury
f) Minor injury
g) Rape without other injuries
h) Intimates
i) Other relatives
j) well known/casual acquaintances
k) Strangers
l) Do not know relationship
m) Do not know number of offenders

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 814680
Number of Observations Used 708493
Number of Obs with Nonpositive Weights 106187
Sum of Weights 1270197175

Statistics

Std Error
Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
viol_weap1 4.831854 0.239978
viol_weap2 16.905851 0.657311
viol_weap3 1.585757 0.126202
viol_inj1 17.472264 0.660110
viol_inj2 1.019748 0.101134
viol_inj3 4.522141 0.241631
viol_inj4 0.294720 0.053443
viol_rel1 3.679001 0.235064
viol_rel2 1.700361 0.168869
viol_rel3 6.996892 0.373455
viol_rel4 8.648548 0.364130
viol_rel5 1.324181 0.088132
viol_rel6 0.968963 0.106546
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

32
 Domain Analysis: sub

Std Error
sub Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 viol_weap1 4.881214 0.279946
viol_weap2 17.021172 0.732904
viol_weap3 1.612903 0.135616
viol_inj1 17.608513 0.731000
viol_inj2 0.991732 0.110364
viol_inj3 4.633226 0.274660
viol_inj4 0.263517 0.065760
viol_rel1 3.774519 0.255819
viol_rel2 1.634968 0.188501
viol_rel3 7.018989 0.395294
viol_rel4 8.728594 0.410245
viol_rel5 1.423006 0.109876
viol_rel6 0.928292 0.124775
1 viol_weap1 4.637771 0.348181
viol_weap2 16.452410 1.176199
viol_weap3 1.479021 0.300596
viol_inj1 16.936533 1.225512
viol_inj2 1.129908 0.191335
viol_inj3 4.085352 0.384823
viol_inj4 0.417409 0.097062
viol_rel1 3.303427 0.499653
viol_rel2 1.957484 0.354815
viol_rel3 6.910004 0.777927
viol_rel4 8.333804 0.686701
viol_rel5 0.935598 0.125684
viol_rel6 1.128885 0.215224
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

33
Example 6: Rate of crime, aggregated years
Example 6A is an estimate of a personal victimization rate with person-level
characteristics of interest. Example 6B is an estimate of a property victimization rate with the
inclusion of a household-level characteristic. Both estimates are computed with variables
available on the modified person and household files.

Example 6A - Rate of violent crimes reported to police by sex, age category, race/ethnicity,
and marital status, 2009–2011
Year(s): 2009-2011
Crime Type(s): violent victimizations reported to police
(rsa_rpt1+rob_rpt1+aast_rpt1+sast_rpt1)
Analysis Classes: sex (sex); age category (agecat); race/ethnicity (race_eth); marital status
(marital)
Weight: wgtpercy
Subpopulation: 2009-2011 ((2009 <= year) and (year <=2011))
Calculated Directly from Adjusted Files?: yes

Code:

data anset;
set adjper;
sub=((2009 <= year) and (year <=2011));
*aggregate violent crimes into a single measure for outcome of
interest;
viol_rpt1=(rsa_rpt1+rob_rpt1+aast_rpt1+sast_rpt1)*adjinc_wt*1000;
run;

proc surveymeans data=anset varmethod=taylor mean;

domain sub*sex sub*agecat sub*race_eth sub*marital;
strata yr_grp v2117;
cluster v2118;
weight wgtpercy;
var viol_rpt1;
run;

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

34
The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 814680
Number of Observations Used 708493
Number of Obs with Nonpositive Weights 106187
Sum of Weights 1270197175

Statistics

Std Error
Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
viol_rpt1 10.942564 0.388276
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*sex

Std Error
sub sex Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Male viol_rpt1 12.263099 0.847527
Female viol_rpt1 11.735712 0.792736
1 Male viol_rpt1 9.682951 0.535355
Female viol_rpt1 10.801333 0.750092
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*agecat

Std Error
sub agecat Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 12 to 14 viol_rpt1 10.706604 1.514160
15 to 17 viol_rpt1 20.325239 2.983613
18 to 20 viol_rpt1 18.508326 2.103361
21 to 24 viol_rpt1 19.158965 2.595491
25 to 34 viol_rpt1 17.998077 2.201344
35 to 49 viol_rpt1 13.197065 1.221089
50 to 64 viol_rpt1 7.657778 0.944521
65 or older viol_rpt1 1.662752 0.359343
1 12 to 14 viol_rpt1 10.142548 1.528288
15 to 17 viol_rpt1 13.041282 1.607437
18 to 20 viol_rpt1 22.576998 4.639424
21 to 24 viol_rpt1 16.933566 1.895897
25 to 34 viol_rpt1 14.075923 1.173494
35 to 49 viol_rpt1 11.413987 0.893979
50 to 64 viol_rpt1 6.593241 0.582309
65 or older viol_rpt1 1.902745 0.371955
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*race_eth

Std Error
sub race_eth Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Non-Hispanic White viol_rpt1 11.779361 0.689952
Non-Hispanic Black viol_rpt1 17.080468 2.129403
Hispanic viol_rpt1 9.606242 1.012843
Non-Hispanic Other viol_rpt1 8.094436 2.965192
Non-Hispanic More than One Race viol_rpt1 20.653312 4.002124
1 Non-Hispanic White viol_rpt1 9.591058 0.556192
Non-Hispanic Black viol_rpt1 16.324494 2.150383

35
 Hispanic viol_rpt1 9.110811 0.905751
Non-Hispanic Other viol_rpt1 6.215577 1.431772
Non-Hispanic More than One Race viol_rpt1 20.785537 3.675282
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*marital

Std Error
sub marital Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Never Married viol_rpt1 18.209104 1.219181
Married viol_rpt1 6.064960 0.536853
Widowed viol_rpt1 5.630742 2.102283
Divorced viol_rpt1 21.873231 2.587565
Separated viol_rpt1 37.178780 7.171082
1 Never Married viol_rpt1 15.248824 1.034491
Married viol_rpt1 5.035802 0.383114
Widowed viol_rpt1 3.738650 0.916177
Divorced viol_rpt1 18.372702 1.868214
Separated viol_rpt1 38.603385 5.885568
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

36
Example 6B - Rate of property crimes reported to police by household income, 2009–2011
Year(s): 2009-2011
Crime Type(s): property crimes reported to police (hburg_rpt1+tft_rpt1+mvtft_rpt1)
Analysis Classes: household income (hincome)
Weight: wgthhcy
Subpopulation: 2009-2011 ((2009 <= year) and (year <=2011))
Calculated Directly from Adjusted Files?: yes

Code:

data anset;
set adjhh;
sub=((2009 <= year) and (year <=2011));
vrproprpt=(hburg_rpt1+tft_rpt1+mvtft_rpt1)*adjinc_wt*1000;
run;

proc surveymeans data=anset varmethod=taylor mean;

domain sub*hincome;
strata yr_grp v2117;
cluster v2118;
weight wgthhcy;
var vrproprpt;
run;

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 527673
Number of Observations Used 398238
Number of Obs with Nonpositive Weights 129435
Sum of Weights 608895975

Statistics

Std Error
Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
vrproprpt 53.182550 0.936036
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*hincome

Std Error
sub hincome Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Less than $7,500 vrproprpt 81.044611 8.658194

37
 $7,500 to $14,999 vrproprpt 74.357555 6.668412
$15,000 to $24,999 vrproprpt 60.660916 3.555991
$25,000 to $34,999 vrproprpt 58.303682 4.122739
$35,000 to $49,999 vrproprpt 60.220186 3.250153
$50,000 to $74,999 vrproprpt 56.186640 3.568633
$75,000 or more vrproprpt 58.226671 2.115554
Unknown vrproprpt 49.863632 1.997445
1 Less than $7,500 vrproprpt 64.052739 4.366416
$7,500 to $14,999 vrproprpt 60.840821 3.678508
$15,000 to $24,999 vrproprpt 54.831987 2.794859
$25,000 to $34,999 vrproprpt 54.832067 3.222495
$35,000 to $49,999 vrproprpt 55.921312 3.030289
$50,000 to $74,999 vrproprpt 49.011134 2.029558
$75,000 or more vrproprpt 49.946119 1.490552
Unknown vrproprpt 43.616987 1.544342
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

38
Example 7: Computing victimization rates based on incident characteristics not included on
the provided file
The two examples below estimate personal and property victimization rates for incident-
level characteristics not contained on the modified person and household files, and thus require
the pre-processing steps outlined in Section 3.2. SAS is used to calculate victimization
summaries from the incident-level file and merge incident summaries onto the household and
person files. Victimization rates are then calculated from the modified person-level file (for
personal crimes) or the modified household-level file (for property crimes). Example 7A is an
estimate of a personal victimization rate and Example 7B is an estimate of a property
victimization rate. Example 7A is for a single year, and Example 7B is for aggregated years.

Example 7A - Rate of violent crimes by location of crime, 2011

Year(s): 2011
Crime Type(s): violent victimizations by location of crime (violent_home; violent_other)
Analysis Classes: n/a
Weight: wgtpercy
Subpopulation: 2011 (year=2011)
Calculated Directly from Adjusted Files?: no

Code:
*step1: identify records with victimization characteristic(s) of
interest on the modified incident-level file ;

data example7a_inc;
set adjincident;
vcrime=(newoff in (1 2 3 4)); *a;

*identify location of crime;

if v4024 in (1 2 3 4) then place_inc=1; *b;
else if v4024 in (5 6 7) then place_inc=2; *c;
else if v4024 in (8 9 10 11) then place_inc=3; *d;
else if v4024 in (12 13 14 24 25 26 27) then place_inc=4; *e;
else if v4024 in (15 16 17) then place_inc=5; *f;
else if v4024 in (18 19) then place_inc=6; *g;
else if v4024 in (20 21 22) then place_inc=7; *h;
else if v4024 in (23) then place_inc=8; *i;
else if v4024=. then place_inc=9; *j;

*collapse location of crime;

if place_inc in (1 2) then place=1; *k;
else place=2; *l;

39
*create binary classifications for each location, excluding
dummy records and crimes that occurred outside of the united
states;

violent_home=(vcrime=1 and place=1 and exclude_outus=0 and

dummy=0);
violent_other=(vcrime=1 and place=2 and exclude_outus=0 and
dummy=0);
run;

Code Comment(s):
a) Identify violent crimes
b) Respondent’s home
c) Near respondents home
d) Friend’s home
e) Commercial
f) Parking lot-garage
g) School
h) Open areas, on street or public transportation
i) Other
j) Missing
k) Respondent’s home/near respondent’s home
l) Other location

*step2: create victimization summary file ;

proc sort data=example7a_inc(keep=yearq idper serieswgt
violent_home violent_other);
by yearq idhh idper;
run;

*calculate the number of victimizations of each type by person

and reporting period ;
proc means data=example7a_inc noprint;
by yearq idper;
weight serieswgt;
var violent_home violent_other;
output out=violent_place_sums(drop=_type_ _freq_)
sum=;
run;
*step3: merge victimization summary file onto the person-level
file ;
proc sort data=adjper; by yearq idper; run;

data example7a_per;
merge adjper(in=a)
violent_place_sums(in=b);
by yearq idper;
if a;
if violent_home=. then violent_home=0; *a;
if violent_other=. then violent_other=0;
run;

40
Code Comment(s):
a) The two incident count variables are missing for persons not
included on the victimization summary file, so they are set to ‘0’ (no
victimizations of this type)

*calculate victimization rates ;

proc sort data=example7a_per; by yr_grp v2117 v2118; run;

data anset;
set example7a_per;
sub=(year=2011);
*a;
vrviolrhome=violent_home*adjinc_wt*1000;
vrviolother=violent_other*adjinc_wt*1000;
run;

proc surveymeans data=anset varmethod=taylor mean;

domain sub;
strata yr_grp v2117;
cluster v2118;
weight wgtpercy;
var vrviolrhome vrviolother;
run;

Code Comment(s):
a) The final step outlined in section 3.2 was to calculate the
victimization adjustment factor, but it is already contained on
the modified person file.

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 814680
Number of Observations Used 708493
Number of Obs with Nonpositive Weights 106187
Sum of Weights 1270197175

Statistics

Std Error
Variable Mean of Mean

41
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
vrviolrhome 8.895617 0.371408
vrviolother 14.427845 0.557310
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub

Std Error
sub Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 vrviolrhome 9.038327 0.425064
vrviolother 14.476962 0.608688
1 vrviolrhome 8.334482 0.631375
vrviolother 14.234720 1.153994
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

42
Example 7B - Rate of property crimes by time of day, 2009-2011
Year(s): 2009-2011
Crime Type(s): property crimes by time of day (property_day; property_night;
property_missing)
Analysis Classes: n/a
Weight: wgthhcy
Subpopulation: 2009-2011 ((2009 <= year) and (year <=2011))
Calculated Directly from Adjusted Files?: no

*step1: identify records with victimization characteristic(s) of

interest on the modified incident-level file ;
data example7b_inc;
set adjincident;
pcrime=(newoff in (6 7 8)); *a;
*identify time of incident;
if v4021b in (1 2 3 4) then time_day=1; *b;
else if v4021b in (5 6 7 8) then time_day=2; *c;
else if v4021b in (9 98) then time_day=3; *d;
*create binary classifications for each time, excluding dummy
records and crimes that occurred outside of the united states;
property_day=(pcrime=1 and time_day=1 and exclude_outus=0 and
dummy=0);
property_night=(pcrime=1 and time_day=2 and exclude_outus=0 and
dummy=0);
property_missing=(pcrime=1 and time_day=3 and exclude_outus=0
and dummy=0);

run;

Code Comment(s):
a) identify property crimes
b) day
c) night
d) missing

*step2: create victimization summary file ;

proc sort data=example7b_inc(keep=yearq idhh serieswgt
property_day property_night property_missing);
by yearq idhh;
run;

43
*calculate the number of victimizations of each type by household and
reporting period ;
proc means data=example7b_inc noprint;
by yearq idhh;
weight serieswgt;
var property_day property_night property_missing;
output out=property_time_sums(drop=_type_ _freq_)
sum=;
run;

*step3: merge victimization summary file onto the HH-level

file ;
proc sort data=adjhh; by yearq idhh; run;
data example7b_hh;
merge adjhh(in=a)
property_time_sums(in=b);
by yearq idhh;
if a;
if property_day=. then property_day=0; *a;
if property_night=. then property_night=0;
if property_missing=. then property_missing=0;
run;

Code Comment(s):
a) The three incident count variables are missing for households not
included on the victimization summary file, so they are set to ‘0’ (no
victimizations of this type)

*calculate victimization rates ;

proc sort data=example7b_hh; by yr_grp v2117 v2118; run;

data anset;
set example7b_hh;
sub=((2009 <= year) and (year <=2011));
*a;
vrpropday=property_day*adjinc_wt*1000;
vrpropnight=property_night*adjinc_wt*1000;
vrpropmiss=property_missing*adjinc_wt*1000;
run;

proc surveymeans data=anset varmethod=taylor mean;

domain sub;
strata yr_grp v2117;
cluster v2118;
weight wgthhcy;
var vrpropday vrpropnight vrpropmiss;
run;

44
Code Comment(s):
a) The final step outlined in section 3.2 was to calculate the
victimization adjustment factor, but it is already contained on
the modified person file.

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 527673
Number of Observations Used 398238
Number of Obs with Nonpositive Weights 129435
Sum of Weights 608895975

Statistics
Std Error
Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
vrpropday 54.546013 1.097167
vrpropnight 59.332574 1.222059
vrpropmiss 24.883746 0.735171
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub

Std Error
sub Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 vrpropday 59.022848 1.518622
vrpropnight 63.701678 1.688110
vrpropmiss 26.018035 1.020754
1 vrpropday 51.620495 1.227080
vrpropnight 56.477455 1.331527
vrpropmiss 24.142513 0.806898
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

45
C.4 Identifying Low Quality Estimates
This section demonstrates how to implement the recommendations for identifying low
quality estimates and the rounding rules discussed in Section 4 of the user’s guide. Three types
of data are needed to identify and flag low quality estimates, regardless of the estimate type: the
estimate, the standard error of the estimate, and the unweighted sample size of the estimate.
Because Examples 1-7 provided details for obtaining estimates and standard errors, the examples
in this section will focus primarily on calculating the percent relative standard error (RSE),
calculating unweighted sample sizes, identifying estimates that should be flagged as unreliable,
and verifying that estimates rounding to zero are not presented.

Unweighted sample sizes should take into account the series adjustment. For example, a
series victimization with a series count of seven would count as seven victimizations in the
unweighted sample size, while a non-series victimization would only count as one. This series
adjustment is already incorporated in the victimization counts on the modified person and
household-level files, so sample sizes for victimization rates are obtained by taking unweighted
sums of victimization counts with the specified characteristic(s) of interest. However, for
victimization totals and proportions, the series count must be included in the calculation of the
sample size, as demonstrated in the examples below. Example 8A demonstrates implementation
of flagging and rounding rules for victimization totals, Example 8B for victimization proportions,
and Example 8C for victimization rates.

Example 8A - Total number of personal thefts by sex and race/ethnicity, 2011

Year(s): 2011
Crime Type: personal thefts (ptft)
Domain(s): sex (sex) and race/ethnicity (race_eth)
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2011 (year=2011)
Code:

*calculate estimates and standard errors;

data anset;
set adjincident;
sub=(exclude_outus=0 and dummy=0 and year=2011);
run;

46
proc surveymeans data=anset varmethod=taylor sum;
domain sub*sex sub*race_eth;
strata yr_grp v2117;
cluster v2118;
weight newwgt;
var ptft;
run;

*calculate sample sizes;

proc means data=anset sum;
where sub;
class sex race_eth;
ways 1;
weight serieswgt; *a;
var ptft;
run;

Code Comment(s):
a) Weight by the series weight to obtain the series-adjusted count
of victimizations

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

The SURVEYMEANS Procedure

Data Summary

Number of Strata 304

Number of Clusters 608
Number of Observations 112932
Sum of Weights 342486009

Statistics

Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
PTFT 2787632 119047
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*sex

sub sex Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Male PTFT 1244808 92801
Female PTFT 1377054 72295
1 Male PTFT 53583 18914
Female PTFT 112187 19314
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
Domain Analysis: sub*race_eth

sub race_eth Variable Sum Std Dev

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Non-Hispanic White PTFT 1615705 86826

47
 Non-Hispanic Black PTFT 435124 52924
Hispanic PTFT 371336 44117
Non-Hispanic Other PTFT 172088 39824
Non-Hispanic More than One Race PTFT 27608 9540.767800
1 Non-Hispanic White PTFT 83956 21479
Non-Hispanic Black PTFT 29507 10989
Hispanic PTFT 30287 10489
Non-Hispanic Other PTFT 13519 5943.552122
Non-Hispanic More than One Race PTFT 8501.872300 5281.778471
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Analysis Variable : PTFT

race_eth N Obs Sum

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
Non-Hispanic White 4660 24.0000000

Non-Hispanic Black 911 9.0000000

Hispanic 1183 9.0000000

Non-Hispanic Other 315 6.0000000

Non-Hispanic More than One Race 186 3.0000000

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
Analysis Variable : PTFT

sex N Obs Sum

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
Male 3358 16.0000000

Female 3897 35.0000000

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Identifying Unreliable Estimates:

As outlined in Section 4, it is recommended that estimated totals meeting either of the following
criteria be identified as unreliable. In addition, any estimate rounding to zero should not be
displayed.

 RSE > 30%

 Count sample size ≤ 10

The percent relative standard error of an estimated total is calculated as:

𝑆𝑆𝑆𝑆(𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇)
% 𝑅𝑅𝑅𝑅𝑅𝑅(𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇) = ∗ 100
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇
Estimated totals, standard errors, and unweighted sample sizes from the output above are
included in Table C-1 (in the “Total,” “SE(Total),” and “n” columns, respectively). Percent
RSEs are calculated based on the formula above. Based on the recommendations, five of the
eight estimates are flagged as unreliable. The estimate for males is flagged because the percent
RSE is greater than 30 percent, while the other four estimates are flagged both because their

48
RSEs exceed 30 percent and because their unweighted sample sizes are less than or equal to 10.
All estimates can be displayed because no estimates round to zero.

Table C-1. Identifying Low Quality Victimization Totals

%RSE Flag as Rounds to
Domain Total SE(Total) n (Total) Unreliable? Zero?
Overall 165770.15 24276.55 51 14.64
Male 53583.48 18913.90 16 35.30 √
Female 112186.67 19313.73 35 17.22
Non-Hispanic White 83955.56 21478.97 24 25.58
Non-Hispanic Black 29506.98 10989.44 9 37.24 √
Hispanic 30286.66 10488.60 9 34.63 √
Non-Hispanic Other 13519.07 5943.55 6 43.96 √
Non-Hispanic More than One 8501.87 5281.78 3 62.12 √
Race

49
Example 8B – Distribution of motor vehicle theft in the Northeast Census Region by report
status, 2011
Year(s): 2011
Domain(s): Northeast Census Region (region=1); motor vehicle theft (mvtft=1)
Covariate(s): report status (notify)
Weight: newwgt
Subpopulation: within the United States (exclude_outUS=0); non-dummy records (dummy=0);
2011 (year=2011); Northeast Census Region (region=1); motor vehicle theft (mvtft=1)

Code:
*calculate estimates and standard errors;
data anset;
set adjincident;
sub=(exclude_outus=0 and dummy=0 and year=2011 and region=1 and
mvtft=1);
run;

proc surveyfreq data=anset varmethod=taylor;

strata yr_grp v2117;
cluster v2118;
weight newwgt;
table sub*notify / row nocellpercent nofreq nowt;
run;

*calculate sample sizes;

proc means data=anset n;
where sub;
weight serieswgt; *a;
class notify;
var notify;
run;

Code Comment(s):
a) Weight by the series weight to obtain the series-adjusted count
of victimizations. Because flagging rules are based on the
denominator sample size, only the overall sample size is needed
(not estimates by report status).

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

50
The SURVEYFREQ Procedure
Data Summary

Number of Strata 304

Number of Clusters 608
Number of Observations 112932
Sum of Weights 342486009

Table of sub by notify

Row Std Err of

sub notify Percent Row Percent
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 reported to police 40.0671 0.3079
not reported to police 58.9204 0.3136
do not know 1.0125 0.0478

Total 100.000
---------------------------------------------------------
1 reported to police 92.6011 5.4032
not reported to police 7.3989 5.4032
do not know . .

Total 100.000
---------------------------------------------------------
Total reported to police
not reported to police
do not know

Total
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
Frequency Missing = 4079

The MEANS Procedure

Analysis Variable : notify

notify N Obs N
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
reported to police 17 17

not reported to police 2 2

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Identifying Unreliable Estimates:

As outlined in Section 4, it is recommended that victimization percentages (P) meeting
either of the following criteria be flagged as unreliable. In addition, any estimate rounding to
zero should not be displayed.

51
  RSE > 30%, where

– If the percentage is ≤ 50%, use the RSE of log(𝑃𝑃)

– If the percentage is > 50%, use the RSE of log(100-𝑃𝑃); or
 Denominator sample size ≤ 10

The percent relative standard errors of the log of the percentages (P) and (1-P) are
calculated as:

𝑆𝑆𝑆𝑆(𝑃𝑃)
% 𝑅𝑅𝑅𝑅𝑅𝑅(log(𝑃𝑃)) = ∗ 100
𝑃𝑃
𝑃𝑃 ∗ 𝑎𝑎𝑎𝑎𝑎𝑎(log �100�)

𝑆𝑆𝑆𝑆(𝑃𝑃)
% 𝑅𝑅𝑅𝑅𝑅𝑅(log(100 − 𝑃𝑃)) = ∗ 100
𝑃𝑃
(100 − 𝑃𝑃) ∗ 𝑎𝑎𝑎𝑎𝑎𝑎(log �1 − 100�)

Estimated percentages, standard errors, and the unweighted denominator sample size
from the output above are included in Table C-2 below (in the “P,” “SE(P),” and “n (denom)”
columns, respectively). Percent RSEs are calculated based on the formulas above. Based on these
recommendations, none of the estimates are flagged as unreliable. The denominator sample size
is 19, which is greater than the recommended threshold of 10. The percent RSEs for the reported
and not reported estimates are 28.03, which are below the 30 percent threshold. However,
because no respondents reported that they did not know whether or not the motor vehicle theft
was reported to the police, the estimated percent is 0. This estimate should not be presented
because it rounds to zero.

Table C-2. Identifying Low Quality Victimization Percentages

%RSE(P) OR Rounds to
Estimate P SE(P) n (denom) %RSE(100-P) Flag as Unreliable? Zero?
Reported to Police 92.60 5.40 19 28.03
Not Reported to Police 7.40 5.40 19 28.03
Do Not Know 0.00 - 19 - √

52
Example 8C – Rate of rape/sexual assault by sex, 2011
Year(s): 2011
Crime Type(s): rape/sexual assault (rsa)
Domain(s): sex (sex)
Weight: wgtpercy
Subpopulation: 2011 (year=2011)
Calculated Directly from Adjusted Files?: yes

Code:

*calculate estimates and standard errors;

data anset;
set adjper;
sub=(year=2011);
vrrsa=adjinc_wt*rsa*1000;
run;

proc surveymeans data=anset varmethod=taylor mean;

domain sub sub*sex;
strata yr_grp v2117;
cluster v2118;
weight wgtpercy;
var vrrsa;
run;

*calculate sample sizes;

proc means data=anset sum; *a;
where sub;
class sex;
var rsa;
run;

Code Comment(s):
b) Unweighted because the victimization count RSA already includes
the series adjustment

Output:

NOTE: Output is provided overall and by subpopulation. Only the output

under “Domain Analysis: sub” with a “sub” value of 1 is relevant
(highlighted row).

53
The SURVEYMEANS Procedure
Data Summary

Number of Strata 160

Number of Clusters 320
Number of Observations 814680
Number of Observations Used 708493
Number of Obs with Nonpositive Weights 106187
Sum of Weights 1270197175

Statistics

Std Error
Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
vrrsa 1.115027 0.141041
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub

Std Error
sub Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 vrrsa 1.157469 0.166403
1 vrrsa 0.948146 0.162529
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Domain Analysis: sub*sex

Std Error
sub sex Variable Mean of Mean
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
0 Male vrrsa 0.173447 0.043650
Female vrrsa 2.095539 0.323622
1 Male vrrsa 0.276019 0.089702
Female vrrsa 1.592869 0.332343
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

The MEANS Procedure

Analysis Variable : RSA

sex N Obs Sum

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
Male 77758 10.0000000

Female 85109 69.0000000

Ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

Identifying Unreliable Estimates:

As outlined in Section 4, it is recommended that victimization rates meeting either of the
following criteria be flagged as unreliable. In addition, any estimate rounding to zero should not
be displayed.

54
  RSE > 30%

 Numerator sample size ≤ 10

The percent relative standard errors of the victimization rates (VR) are calculated as:

𝑆𝑆𝑆𝑆(𝑉𝑉𝑉𝑉)
% 𝑅𝑅𝑅𝑅𝑅𝑅(𝑉𝑉𝑉𝑉) = ∗ 100
𝑉𝑉𝑉𝑉
Estimates, standard errors, and unweighted sample sizes from the output above are
included in Table C-3 below (in the “VR,” “SE(VR),” and “n (num)” columns, respectively).
Percent RSEs are calculated based on the formula above. Based on these recommendations, the
estimated victimization rate for males is flagged as unreliable because the percent RSE is greater
than 30 percent and the sample size of the numerator is less than or equal to 10. Because none of
the estimates round to zero, all three estimates can be presented.

Table C-3. Identifying Low Quality Victimization Rates

Flag as Rounds to
Domain VR SE(VR) n (num) %RSE(VR) Unreliable? Zero?
Overall 0.94815 0.16253 79 17.14
Male 0.27602 0.0897 10 32.50 √
Female 1.59287 0.33234 69 20.86

55