Literature Review

August 13, 2009

Literature Reviews:
Examination of Data Collection
Methods for the NCVS
Report

Prepared for
Bureau of Justice Statistics
810 7th Street, NW
Washington, DC 20531

Prepared by
RTI International
3040 Cornwallis Road
Research Triangle Park, NC 27709

RTI Project Number 0211889

RTI Project Number
0211889

Literature Reviews:
Examination of Data Collection
Methods for the NCVS
Report
August 13, 2009

Prepared for
Bureau of Justice Statistics
810 7th Street, NW
Washington, DC 20531

Prepared by
RTI International
3040 Cornwallis Road
Research Triangle Park, NC 27709

_________________________________
RTI International is a trade name of Research Triangle Institute.

Contents
Section
1.

2.

3.

4.

5.

Page

Address-Based Sampling

1-1

1.1

Summary ................................................................................................1-2

1.2

References ..............................................................................................1-3

Mixed-Mode Surveys

2-1

2.1

Attractiveness of Mixed-Mode Data Collection ..............................................2-1

2.2

Mode Effects ............................................................................................2-2

2.3

Things to Consider When Mixing Modes.......................................................2-3

2.4

Summary ................................................................................................2-4

2.5

References ..............................................................................................2-4

Self-Administered Modes of Data Collection

3-1

3.1

Mail Surveys ............................................................................................3-1

3.2

Self-Administered Modes and Sensitive Questions ........................................3-2

3.3

Web Surveys ...........................................................................................3-3

3.4

Summary ................................................................................................3-4

3.5

References ..............................................................................................3-5

Use of Incentives

4-1

4.1

Theories on Incentive Effectiveness ............................................................4-1

4.2

Impact on Response Rate and Nonresponse Bias..........................................4-2

4.3

Prepaid Versus Postpaid Incentives.............................................................4-3

4.4

Incentives and Survey Mode ......................................................................4-3

4.5

Summary ................................................................................................4-4

4.6

References ..............................................................................................4-4

Additional Issues in Measuring Crime Victimization in Surveys

5-1

5.1

Event Recall ............................................................................................5-1

5.2

Proxy Respondents ...................................................................................5-2

5.3

Crime Severity, Survey Context, Stigma, and Terminology ............................5-3

iii

Report Title

iv

5.4

Summary ................................................................................................5-4

5.5

References ..............................................................................................5-4

1. ADDRESS-BASED SAMPLING
The development and improvement of a database of addresses in the United States has
provided a potential alternative to the costly creation of sampling frames for area probability
surveys through field listing. Address-based sampling (ABS) is possible using the Delivery
Sequence File (DSF), a computerized file that contains all delivery point addresses serviced
by the U.S. Postal Service (with the exception of general delivery). So far, evaluations of
DSF for replacing enumeration of household units have shown promise, with potential
household coverage as high as 97% on average. All evaluations have shown higher
household coverage in urban areas than in rural areas.
The survey literature so far has focused on various approaches to a sampling frame
construction from an address list and evaluation of its coverage and usability properties.
The different approaches yield a uniform finding: using mailing addresses to develop a
sampling frame for metropolitan households is a good and less costly alternative to
household enumeration. For example, Iannacchione, Staab, and Redden (2003) applied
Kish’s half-open interval (Kish, 1965) frame-linking procedure to evaluate the coverage of
an ABS frame using DSF. It was estimated that half-open intervals could be constructed and
located for 94% of the addresses in the newly constructed frame. In another study designed
to compare the coverage of ABS to field enumeration, Iannacchione et al. (2007) used
global positioning system (GPS) technology to match the housing units from each frame.
Even though field enumeration yielded higher overall coverage (98% vs. 82% in ABS), there
was no difference when the matching was restricted to occupied urban housing units.
Morton et al. (2007) applied Geographic information system (GIS) and GPS technologies to
match postal (mailing address lists by postal carrier routes) to census geography (tracts and
blocks). Not surprisingly, housing units in urban areas were more likely to geocode to the
correct census block than housing units in rural areas (73% vs. 38%). O’Muircheartaigh et
al. (2006) compared the coverage and cost-benefit tradeoffs of traditional enumeration and
ABS on a national scale, employing a process in which a benchmark frame was constructed
and ABS and traditional enumeration were evaluated against it. Overall, ABS was found to
be more effective than the traditional enumeration, with the exception of areas with
irregular street patterns and high population growth rates.
A few studies present methods for improving the coverage of ABS. Dohrmann, Han, and
Mohadjer (2006) proposed enhancing the existing “Waksberg approach” to select segments
with high growth rates at higher probabilities and applying lower subsampling rates for
inclusion of missed units in such segments. O’Muircheartaigh, English, and Eckman (2007)
proposed a model-based approach to inform decisions prior to data collection on whether
field enhancement to ABS would be needed in particular segments. ABS was found
appropriate for small-scale, low-cost surveys but was seen as not yet ready to fully replace
traditional enumeration for high-quality national surveys. McMichael, Ridenhour, and

1-1

Report Title

Shook-Sa (2008) proposed an alternative to HOI—a three-component procedure called
Check for Housing Units Missed (CHUM). In initial evaluation, the first component of CHUM
picked 79% of the missing units, while the second component picked the remaining 21%.
The quality of the address lists and of their coverage varies by vendors. Various vendors
maintain and provide current versions of DSF that could be purchased for surveys (USPS
does not offer it to survey organizations). Dohrmann, Han, and Mohadjer (2006) compared
list quality by vendors (Compact Information Systems [CIS], Donnelly Marketing, and
ADVO) for an urban/suburban area and compared ABS to traditional enumeration in
urban/suburban, very urban, and rural areas. CIS and ADVO were found to be comparable.
Consistent with other findings, high match rates between ABS and traditional enumeration
were reported mainly for urban areas.
Alternatives to DSF have also been considered in investigating alternative methods for
sampling frame construction. For example, Kalsbeek, Kavanagh, and Wu (2004) examined
the utility of using lists of property tax parcels in U.S. counties. A test of the proposed
approach yielded high levels of validity and reliability, similar to the levels associated with
the traditional housing unit enumeration.
Finally, the evaluation of ABS for sampling frame creation for the general population has
been expanded by a comparison to random-digit dialing (RDD) sampling methods (Link et
al., 2008). In addition to the lower cost of the ABS mail survey, ABS reported significantly
higher response rates than RDD in five of the six studied states.

1.1

Summary

Overall, the existing research presents a promising future for ABS in survey design and
suggests that its true potential may be in mixed-mode surveys. The attractiveness of ABS is
that it is cost efficient and time efficient. Large-scale surveys often require several months
to list all dwelling units in the selected segments (usually, census blocks). In contrast, ABS
offers greater geographic diversity (selection of housing units is not restricted to small
segments based on census blocks) and thus presents a potential for improving statistical
efficiency. There are some drawbacks associated with the construction of an address-based
sampling frame related to the overall completeness of the list, the current status of the
addresses, and the adequacy of the list coverage in rural areas. The typical sources of
undercoverage for ABS are post office boxes, when used as the only method for mail
delivery (making up 1.3% of households in the United States, according to Staab and
Iannacchione [2003]); rural routes (making up 3.9% of households nationwide); and
noninstitutional group quarters (e.g., dormitories, assisted living facilities, shelters) that are
not identified on the USPS lists because they operate their own post office or because mail
is delivered to the business unit.

1-2

Section 1 — Address-Based Sampling

1.2

References

Dohrmann, Sylvia, Daifeng Han, and Leyla Mohadjer. 2006. “Residential Address Lists vs.
Traditional Listing: Enumerating Households and Group Quarters.” In Proceedings of
the American Statistical Association, Section on Survey Research Methods, pp.295964.
Iannacchione, Vincent, Katherine Morton, Joseph McMichael, David Cunningham, James
Cajka, and James Chromy. 2007. “Comparing the Coverage of a Household Sampling
Frame Based on Mailing Addresses to a Frame Based on Field Enumeration.” In
Proceedings of the American Statistical Association, Section on Survey Research
Methods, pp.3323-32.
Iannacchione, Vincent, Jennifer Staab, and David Redden. 2003. “Evaluating the Use of
Residential Mailing Lists in a Metropolitan Household Survey.” Public Opinion
Quarterly 67(2):202-10.
Kalsbeek, William, Sarah Kavanagh, and Jingjing Wu. 2004. “Using GIS-Based Property Tax
Records as an Alternative to Traditional Household Listing in Area Samples.” In
Proceedings of the American Statistical Association, Section on Survey Research
Methods, pp.3750-57.
Kish, Leslie. 1965. Survey Sampling. New York: Wiley
Link, Michael, Michael Battaglia, Martin Frankel, Larry Osborn, and Ali Mokdad. 2008. “A
Comparison of Address-Based Sampling (ABS) Versus Random-Digit Dialing (RDD)
for General Population Surveys.” Public Opinion Quarterly 72(1):6-27.
McMichael, Joseph, Jamie Ridenhour, and Bonnie Shook-Sa. 2008. “A Robust Procedure to
Supplement the Coverage of Address-Based Sampling Frames for Household
Surveys.” In Proceedings of the American Statistical Association, Section on Survey
Research Methods, pp. 4329-35.
Morton, Katherine, Vincent Iannacchione, Joseph McMichael, James Cajka, Ross Curry, and
David Cunningham. 2007. “Linking Mailing Addresses to a Household Sampling
Frame Based on Census Geography.” In Proceedings of the American Statistical
Association, Section on Survey Research Methods, pp. 3971-74.
O’Muircheartaigh, Colm, Edward English, and Stephanie Eckman. 2007. “Predicting the
Relative Quality of Alternative Sampling Frames.” In Proceedings of the American
Statistical Association, Section on Survey Research Methods, pp.3239-48.
O’Muircheartaigh, Colm, Edward English, Stephanie Eckman, Heidi Upchurch, Erika Garcia,
and James Lepkowski. 2006. “Validating a Sampling Revolution: Benchmarking
Address Lists Against Traditional Listing.” In Proceedings of the American Statistical
Association, Section on Survey Research Methods, pp. 4189-96.
Staab, Jennifer, and Vincent Iannacchione. 2003. “Evaluating the Use of Residential Mailing
Addresses in a National Household Survey.” In Proceedings of the American
Statistical Association, Section on Survey Research Methods, pp.4028-33.

1-3

2. MIXED-MODE SURVEYS
Researchers are continually trying to find the optimal mix of methods to minimize total
survey error in survey estimates. Declining response rates, increasing costs, coverage
issues, and data collection deadlines have all led to the increasing use of mixed-mode
survey designs. With the popularity of telephone surveys in the 1970s, the mix of face-toface and telephone data collection modes soon became attractive for large national surveys
(e.g., the Current Population Survey). The development of computer technology marked the
next change in data collection—computer-assisted equivalents were implemented in all
major modes of data collection (de Leeuw and Collins, 1997; Couper and Nicholls, 1998).
The development of web surveys gave rise to a combination of mail and web surveys.
When discussing mixed-mode surveys, it is important to investigate the reasons for mixing
modes, mode effects, and issues to consider when mixing modes. These items are discussed
in more detail below.

2.1

Attractiveness of Mixed-Mode Data Collection

Groves et al. (2004) identified three main reasons for using mixed-mode data collection:
cost reduction, response rate maximization, and money saving in longitudinal surveys. The
use of a combination of data collection methods reduces cost, as it typically involves an
attempt to collect data in a cheaper mode (e.g., mail), followed by a more expensive mode
(e.g., telephone), and possibly moving to an even more costly mode (e.g., face-to-face
interviewing) for the nonrespondent sample persons. The American Community Survey is an
example of this approach: it starts in a mail mode; this is followed by telephone follow-up of
nonrespondents; and then there are face-to-face follow-ups with a subsample of the
remaining nonrespondents (Alexander and Wetrogan, 2000). Maximization of response rates
is often achieved through mixed-mode data collection. For example, the Current
Employment Statistics program offers multiple modes of data collection, such as web, fax,
inbound interactive voice response (IVR), telephone, and mail. While the Current
Employment Statistics survey, which includes 390,000 business establishments, employs six
methods of data collection, the use of two or three modes is more common in increasing
response rates and decreasing costs.
Longitudinal surveys also employ mixed-mode data collection to reduce cost in later waves,
when rapport between the interviewer and the respondent has already been established in
the first wave, usually administered in face-to-face mode. An example of this approach is
the Current Population Survey, where interviewers obtain telephone numbers in the first
wave of data collection that are to be used in subsequent rounds.
Biemer and Lyberg (2003) note that mixed-mode designs have now become the norm of
data collection in the United States and Western Europe. The attractiveness of mixed-mode

2-1

Report Title

designs is in their ability to compensate for the weaknesses of individual modes. For
example, to reduce coverage bias in the early days of telephone data collection, mixedmode dual frame designs were often employed, benefiting from the cost savings of
telephone interviewing and the complete coverage of face-to-face data collection (e.g.,
Massey, Marquis, and Tortora, 1982; Marquis and Blass, 1985; for a detailed discussion, see
Groves and Lepkowski, 1985). Another feature that makes mixed-mode designs attractive is
their application in reducing nonresponse bias. Since nonresponse includes both
noncontacted respondents and those who refuse to cooperate under the initial protocol,
implementing a different mode of data collection can be addressed both by changing the
method of contact and using different persuasive techniques, particularly through the use of
interviewers. It is not necessarily that some modes are better than others for a particular
population; to the extent that individuals vary in their likelihood to participate across modes
and that respondents to different modes are somewhat different, the threat of nonresponse
bias is minimized through the use of multiple modes.
The possibility that some respondents prefer one mode over another has been recognized.
Often, however, the mode in which respondents are asked about their mode of preference is
selected as the mode of choice. For example, Groves and Kahn (1979) reported that among
respondents in a national telephone survey, 39% expressed a preference to be interviewed
by phone, 23% in a face-to-face setting, and 28% by mail. The preferred mode of interview
in a face-to-face survey was overwhelmingly face-to-face (78%), followed by mail (only
17%). Some studies suggest that giving the choice of mode to the respondent does not
necessarily improve response rates. For example, Dillman, Clark, and West (1995) showed
that offering the respondent the choice of returning a questionnaire by mail or calling in to
be interviewed did not improve response rates. On the other hand, sequential change of
modes has been reported to significantly improve response rates. For example, Shettle and
Mooney (1999) reported a response rate of 68% after four mailings and an incentive, which
increased to 81% with telephone follow-up and to 88% with a final switch to face-to-face
interviewing.

2.2

Mode Effects

Different data collection modes possess different strengths and weaknesses. In searching
for reasonable alternatives, studies have contrasted pairs of modes. Compared with face-toface surveys, telephone surveys have been found to yield lower response rates (Groves and
Kahn, 1979; Cannell et al., 1987; Sykes and Collins, 1988), shorter responses to openended questions (Groves and Kahn, 1979; Sykes and Collins, 1988; Kormendi and
Noordhoek, 1989), and higher rates of satisficing and socially desirable responding
(Holbrook, Green, and Krosnick, 2003; Kirsch, McCormack, and Saxon-Harrold, 2001). In
addition, sensitive questions have been found to increase mode differences (Aquilino and
LoSciuto, 1990). Similarly, comparisons between mail and telephone modes show higher

2-2

Section 2 — Mixed-Mode Surveys

social desirability effects (Dillman and Tarnai, 1991; Walker and Restuccia, 1984) and
increased response order and question order effects (Bishop et al., 1988) for telephone
surveys. A meta-analysis of face-to-face versus mail response rates did not find significant
differences between modes (Goyder, 1986). Research so far has produced mixed results on
the effect of these modes on reports of sensitive behaviors. For example, Bongers and van
Oers (1998) found no difference between mail and face-to-face interviewing on responses to
alcohol-related questions, but Hochstim (1967) and Tourangeau and Smith (1996) found
greater reporting of sensitive behaviors in self-administered surveys.

2.3

Things to Consider When Mixing Modes

There are potential drawbacks to using mixed-mode survey designs, affecting different
sources of survey error: coverage, nonresponse, measurement, and processing. Coverage
error can be affected in mixed-mode designs when multiple sampling frames are needed.
Although the use of multiple frames can reduce undercoverage, it involves the use of
statistical adjustments to sample weights to merge data from each mode—a procedure that
can induce varying, and often unknown, amounts of error, depending on the particular
frames and study design.
As noted earlier, mixed-mode designs are often used to increase response rates, but when
they are used to reduce costs, they can lead to lower response rates—likely respondents to
face-to-face survey requests may be less likely to participate if first asked in a different
mode, such as by mail. Apart from the choice of modes to be implemented in a study, the
order of modes can also have an impact on cost and response rates—and may likely result
in a different mix of survey errors. An equally important decision is whether to implement
modes simultaneously, giving the choice of mode to the respondent, or sequentially, often
offering the lower-cost modes first. While this is an important design decision, one that
could affect response rates, nonresponse bias, and the measurement properties of the data,
it is still in need of empirical research.
Perhaps the greatest source of error from implementing a mixed-mode design is from
measurement. Differences across modes have been identified in the research literature,
which for the most part can be attributed to three factors: interviewer versus self, visual
versus auditory, and computer versus paper-and-pencil administration. In a seminal paper
covering two of these three dimensions, Tourangeau and Smith (1996) found greater
reporting of sensitive behaviors in computer-assisted self-interviewing than in computerassisted personal interviewing, and even greater reporting of sensitive behaviors in audio
computer-assisted self-interviewing.
A large body of literature reports that interviewer-administered modes evoke socially
desirable reporting to a greater extent than do self-administered modes (Aquilino, 1994; de
Leeuw, 1992; De Maio, 1984; Hochstim, 1967). It has also been suggested that

2-3

Report Title

respondents are more likely to acquiesce in the presence of an interviewer (Schuman and
Presser, 1981). Additionally, the presentation of the survey questions (visual vs. auditory)
in each mode contributes to primacy or recency effects, as described by Krosnick and Alwin
(1987).
Finally, the mix of modes in a survey can result in different processing errors. Often
overlooked, the errors made by interviewers (e.g., coding of occupation) are different from
the errors made in the processing of paper questionnaires, which in turn are different from
those in computerized self-administered modes. Like measurement error, this is particularly
threatening when these mode-specific errors are not randomly distributed across different
sample members—and the interview mode is seldom, if ever, a random choice or
assignment.

2.4

Summary

Overall, mixed-mode designs will continue to gain popularity mainly because of their ability
to reduce costs and maximize response rates. However, careful consideration should be
given to the potential impact of such designs on the coverage, nonresponse, and
measurement properties of the data.

2.5

References

Alexander, C.H., Jr., and S. Wetrogan. (2000). “Integrating the American Community
Survey and the Intercensal Demographic Estimates Program.” In Proceedings of the
American Statistical Association.

Aquilino, W.S. 1994. “Interview Mode Effects in Surveys of Drug and Alcohol Use.” Public
Opinion Quarterly 58:210-40.
Aquilino, W.S., and L.A. LoSciuto. 1990. “Interview Mode Effects in Drug Use Surveys.”
Public Opinion Quarterly 54(3):362-95.
Biemer, P.P., and L.E. Lyberg. 2003. Introduction to Survey Quality. New York: John Wiley.
Bishop, G., H. Hippler, N. Schwarz, and F. Strack. 1988. “A Comparison of Response Effects
in Self-Administered and Telephone Surveys.” In Telephone Survey Methodology, R.
Groves, P. Biemer, L. Lyberg, J. Massey, W. Nicholls, II, and J. Waksberg, eds., pp.
321-40. New York: Wiley.
Bongers, I.M.B., and J.A.M. van Oers. 1998. “Mode Effects on Self-Reported Alcohol Use and
Problem Drinking: Mail Questionnaires and Personal Interviewing Compared.” Journal
of Studies on Alcohol 59:280-85.
Cannell, C.F., R.M. Groves, L. Magilavy, N. Mathiewetz, P. Miller, and O. Thornberry. 1987.
An Experimental Comparison of Telephone and Personal Health Interview Surveys.
Vital and Health Statistics, series 2, no. 106. DHHS Pub. No. (PHS)87-1380.
Washington, DC: U.S. Government Printing Office.

2-4

Section 2 — Mixed-Mode Surveys

Couper, M.P., and W.L. Nicholls, II. 1998. “The History and Development of ComputerAssisted Survey Information Collection Methods.” In Computer-Assisted Survey
Information Collection, M.P. Couper, R.P. Baker, J. Bethlehem, C.Z.F. Clark, J.
Martin, W.L. Nicholls, II, and J.M. O’Reilly, eds., pp. 1-22. New York: John Wiley.
de Leeuw, E.D. 1992. Data Quality in Mail, Telephone, and Face-to-Face Surveys.
Amsterdam: TT-Publicaties.
de Leeuw, E., and M. Collins. 1997. “Data Collection Methods and Survey Quality: An
Overview.” In Survey Measurement and Process Quality, L. Lyberg, P. Biemer, M.
Collins, E. de Leeuw, C. Dippo, N. Schwarz, and D. Trewin, eds., pp. 199-220. New
York: John Wiley.
De Maio, T.J. 1984. “Social Desirability and Survey Measurement: A Review.” In Surveying
Subjective Phenomena, Vol. 2, C. Turner and E. Martin, eds., pp. 257-82. New York:
Russell Sage Foundation.
Dillman, D.A., J.R. Clark, and K.K. West. 1995. “Influence of an Invitation to Answer by
Telephone on Response to Census Questionnaires.” Public Opinion Quarterly 51:20119.
Dillman, D., and J. Tarnai. 1991. “Mode Effects of Cognitively Designed Recall Questions: A
Comparison of Answers to Telephone and Mail Surveys.” In Measurement Errors in
Surveys, P. Biemer, R. Groves, L. Lyberg, M. Mathiowetz, and S. Sudman., eds., pp.
73-93. New York: Wiley.
Goyder, J. 1986. “Surveys on Surveys: Limitations and Potentialities.” Public Opinion
Quarterly 50:27-41.
Groves, R.M., F. Fowler, M. Couper, J. Lepkowski, E. Singer, and R. Tourangeau. 2004.
Survey Methodology. New York: Wiley.
Groves, R.M., and R. Kahn. 1979. Surveys by Telephone: A National Comparison With
Personal Interviews. New York: Academic Press.
Groves, R.M., and J.M. Lepkowski. 1985. “Dual Frame, Mixed-Mode Survey Designs.”
Journal of Official Statistics 1:263-86.
Hochstim, J.R. 1967. “A Critical Comparison of Three Strategies of Collecting Data From
Households.” Journal of the American Statistical Association 62:976-89.
Holbrook, A.L., M.C. Green, and J.A. Krosnick. 2003. “Telephone vs. Face-to-Face
Interviewing of National Probability Samples With Long Questionnaires: Comparisons
of Respondent Satisficing and Social Desirability Response Bias.” Public Opinion
Quarterly 67:79-125.
Kirsch, A.D., M.T. McCormack, and S.K.E. Saxon-Harrold. 2001. “Evaluation of Differences
in Giving and Volunteering Data Collected by In-Home and Telephone Interviewing.”
Nonprofit and Voluntary Sector Quarterly 30:495-504.
Kormendi, E., and J. Noordhoek. 1989. Data Quality in Telephone Surveys. Copenhagen:
Danmarks Statistik.

2-5

Report Title

Krosnick, Jon A., and D.F. Alwin. 1987. “An Evaluation of a Cognitive Theory of ResponseOrder Effects in Survey Measurement.” Public Opinion Quarterly 51:201-19.
Marquis, K.H., and R. Blass. 1985. “Nonsampling Error Considerations in the Design and
Operation of Telephone Surveys.” In Proceedings of the First Annual Research
Conference of the U.S. Bureau of the Census, pp. 301-29.
Massey, J.T., K. Marquis, and R. Tortora. 1982. “Methodological Issues Related to Telephone
Surveys by Federal Agencies.” In Proceedings of the Social Statistics Section,
American Statistical Association, pp. 63-72.
Schuman, H., and S. Presser. 1981. Questions and Answers in Attitude Surveys:
Experiments in Question Form, Wording, and Context. New York: Academic Press.
Shettle, C., and G. Mooney. 1999. “Monetary Incentives in U.S. Government Surveys.”
Journal of Official Statistics 15:231-50.
Sykes, W., and M. Collins. 1988. “Effects of Mode of Interview: Experiments in the UK.” In
Telephone Survey Methodology, R. Groves, P. Biemer, L. Lyberg, J. Massey, W.
Nicholls, II, and J. Waksberg, eds., pp. 301-20. New York: John Wiley and Sons.
Tourangeau, R., and T. Smith. 1996. “Asking Sensitive Questions: The Impact of Data
Collection, Question Format, and Question Context.” Public Opinion Quarterly
60:275-304.
Walker, A.H., and J.D. Restuccia. 1984. “Obtaining Information on Patient Satisfaction With
Hospital Care: Mail Versus Telephone.” Health Services Research 19:291-306.

2-6

3. SELF-ADMINISTERED MODES OF DATA COLLECTION
Self-administered surveys involve indirect contact with the respondent, may utilize both
visual (e.g., mail) and aural (e.g., audio computer-assisted self-interviewing [ACASI])
channels of communication, and usually do not allow for complex instruments (unless
computer administered). Self-administered modes can be used as stand-alone modes, in
mixed-mode designs, or in portions of face-to-face surveys where sensitive questions are
asked. A common feature in self-administered modes (when used as stand-alone modes or
in mixed-mode designs) is the sequence of the distribution of materials—such as advance
letters; the cover letter and questionnaire; and the reminder message and follow-up
questionnaire—used to maximize response rates (see Dillman, 1978; Dillman, 2000).
There are various types of self-administered methods of data collection that differ largely in
the extent to which they employ technology and utilize aural and visual presentation. Mail
surveys remain one of the most popular modes, in part due to the ability to use address
sampling frames. Other self-administered modes include e-mail, web, fax, optical character
recognition (OCR), disk-by-mail (DBM), touchtone data entry (TDE), voice recognition entry
(VRE), automatic speech recognition (ASR), and inbound interactive voice response (IVR).
Several self-administered modes are used as part of an interviewer-administered survey,
where the interviewer sets up the equipment, instructs the respondent in how to use it, and
is available during the interview to assist, if necessary: computer-assisted self-interviewing
(text-CASI), audio-CASI (ACASI), video-CASI (V-CASI), and audio-visual CASI (AV-CASI).
From a cost-and-error perspective, self-administered modes are often characterized by
relatively low costs when used as the primary mode but are associated with lower response
rates than interviewer-administered surveys. This leaves a substantial potential for
nonresponse bias in self-administered surveys. Often, it is not possible to disentangle
refusals from noncontacts—for example, Mathiowetz, Couper, and Singer (1994) reported
that in 63% of households in the United States, one person is responsible for opening mail
and that 63% of households throw some mail away without opening. In addition, even in
interviewer-administered surveys, breakoff rates are very high: for example, in their review
of IVR studies, Tourangeau, Steiger, and Wilson (2002) reported breakoff rates as high as
31%; similarly, Couper, Singer, and Tourangeau (2004) reported a 24% overall breakoff
rate in outbound IVR, and Gribble et al. (2000) reported a 24% breakoff rate for telephoneCASI, compared with 2% for computer-assisted telephone interviewing (CATI).

3.1

Mail Surveys

Mail surveys continue to be one of the most popular methods for data collection. The
research on mail surveys is voluminous: a bibliography compiled in the 1990s on research
to improve mail survey procedures published since 1970 included more than 400 entries
(Dillman and Sangster, 1990). Nonresponse has been the biggest challenge to mail surveys

3-1

Report Title

so far; thus, various studies have focused on procedures and techniques for maximizing
response rates. Such techniques include incentives; personalization of correspondence;
content of cover letter; questionnaire layout, length and color; follow-up reminders; and so
forth (see Dillman, 1978; Dillman, 2000). The Total Design Method (TDM) proposed by
Dillman (Dillman, 1978) and later renamed the Tailored Design Method (Dillman, 2000)
utilizes social exchange theory to guide the integration of specific procedures. The theory
posits that sample members are more likely to return the questionnaire if the perceived
benefit of doing so outweighs the perceived cost of responding. This has led to practical
recommendations on how to design a mail survey that appears interesting, trustworthy,
easy, and less time consuming to complete.
In terms of coverage, mail surveys so far have not enjoyed the degree of coverage
accomplished by face-to-face surveys. However, with the development and improvement of
a database of addresses and the promising future of the Delivery Sequence File1 for
address-based sampling, mail surveys may become a mode that offers almost complete
coverage of households in the United States at a relatively low price.
From a measurement error perspective, mail surveys have been reported to be less
susceptible to response order effects (mainly recency effects, i.e., choosing the last
response category) relative to telephone surveys (Bishop, Hippler, and Schwarz, 1988;
Ayidiya and McClendon, 1990). Another difference between mail and intervieweradministered modes that is frequently observed in research is the tendency for mail
respondents to use the entire scale when vague quantifiers are used as scale categories
rather than selecting the extremes. Such an effect was first reported by Hochstim (1967)
and was later supported by studies on mode comparisons by Dillman and Mason (1984),
Mangione, Hingson, and Barrett (1982), Talley et al. (1983), Walker and Restuccia (1984),
and Zapka, Stoddard, and Lubin (1988). One possible explanation for the observed
differences is that respondents do not interact with an interviewer and thus are less
concerned about self-representation and less likely to provide socially desirable responses
(extremes on scales). In fact, self-administered surveys in general have been reported to
yield higher rates of sensitive and socially undesirable behaviors and attitudes, possibly due
to the increased social distance between respondent and researcher and the private
environment in which the survey can be filled out.

3.2

Self-Administered Modes and Sensitive Questions

In response to the need for a private data collection environment, various (usually CASI)
techniques in which the respondents interact directly with a laptop computer for a portion of
the face-to-face interview have been utilized. A seminal article by Tourangeau and Smith
1

A computerized file that contains all delivery point addresses serviced by the U.S. Postal Service with
the exception of general delivery.

3-2

Section 3 — Self-Administered Modes of Data Collection

(1996) examined responses to computer-assisted personal interviewing (CAPI) and
interviews conducted using text-CASI and ACASI. Topics ranged from illicit drug use to
sexual behavior. The findings supported the notion that the privacy of the CASI setting
encouraged respondents’ honesty in reporting such sensitive behaviors. It was also
demonstrated that the audio component of the interview (ACASI) enhanced the feeling of
privacy, thereby increasing the level of reporting. Similar findings were reported by Aitken
et al. (2000), Hewett et al. (2004), Fu et al. (1998), Kissinger et al. (1999), and Moskowitz
(2004). A recent study by Couper, Tourangeau, and Marvin (2009) demonstrates that the
gains from using ACASI are modest relative to text-CASI and that most respondents make
limited use of the audio component.
Many national surveys that gather data about sensitive topics employ self-administration for
part of the interview. For example, the National Survey of Family Growth administers items
about pregnancies and abortions in ACASI and also in the main CAPI module. A difference of
17% in reports of abortions has been reported between ACASI and CAPI (Fu et al., 1998).
Similar findings have been reported on illicit drug use in the National Longitudinal Study of
Youth (Schoeber et al., 1992) and in a randomized experiment embedded in the 1990
National Household Survey on Drug Abuse (NHSDA) field test.
Recently, the effects of self-administered modes on socially desirable and sensitive reporting
were reexamined by Kreuter, Presser, and Tourangeau (2008). The authors used survey
and university record data to look at mode effects on the reporting of potentially sensitive
information by a sample of recent university graduates. Conventional CATI, IVR, and web
modes were compared. Web administration was found to increase the level of reporting of
sensitive information and reporting accuracy relative to conventional CATI, followed by IVR.
No significant differences in reports to sensitive and socially desirable questions have been
reported across self-administered modes (e.g., Dillman and Tarnai, 1991; Knapp and Kirk,
2003; Lensvelt-Mulders et al., 2006). Generally, computerization does not add an additional
advantage (e.g., Dillman and Tarnai, 1991), even though the use of ACASI can be
invaluable for low-literate populations.

3.3

Web Surveys

With the mass use of the Internet, web surveys became popular very fast. Web surveys
offer access to millions of potential respondents, at low cost and with rapid turnaround.
Coverage remains the biggest threat to inference from web surveys (unless the target
population is made up entirely of web users). Sampling frames for web surveys are hard to
construct because the “internet population” is different in many aspects from the general
population in the United States (Couper, 2000). Thus, web surveys often use nonprobability-based sample designs. Many survey organizations create panels of web
respondents that are recruited via a probability mode, such as phone, face-to-face, or mail.

3-3

Report Title

However, this strategy adds another layer of concern—panel conditioning that occurs with
continuous experience with a survey over time (Kalton and Citro, 1993; Kalton, Kasprzyk,
and McMillen, 1989).
When frames are available and probability methods employed (e.g., lists of e-mail
addresses of university students), web surveys generally produce lower response rates than
mail surveys (e.g., Guterbock et al., 2000; Kwak and Radler, 2002; Lesser and Newton,
2001; Lesser and Newton, 2002). The reasons for this may be many—the fact that
techniques that have proven successful in increasing response rates in mail surveys may not
work for web surveys, technical difficulties, and so forth. Concerns of privacy and
confidentiality may be a crucial factor affecting not only web survey response rates but also
the ability to collect sensitive information with less social desirability bias (Couper, 2000).
We are not familiar with research that examines the extent to which the use of web surveys
negates the ability of self-administered surveys to collect sensitive information.
From a measurement error perspective, web surveys possess unique features, such as the
ability to deliver multimedia content to respondents; however, there may be variation in
how a survey appears on a respondent’s screen (dependent on browser settings, screen
size, etc.). Various aspects of visual design features have been tested, including the use of
progress indicators (e.g., Crawford, Couper, and Lamias, 2001; Conrad et al., 2005;
Heerwegh and Loosveldt, 2006); paging versus scrolling web survey design (e.g., Peytchev
et al., 2006); definitions (e.g., Conrad et al., 2006); visual analog scales (e.g., Couper et
al., 2006); response formats (e.g., Heerwegh and Loosveldt, 2002); and interviewer
pictures, scale colors, and other visual features (e.g., Couper, Conrad, and Tourangeau,
2007; Tourangeau, Couper, and Conrad, 2007).
Web surveys are increasingly becoming a popular option in mixed-mode designs using the
choice of completion method, where the focus is on minimizing respondent burden and cost
(rather than concern about possible mode effects). Many government agencies have
introduced a web option (usually in panel surveys of establishments): for example, the
Current Employment Statistics program at the Bureau of Labor Statistics (Clayton and
Werking, 1998); and the U.S. Census Bureau’s Library Media Center survey (see Tedesko,
Zuckerberg, and Nichols, 1999; Zuckerberg, Nichols, and Tedesco, 1999).

3.4

Summary

Self-administration is a preferred mode of data collection for survey questions related to
sensitive or socially undesirable events and behaviors. This is usually achieved through the
use of various CASI techniques, even though research suggests that it is the use of selfadministration rather than computerization of the survey interview and audio components
that is believed to enhance a respondent’s privacy. Mail and web modes are the dominating
self-administered options that are used as stand-alone modes or in mixed-mode designs.

3-4

Section 3 — Self-Administered Modes of Data Collection

3.5

References

Aitken, Sherrie S., James DeSantis, Thomas C. Harford, M. Fe Caces. 2000. “Marijuana Use
Among Adults: A Longitudinal Study of Current and Former Users.” Journal of
Substance Abuse 12(3):213-26.
Ayidiya, S.A., and M.J. McClendon. 1990. “Response Effects in Mail Surveys.” Public Opinion
Quarterly 54:229-47.
Bishop, G.G., H. Hippler, and F. Schwarz. 1988. “A Comparison of Response Effects in SelfAdministered and Telephone Surveys.” In Telephone Survey Methodology, R.M.
Groves, P. Biemer, L. Lyberg, J. Massey, W. Nicholls, II, and J. Waksberg , eds., pp.
321-40. New York: Wiley and Sons.
Clayton, R.L., and G.S. Werking. 1998. “Business Surveys of the Future: The World Wide
Web as a Data Collection Methodology.” In Computer Assisted Survey Information
Collection, M.P. Couper, R.P. Baker, J. Bethlehem, C.Z.F. Clark, J. Martin, W.L.
Nicholls, II, and J. O’Reilly, eds., pp. 543-562. New York: Wiley.
Conrad, F.G., M.P. Couper, R. Tourangeau, and A. Peytchev. 2005. “Impact of Progress
Feedback on Task Completion: First Impressions Matter.” In SIGCHI 2005: Human
Factors in Computing Systems, pp. 1921-24. Portland, OR: ACM Press.
Conrad, F.G., M.P. Couper, R. Tourangeau, and A. Peytchev. 2006. “Use and Non-Use of
Clarification Features in Web Surveys.” Journal of Official Statistics 22(2):245-69.
Couper, M.P. 2000. “Web Surveys: A Review of Issues and Approaches.” Public Opinion
Quarterly 64:464-94.
Couper, M.P., F.G. Conrad, and R. Tourangeau. 2007. “Visual Context Effects in Web
Surveys.” Public Opinion Quarterly 71(4):623-34.
Couper, M.P., E. Singer, and R. Tourangeau. 2004. “Does Voice Matter? An Interactive Voice
Response (IVR) Experiment.” Journal of Official Statistics 20(3):551-70.
Couper, M.P., R. Tourangeau, F.G. Conrad, and E. Singer. 2006. “Evaluating the
Effectiveness of Visual Analog Scales: A Web Experiment.” Social Science Computer
Review 24(2):227-45.
Couper, M. P., R. Tourangeau, and T. Marvin. 2009. “Taking the Audio Out of Audio-CASI.”
Public Opinion Quarterly 73(2):281-303.
Crawford, S.D., M.P. Couper, and M.J. Lamias. 2001. “Web Surveys—Perceptions of
Burden.” Social Science Computer Review 19(2):146-62.
Dillman, D.A. 1978. Mail and Telephone Surveys: The Total Design Method. New York: Wiley
and Sons.
Dillman, D.A. 2000. Mail and Internet Surveys: The Tailored Design Method. New York:
Wiley and Sons.

3-5

Report Title

Dillman, D.A., and R.G. Mason. 1984. “The Influence of Survey Methods on Question
Response.” Paper presented at the Annual Meeting of the American Association for
Public Opinion and Research, Delavan, WI, May 17-20.
Dillman, D.A., and R.L. Sangster. 1990. Mail Surveys: A Comprehensive Bibliography 1974–
1989. Technical Report. Pullman, WA: Social and Economic Sciences Research
Center, Washington State University.
Dillman, D., and J. Tarnai. 1991. “Mode Effects of Cognitively Designed Recall Questions: A
Comparison of Answers to Telephone and Mail Surveys.” In Measurement Errors in
Surveys, P. Biemer, R. Groves, L. Lyberg, N. Mathiowetz, and S. Sudman, eds., pp.
73-93. New York, NY: John Wiley and Sons.
Fu, Haishan, Jacqueline E. Darroch, Stanley K. Henshaw, Elizabeth Kolb. 1998. “Measuring
the Extent of Abortion Underreporting in the 1995 National Survey of Family
Growth.” Family Planning Perspectives 30(3):128-38.
Gribble, James N., Heather G. Miller, Joseph A. Catania, Lance Pollack, and Charles F.
Turner. 2000. “The Impact of T-ACASI Interviewing on Reported Drug Use among
Men Who Have Sex with Men.” Substance Use and Misuse 35:869–90.
Guterbock, T.M., B.J. Meekins, A.C. Weaver, and J.C. Fries. 2000. “Web Versus Paper: A
Mode Experiment in a Survey of University Computing.” Paper presented at the
Annual Meeting of the American Association for Public Opinion Research, Portland,
OR, May 18-21.
Heerwegh, D., and G. Loosveldt. 2002. “An Evaluation of the Effect of Response Formats on
Data Quality in Web Surveys.” Social Science Computer Review 20(4):471-84.
Heerwegh, D., and G. Loosveldt. 2006. “An Experimental Study on the Effects of
Personalization, Survey Length Statements, Progress Indicators, and Survey Sponsor
Logos in Web Surveys.” Journal of Official Statistics 22:191-210.
Hewett, P.C., B. S. Mensch, and A. Erulkar. 2004. “Consistency in the Reporting of Sexual
Behaviour by Adolescent Girls in Kenya: A Comparison of Interviewing Methods.”
Sexually Transmitted Infections 80(Suppl.2):43-8.
Hochstim, J. R. 1967. “A Critical Comparison of Three Strategies of Collecting Data From
Households.” Journal of the American Statistical Association 62:976-89.
Kalton, G., and C.F. Citro. 1993. “Panel Surveys: Adding the Fourth Dimension.” Survey
Methodology 19(2):205-15.
Kalton, G., D. Kasprzyk, and D.B. McMillen. 1989. “Nonsampling Errors in Panel Surveys.”
In Panel Surveys, D. Kasprzyk, G. Duncan, G. Kalton, and M.P. Singh, eds., pp. 24970. New York: Wiley.
Kissinger, Patricia, Janet Rice, Thomas Farley, Shelly Trim, Kayla Jewitt, Victor Margavio,
and David H. Martin. 1999. “Application of Computer-Assisted Interviews to Sexual
Behavior Research.” American Journal of Epidemiology 149(10):950-54.

3-6

Section 3 — Self-Administered Modes of Data Collection

Knapp, H., and S.A. Kirk. 2003. “Using Pencil and Paper, Internet and Touch-Tone Phones
for Self-Administered Surveys: Does Methodology Matter?” Computers in Human
Behavior 19(1):117-34.
Kreuter, F, S. Presser, and R. Tourangeau. 2008. “Social Desirability Bias in CATI, IVR, and
WEB Surveys: The Effects of Mode and Question Sensitivity.” Public Opinion
Quarterly 72(5):847-65.
Kwak, N., and B. Radler. 2002. “A Comparison Between Mail and Web Surveys: Response
Pattern, Respondent Profile, and Data Quality.” Journal of Official Statistics
18(2):257-74.
Lensvelt-Mulders, G.J.L.M., P.G.M. van der Heijden, O. Laudy, and G. van Gils. 2006. “A
Validation of a Computer-Assisted Randomized Response Survey to Estimate the
Prevalence of Fraud in Social Security.” Journal of the Royal Statistical Society Series
A–Statistics in Society 169:305-18.
Lesser, V.M., and L. Newton. 2001. “Mail, Email, and Web Surveys. A Cost and Response
Rate Comparison in a Study of Undergraduate Research Activity.” Paper presented at
the Annual Meeting of the American Association for Public Opinion Research,
Montreal, Canada, May 17-20.
Lesser, V.M., and L. Newton, L. 2002. “Comparison of Response Rates and Quality of
Response in a Survey Conducted by Mail, E-mail, and Web.” Paper presented at the
Annual Meeting of the American Association for Public Opinion Research, St. Pete
Beach, FL, May 16-19.
Mangione, T.W., R. Hingson, and J. Barrett. 1982. “Collecting Sensitive Data: A Comparison
to Three Survey Strategies.” Sociological Methods Research 10:337-46.
Mathiowetz, Nancy, Mick P. Couper, and Eleanor Singer. 1994. “Where Does All the Mail Go?
Mail Receipt and Handling in U.S. Households.” Survey Methodology Program
Working Paper No. 25. Ann Arbor: University of Michigan.
Moskowitz, J.M. 2004. “Assessment of Cigarette Smoking and Smoking Susceptibility Among
Youth—Telephone Computer-Assisted Self-Interviews Versus Computer-Assisted
Telephone Interviews.” Public Opinion Quarterly 68(4):565-87.
Peytchev, A., M.P. Couper, S.E. McCabe, and S. Crawford. 2006. “Web Survey Design:
Paging vs. Scrolling.” Public Opinion Quarterly 70(4):596-607.
Schoeber, Susan E., M. Fe Caces, Michael R. Pergamit, and Laura Branden. 1992. “Effect of
Mode of Administration on Reporting of Drug Use in the National Longitudinal
Survey.” In Survey Measurement of Drug Use: Methodological Studies, Charles F.
Turner, Judith T. Lessler, and Joseph C. Gfroerer, eds., pp. 267-76. DHHS Publication
no. (ADM) 92-1929. Washington, DC: U.S. Department of Health and Human
Services, Public Health Service, Alcohol, Drug Abuse, and Mental Health
Administration.
Talley, J.E., J.C. Barrow, K.F. Fulkerson, and C.A. Moore. 1983. “Conducting a Needs
Assessment of University Psychological Services: A Campaign of Telephone and Mail
Strategies.” Journal of American College Health 32:101-03.

3-7

Report Title

Tedesco, H., R.L. Zuckerberg, and E. Nichols. 1999. “Designing Surveys for the Next
Millennium: Web-Based Questionnaire Design Issues.” In Proceedings of the Third
Association for Survey Computing International Conference, pp. 103-12. Edinburgh,
September 22-4.
Tourangeau, R., M.P. Couper, and F. Conrad. 2007. “Color, Labels, and Interpretive
Heuristics for Response Scales.” Public Opinion Quarterly 71(1):91-112.
Tourangeau, R., and T.W. Smith. 1996. “Asking Sensitive Questions: The Impact of Data
Collection Mode, Question Format, and Question Context.” Public Opinion Quarterly
60:275-304.
Tourangeau, R., D. Steiger, and D. Wilson. 2002. “Self-Administered Questions by
Telephone: Evaluating Interactive Voice.” Public Opinion Quarterly 66:265-78.
Walker, A.H., and J.D. Restuccia. 1984. “Obtaining Information on Patient Satisfaction With
Hospital Care: Mail Versus Telephone.” Health Service Research 19:291-306.
Zapka, J.G., A.M. Stoddard, and H. Lubin. 1988. “A Comparison of Data From Dental Charts,
Client Interview, and Client Mail Survey.” Medical Care 26:27-33.
Zuckerberg, A., E. Nichols, and H. Tedesco. 1999. “Designing Surveys for the Next
Millennium: Internet Questionnaire Design Issues.” Paper presented at the annual
conference of the American Association for Public Opinion Research, St. Petersburg
Beach, FL, May 16-19.

3-8

4. USE OF INCENTIVES
The use of incentives in surveys has been studied for decades. The literature surrounding
the use of incentives details multiple dimensions that impact the effectiveness of incentives.
These include theories on why incentives work, impact on response rate and nonresponse
bias, prepaid versus postpaid incentives, and mode differences.

4.1

Theories on Incentive Effectiveness

Different reasons for the effectiveness of incentives have been provided in the literature.
The theory of social exchange in the field of social psychology suggests a mechanism of
social indebtedness, in response to which the individual cooperates with a survey request
(Dillman, 1978). While social exchange would require that the sample member does not link
the incentive to the survey request, a feature of the use of social exchange is that the
incentive is rather small, so it is construed as a token of appreciation rather than a form of
compensation for time and effort. This would suggest an incentive amount that has a small
value. Kulka (1994) conducted an extensive overview of the existing literature and
concluded that there was support for the belief that small monetary incentives increased
response rates—a phenomenon largely attributed to social exchange.
Another reason for the effectiveness of incentives is more direct and can be described by
theories such as economic exchange: an incentive is a form of compensation for
participating in the survey. For some respondents, a particular compensation amount may
be below a threshold level at the time of a survey request, but the higher the incentive, the
more respondents decide to participate in the survey. Indeed, multiple studies have
demonstrated that, for incentives, more is usually better. Trussel and Lavrakas (2004)
examined the effect of incremental incentive increase in an experiment launched in a largescale, mixed-mode survey. The levels of tested incentives ranged from $0 to $10.
Consistent with previous findings, sending $1 versus not sending an incentive at all resulted
in higher response rates. The incremental increase in the incentive amount had a differential
effect, depending on the outcome of the prior contact with the household. For households
with positive outcome, it was not until the amount of $5 was reached when the response
and cooperation rates became significantly higher, relative to $1. More interestingly, the $7
to $10 condition did not differ significantly from $6. In contrast, in households that were
never initially contacted or had negative outcome, each incremental dollar had a larger
impact on response and cooperation than the previous dollar amount. The result suggests
that when there is negative previous contact with the sample person, researchers should
spend the maximum allowed in the budget on incentives.
Brick et al. (2005) compared the effectiveness of prepaid $0, $2, and $5 incentives at
various stages of a random-digit dialing (RDD) survey on educational topics. Brick et al.
(2005) found that $5 was more effective than $2 in achieving initial cooperation, but the

4-1

Report Title

relative effectiveness of the incentive (defined as the percentage point increase in the initial
cooperation rate per dollar when compared to no incentive) was higher for the $2 condition.
Furthermore, incentives provided at the refusal conversion stage (a letter was mailed before
calling) were more effective than incentives provided at the recruitment stage of the survey.
It is yet unknown whether respondents really construe a small incentive as a token of
appreciation as opposed to a small amount of compensation, but in addition to the cognitive
mechanism at play, a small token of appreciation can have a very different impact on
survey costs compared to a larger compensation.

4.2

Impact on Response Rate and Nonresponse Bias

Offering respondent incentives is a demonstrated method to increase cooperation and
response rates, but more importantly, it is also a method to decrease nonresponse bias.
Sample members participate in surveys for various reasons. The leverage-salience theory
(Groves, Singer, and Corning, 2000) posits that different people place different levels of
importance on features of the survey request, such as the survey topic, survey sponsor,
interview length, and so forth. Depending on what is made salient when the sample person
is approached, the outcome of the survey request can be a refusal or an acceptance. For
example, those less interested or involved in the survey topic can cooperate at a lower rate,
leading to nonresponse bias in estimates based on the respondents. Incentives have been
shown to increase cooperation particularly among sample persons with lower topic
involvement. In a study that tests the theoretical framework based on the leverage-salience
hypothesis, Groves, Singer, and Corning (2000) compared incentive and no-incentive
treatments in a survey about political and community involvement. As expected, incentives
significantly increased response rates. More interestingly, however, the effect of incentives
was diminished for sample persons with high community involvement. Similar results were
reported earlier by Baumgartner and Rathbun (1996), who found that monetary incentives
increased cooperation more among those less interested in the survey topic. Such findings
suggest that by attracting respondents who normally would not take part in the survey,
incentives also changed the mix of sample persons who are measured, thus presenting a
potential for reducing nonresponse error. However, in another test of the leverage-salience
theory, Groves, Presser, and Dipko (2004) failed to find significant effects of monetary
incentives in reducing the effect of topic interest on survey participation.
The link between response rates and nonresponse bias arises when there is a clear
connection between response propensity and a survey variable of interest. The use of
incentives may influence both the participation decision and survey variables. In a series of
experiments launched to test whether those interested in the survey topic participate at
higher rates and whether nonresponse bias on estimates of variables reflecting the survey
topic was affected by this, Groves et al. (2006) also examined whether the use of incentives
affected the link between topic interest and nonresponse bias. Incentives did not reliably

4-2

Section 4 — Use of Incentives

dampen the effect of topic interest, even though the results were in the hypothesized
direction.

4.3

Prepaid Versus Postpaid Incentives

Another important factor when considering incentives is whether to offer them in advance,
regardless of the sample person’s decision to participate in the survey (prepaid), or after
the respondent has agreed and completed the survey (promised). Some studies have found
only prepaid incentives to be effective in reducing nonresponse in interviewer-administered
surveys (Berk et al., 1987; Cantor et al., 1998; Singer, Van Hoewyk, and Maher, 2000),
while Bosnjak and Tuten (2003) have found no difference between prepaid and promised
incentives in web surveys.
Various studies have demonstrated the stronger effect of prepaid versus promised monetary
incentives in mail surveys (for an overview, see Linsky, 1975; Armstrong, 1975). A metaanalysis of the experimental work on incentives in mail surveys by Church (1993) concluded
that prepaid incentives yielded higher response rates than promised incentives or gifts sent
with the initial mailing (65% average increase). Furthermore, it was concluded that an
increase in the amount of money sent translated to an increase in response rates (but as
Armstrong [1975] and Fox, Crask, and Kim [1998] suggest, at a decreasing rate).
Certain designs do not allow for prepaid incentives (e.g., most RDD surveys, or surveys of
the whole household when the number of household members is unknown). In such cases,
the amount offered may determine to a large degree the effectiveness of the incentive. For
example, Cantor, Wang, and Abi-Habibm (2003) found an almost 10% increase in the
response rate when promising $20 (vs. no incentive) in an RDD survey of caregivers to
children aged newborn to 17. Strouse and Hall (1997) recommend that for a survey to be
successful, promised incentives have to be quite large (in the $15 to $35 range).
Promised incentives are fairly common at the refusal conversion stage. A number of studies
have reported gains in response rates through offering relatively large amounts of money
($25 or greater) at the end of the data collection period (e.g., Olson et al., 2004; Curtin,
Presser, and Singer, 2005).

4.4

Incentives and Survey Mode

Comparison of the respondent conditions in self-administered versus intervieweradministered surveys suggests that the need for incentives will be greater in selfadministered modes, where the persuasive presence of an interviewer is missing. In a
meta-analysis that included face-to-face, telephone, and mixed-mode surveys, Singer et al.
(1999) found that the effect of incentives was largely the same across modes. The results
suggested that prepaid incentives yielded significant improvement in response rates, and

4-3

Report Title

gifts were found to be significantly less effective than monetary incentives, even controlling
for the value of the incentive.
It remains unknown whether nonmonetary incentives that appeal only to some respondents
produce the same expected reduction in bias that is usually associated with monetary
incentives. To an extreme, it is unclear whether such incentives may even induce bias in
survey estimates—similarly to the bias induced through topic interest.

4.5

Summary

Despite these arguments and empirical findings, incentives may not be included in a study
design due to their cost. Yet incentives can reduce the cost per case through the need for
fewer interviewer call attempts to sample members and for the more costly refusal
conversion attempts, as evidenced by the incentive experiments conducted for the National
Survey on Drug Use and Health. The cost per interview in the $20 group was 5% lower than
the control; in the $40 group, costs were 4% lower than the control. The cost savings were
gained by interviewers spending less time trying to obtain cooperation from respondents
(Kennet et al., 2005).

4.6

References

Armstrong, J. Scott. 1975. “Monetary Incentives in Mail Surveys.” Public Opinion Quarterly
39:111-16.
Baumgartner, Robert, and Pamela Rathbun. 1996. “Prepaid Monetary Incentives and Mail
Survey Response Rates.” Paper presented at the Joint Statistical Meetings, Chicago,
IL, Aug 4-8.
Berk, M.L., N.A. Mathiowetz, E.P. Ward, and A.A. White. 1987. “The Effect of Prepaid and
Promised Incentives: Results of a Controlled Experiment.” Journal of Official
Statistics 3(4):449-57.
Bosnjak, Michael, and Tracy Tuten. 2003. “Prepaid and Promised Incentives in Web
Surveys: An Experiment.” Social Science Computer Review 21(2):208-17.
Brick, Michael, Jill Montaquila, Mary Collins Hagedorn, Shelley Brock Roth, and Christopher
Chapman. 2005. “Implications for RDD Design From an Incentive Experiment.”
Journal of Official Statistics 21(4):571-89.
Cantor, D., Allen B., P. Cunningham, J.M. Brick, R. Slobasky, P. Giambo, and G. Kenny.
1998. “Promised Incentives on a Random Digit Dial Survey, Appearing in NonResponse in Survey Research.” In Proceedings of the Eighth International Workshop
on Household Survey Non-Response, A. Koch and R. Porst, eds., pp. 219-28.
Mannheim, Germany: ZUMA.
Cantor, D., K. Wang, and N. Abi-Habibm. 2003. “Comparing Promised and Prepaid
Incentives for an Extended Interview on a Random Digit Dial Survey.” Paper
presented at the Annual Conference at the American Association for Public Opinion,
Nashville, TN, May 15-18.

4-4

Section 4 — Use of Incentives

Church, Allan H. 1993. “Estimating the Effect of Incentives on Mail Survey Response Rates:
A Meta-Analysis.” Public Opinion Quarterly 57:62-79.
Curtin, R., S. Presser, and E. Singer. 2005. “Changes in Telephone Survey Nonresponse
Over the Past Quarter Century.” Public Opinion Quarterly 69:87-98.
Dillman, Don A. 1978. Mail and Telephone Surveys. New York: Wiley.
Fox, R.J., M.R. Crask, and J. Kim. 1988. “Mail Survey Response Rate: A Meta-Analysis of
Selected Techniques for Inducing Response.” Public Opinion Quarterly 52:467-91.
Groves, Robert M., Mick P. Couper, Stanley Presser, Eleanor Singer, Roger Tourangeau,
Giorgina Piani Acosta, and Lindsay Nelson. 2006. “Experiments in Producing
Nonresponse Bias.” Public Opinion Quarterly 70(5):720-36.
Groves, Robert M., Stanley Presser, and Sarah Dipko. 2004. “The Role of Topic Interest in
Survey Participation Decisions.” Public Opinion Quarterly 68(1):2-31.
Groves, Robert M., Eleanor Singer, and Amy Corning. 2000. “Leverage-Saliency Theory of
Survey Participation—Description and an Illustration.” Public Opinion Quarterly
64(3):299-308.
Kennet, Joel, Joseph Gfroerer, Katherine R. Bowman, Peilan C. Martin, and David
Cunningham. 2005. “Introduction of an Incentive and Its Effects on Response Rates
and Costs in NSDUH.” In Evaluating And Improving Methods Used in the National
Survey on Drug Use and Health, J. Kennet and J, Gfroerer, eds., pp. 7-17. DHHS
Publication No. SMA 05-4044, Methodology Series M-5. Rockville, MD: Substance
Abuse and Mental Health Services Administration, Office of Applied Studies.
Kulka, Richard. 1994. “The Use of Incentives to Survey ‘Hard-to-Reach’ Respondents: A
Brief Review of Empirical Research and Current Research Practices.” Paper presented
at the Seminar on New Directions in Statistical Methodology, Council of Professional
Agencies on Federal Statistics, Washington, DC, May 25-26.
Linsky, Arnold. 1975. “Stimulating Responses to Mailed Questionnaires: A Review.” Public
Opinion Quarterly 39:82-101.
Olson, L., M. Frankel, K.S. O’Connor, S.J. Blumberg, M. Kogan, and S. Rodkin. 2004. “A
Promise or a Partial Payment: The Successful Use of Incentives in an RDD Survey.”
Paper presented at the Annual Conference of the American Association for Public
Opinion Research, Phoenix, AZ, May 13-16.
Singer, Eleanor, John Van Hoewyk, Nancy Gebler, Trivellore Raghunathan, and Katherine
McGonagle. 1999. “The Effect of Incentives on Response Rates in InterviewerMediated Surveys.” Journal of Official Statistics 15:217-30.
Singer, E., J. Van Hoewyk, and M.P. Maher. 2000. “Experiments With Incentives in
Telephone Surveys.” Public Opinion Quarterly 64(2):171-88.
Strouse, R.C., and J.W. Hall. 1997. “Incentives in Population Based Health Surveys.” In
Proceedings of the American Statistical Association, Survey Research Section,
pp.952-57.

4-5

Report Title

Trussel, Norm, and Paul Lavrakas. 2004. “The Influence of Incremental Increase in Token
Cash Incentives on Mail Survey Response: Is There an Optimal Amount?” Public
Opinion Quarterly 68(3):349-67.

4-6

5. ADDITIONAL ISSUES IN MEASURING CRIME VICTIMIZATION
IN SURVEYS
Several methodological issues are of particular relevance to surveys collecting data on crime
victimization, including problems with respondents’ recalling and dating victimization
incidents correctly, the use of proxy respondents, perceptions of crime severity, survey
context, stigma, and terminology used in survey questions. The purpose of this document is
to provide the Bureau of Justice Statistics with an overview of additional issues in measuring
crime victimization surveys. This information will be used to inform the data collection
methods project being conducted as part of the overall redesign of the National Crime
Victimization Survey (NCVS).

5.1

Event Recall

Survey designers rely on respondents’ recall when collecting reports of past behaviors and
events. Accuracy of self-reports of past behaviors and autobiographical events is challenged
by the failure to encode the event initially, telescoping (reporting of events outside the
reference frame), or other sources of recall loss (Tourangeau, Rips, and Rasinski, 2000).
Not all encoded events are easily retrieved. Various studies have demonstrated that
accuracy of responses to autobiographical questions depends on passage of time (Cannell,
Miller, and Oksenberg, 1981; Loftus et al., 1992; Means et al., 1989; Smith and Jobe,
1994), length of reference period (for a meta-analysis, see Sudman and Bradburn, 1973),
event salience characteristics (e.g., Thompson et al., 1996; Wagenaar, 1986), and question
aids used to improve recall (e.g., Brewer, 1988; Wagenaar, 1986). Commonly used
question aids are situational cues (e.g., physical context, date) and retrieval cues (e.g.,
examples of similar events). To improve crime report accuracy, the NCVS 1992 redesign
introduced the short-cue screener strategy. The short-cue screener model attributed the
failure to report crime incidents to a lack of conceptual question understanding, memory
failure, or intentional misreporting; the redesign attempted to address the first two sources
by using person and location reference frames and by increasing the number and variety of
cues presented to the respondent. Preliminary tests of the short-cue screener yielded crime
report rates 19% greater than the rates produced when the original screening questions
were used (Martin et al., 1986). Several field tests conducted by the U.S. Census Bureau
reported similar findings—significantly higher rates of violence and crime reporting for the
short-cue screener group relative to the original screener group (Hubble, 1990). As
expected, the introduction of the short-cue screener in 1992 yielded more reports of
victimizations and captured types of crimes that were previously undetected (Rand, Lynch,
and Cantor, 1997). It improved the measurement of traditionally underreported crimes
(e.g., rape and aggravated assault) and crimes committed by family members and
acquaintances (Kindermann, Lynch, and Cantor, 1997). The differences were largely

5-1

Report Title

attributed to explicit cueing of certain crime types (e.g., rape and sexual assaults) and the
addition of two reference frames to aid recall: the first, related to crimes committed by
someone the respondent knew; the second, related to the location of the crime (U.S.
Bureau of the Census, 1994).
In the search for strategies that improve recall, a number of studies have examined the
issues of forward telescoping, the reporting of events that occurred prior to the reference
period (e.g., Neter and Waksberg, 1964; Loftus and Marburger, 1983; Brown, Rips, and
Shevell, 1985; Loftus et al., 1990), and backward telescoping, the reporting of events that
occurred after the reference period (e.g., Sudman and Bradburn, 1973; Means et al., 1989).
One of the design strategies used to reduce telescoping in panel surveys and employed by
NCVS is bounded recall (Neter and Waksberg, 1964), a technique where the responses from
the first interview are used to anchor responses from following interviews.
Another approach to assist event recall is known as anchoring, which uses events such as
holidays, major public events, personal landmarks, and so forth (Linton, 1975; Loftus and
Marburger, 1983; Brown, Shevell and Rips, 1986; Means et al., 1989). Yet another
approach is to vary the length of the reference period depending on how salient and rare
the event is judged by the researcher, the premise being that longer reference periods can
be used for rare and salient events (Sudman and Bradburn, 1974; Mathiowetz, 1988;
Warner et al., 2005). An examination of recall biases in NCVS revealed that rates of
victimization decreased significantly as the length of the reference period increased
(Bushery, 1981). In a reverse record-check study of victims of robbery, burglary, and
assault, Czaja et al. (1994) found that the length of reference period and anchoring did not
affect victimization rates; however, both factors influenced reports of victimization dates.
Furthermore, Event History Calendars (EHCs) have been employed to facilitate recall. EHCs
facilitate the use of all memory retrieval mechanisms (top-down, sequential, and parallel).
Such calendars rely on inherent cueing mechanisms: noteworthy events can be dated
precisely and used as landmarks for other events; events remembered in one life domain
can cue events that happened in another; and inconsistencies can be spotted easily and
addressed. Freedman et al. (1988) found almost 90% agreement between monthly reports
in EHC for events that occurred 5 years prior and validation data. Similar rates were
reported by Caspi et al. (1996) when retrospective reports were matched to concurrent
reports 3 years prior. Further, Belli, Shay, and Stafford (2001) found that EHC led to betterquality retrospective reports on key social and economic events measured by the Panel
Survey of Income Dynamics.

5.2

Proxy Respondents

Many surveys use proxy respondents when the sample member is not available for an
interview. Proxy reports offer time and cost savings, but they often do so at the price of

5-2

Section 5 — Additional Issues in Measuring Crime Victimization in Surveys

data quality. The validity of a proxy report depends largely on the relationship of the proxy
to the respondent, the saliency of the event being reported, and the proxy’s knowledge of
the event. Cantor and Lynch (2000) discussed the results of a pilot test of NCVS that found
far greater reporting of victimization with self-reports compared with proxy reports. Such
findings are consistent with other studies comparing self-reports and proxy reports in other
surveys (e.g., Hyland et al., 1997; Perruccio and Badley, 2004; Rajmil et al.,1999).

5.3

Crime Severity, Survey Context, Stigma, and Terminology

Respondents may not report smaller, less severe crimes, such as simple assault (attack
without a weapon resulting in minor or no injury), because they may not believe the
incident was serious enough to be considered a crime. Respondents may fail to recall the
incident or may choose not to report it due to the perceived ambiguity of the crime. Crimes
committed by nonstrangers (e.g., family members, intimates, acquaintances) may also be
underreported for this reason (Kinderman, Lynch, and Cantor, 1997).
Surveys that measure victimization outside the context of criminal behavior, such as the
National Violence Against Women Survey, have produced higher estimates of rape,
domestic violence, and assault than the crime-focused NCVS. The contextual differences in
the surveys may contribute to the different estimates of victimization because respondents
are less likely to report incidents to NCVS that they do not consider to be criminal (Rand
and Rennison, 2004). Additionally, the social and cultural stigmas attached to rape and
domestic violence may result in underreporting.
Much attention has been given to the measurement of rape, including wording in survey
questions (Fisher and Cullen, 2000). Research has demonstrated that the terms used and
the specificity of questions can influence victimization reports. Different terms used to ask
about sexual victimization may have different meanings to different respondents and, as a
result, may influence respondents’ understanding of the question and, ultimately, their
reporting (Hamby and Koss, 2003). The use of legal terms may also impede comprehension.
Behaviorally specific questions and specific descriptions of sexual acts produce higher rates
of sexual victimization than the use of legalistic terms such as “rape” and “sexual assault”
(Fisher, 2004; Hamby and Koss, 2003).
In addition, the way a crime is enumerated affects the accuracy of the survey estimates. For
example, repeated victimizations are common in cases of domestic violence. Concerns have
been raised on how to accurately count repeated victimizations or series of victimizations.
NCVS counts six or more similar victimizations that happened within a 6-month reference
period as one incident (based on the most recent incident). Other surveys that calculate
each incident separately produce higher estimates (e.g., Rand and Rennison, 2005).

5-3

Report Title

5.4

Summary

A wide body of research addresses issues relevant to collecting valid survey data on crime
victimization. The length of the reference period, questionnaire design aids used to improve
recall, proxy respondents, perceptions of crime, stigma associated with the crime, and the
choice of words in survey questions are among the factors that can affect the accuracy of
crime reports. Careful consideration of such features at the survey design stage and the
selection of the most appropriate mode of data collection may drastically improve report
accuracy.

5.5

References

Belli, Robert F., William L. Shay, and Frank P. Stafford. 2001. “Event History Calendars and
Question List Surveys: A Direct Comparison of Interviewing Methods.” Public Opinion
Quarterly 65:45-74.
Brewer, W.F. 1988. “Memory for Randomly Sampled Autobiographical Events.” In
Remembering Reconsidered: Ecological and Traditional Approaches to the Study of
Memory, U. Neisser and E. Winograd, eds., pp. 21-90. Cambridge: Cambridge
University Press.
Brown, N.R., L.J. Rips, and S.M. Shevell.1985. “Subjective Dates of Natural Events in VeryLong-Term Memory.” Cognitive Psychology 17:139-77.
Brown, N,R., Shevell, S.K, and Rips, L.J.(1986). Public Memories and Their Personal
Context, In Autobiographical Memory, D.G. Rubin (ed.). Cambridge: Cambridge University
Press.
Bushery, J.M. 1981. “Recall Biases for Different Reference Periods in the National Crime
Survey.” In Proceedings of the Section on Survey Research Methods, American
Statistical Association, pp. 238-43.
Cannell, C., P. Miller, and L. Oksenberg. 1981. “Research on Interviewing Techniques.” In
Sociological Methodology 1981, S. Leinhardt, ed., pp. 389-437. San Francisco:
Jossey-Bass.
Cantor, David, and James P. Lynch. 2000. “Self-Report Surveys as Measures of Crime and
Criminal Victimization.” Measurement and Analysis of Crime and Justice 4:85-138.
Caspi, A., T.E. Moffitt, A. Thornton, D. Freedman, J.W. Amel, H. Harrington, J. Smeiijers,
and P.A. Silva. 1996. “The Life History Calendar: A Research and Clinical Assessment
Method for Collecting Retrospective Event-History Data.” International Journal of
Methods in Psychiatric Research 6:101-14.
Czaja, Ronald, Johnny Blair, Barbara Bickart, and Elizabeth Eastman. 1994. “Respondent
Strategies for Recall of Crime Victimization Incidents.” Journal of Official Statistics
10(3):257-76.
Fisher, B.S. 2004. Measuring Rape Against Women: The Significance of Survey Questions.
NCJ 199705. Washington, DC: U.S. Department of Justice, National Institute of
Justice.

5-4

Section 5 — Additional Issues in Measuring Crime Victimization in Surveys

Fisher, B.S., and F.T. Cullen. 2000. “Measuring the Sexual Victimization of Women:
Evolution, Current Controversies and Future Research.” In Criminal Justice 2000, Vol.
4: Measurement and Analysis of Crime, D. Duffee, ed., pp. 317-90. Washington, DC:
U.S. Department of Justice, National Institute of Justice.
Freedman, D., A. Thornton, D. Camburn, D. Alwin, and L. Young-DeMarco. 1988. “The Life
History Calendar: A Technique for Collecting Retrospective Data.” In Sociological
Methodology, C. Clogg, ed., pp. 37-68. New York: Academic Press.
Hamby, S.L., and M.P. Koss. 2003. “Shades of Gray: A Qualitative Study of Terms Used in
the Measurement of Sexual Victimization.” Psychology of Women Quarterly
27(3):243-55.
Hubble, D.L. 1990. “National Crime Survey New Questionnaire Phase-in Research:
Preliminary Results.” Paper presented at the International Conference on
Measurement Errors in Surveys, Tucson, AZ, November 11-14.
Hyland A., K.M. Cummings, W.R. Lynn, D. Corle, and C.A. Giffen. 1997. “Effect of ProxyReported Smoking Status on Population Estimates of Smoking Prevalence.” American
Journal of Epidemiology 145(8):746-51.
Kinderman, J.L., J. Lynch, and D. Cantor. 1997. Effects of the Redesign on Victimization
Estimates. NCJ 164381. Washington, DC: U.S. Department of Justice, Bureau of
Justice Statistics.
Linton, M. 1975. “Memory for Real World Events.” In Explorations in Cognition, D.A. Norman
and D.E. Rumelhart, eds., pp. 376-404. San Francisco: Freeman.
Loftus, E.F., M.R. Klinger, K.D. Smith, and J. Fiedler. 1990. “Tale of Two Questions.” Public
Opinion Quarterly 54:330-35.
Loftus, E.F., and W. Marburger. 1983. “Since the Eruption of Mt. St. Helens, Has Anyone
Beaten You Up? Improving the Accuracy of Retrospective Reports With Landmark
Events.” Memory and Cognition 11:114-20.
Loftus, E.F., K.D. Smith, M.R. Klinger, and J. Fiedler. 1992. “Memory and Mismemory for
Health Events.” In Questions About Questions: Inquiries Into the Cognitive Basis of
Surveys, J.M. Tanur, ed., pp. 102-37. New York: Russell Sage Foundation.
Martin, E., R.M. Groves, V.J. Matlin, and C. Miller. 1986. Report on the Development of
Alternative Screening Procedures for the National Crime Survey. Washington, DC:
Bureau of Social Science Research.
Mathiowetz, N.A. 1988. “Forgetting Events in Autobiographical Memory: Findings From a
Health Care Survey.” In Proceedings of the Section on Survey Research Methods,
American Statistical Association, pp. 167-72.
Means, B., A. Nigam, M. Zarrow, E.F. Loftus, and M. Donaldson. 1989. Autobiographical
Memory for Health-Related Events Series 6, No. 2. Department of Health and Human
Services (DHHS) Publication No. (PHS) 89-1077. Washington, DC: U.S. Government
Printing Office.

5-5

Report Title

Neter, J., and J. Waksberg. 1964. “A Study of Response Errors in Expenditures Data From
Household Interview.” Journal of the American Statistical Association 59:18-55.
Perruccio, A.V., and E.M. Badley. 2004. “Proxy Reporting and the Increasing Prevalence of
Arthritis in Canada.” Canadian Journal of Public Health 95(3):169-73.
Rajmil, L., E. Fernandez, R. Gispert, M. Rue, J.P. Glutting, A. Plasencia, and A. Segura.
1999. “Influence of Proxy Respondents in Children’s Health Interview Surveys.”
Journal of Epidemiology and Community Health 53(1):38-42.
Rand, M., J. Lynch, and D. Cantor. 1997. Criminal Victimization, 1973-95. NCJ-163069.
Washington, DC: Bureau of Justice Statistics.
Rand, M.R., and C.M. Rennison. 2004. How Much Violence Against Women Is There? NCJ
199702. Washington, DC: U.S. Department of Justice, National Institute of Justice.
Rand, M.R., and C.M. Rennison. 2005. “Bigger Is Not Necessarily Better: An Analysis of
Violence Against Women Estimates From the National Crime Victimization Survey
and the National Violence Against Women Survey.” Journal of Quantitative
Criminology 21(3):267-91.
Smith, A.F., and J.B. Jobe. 1994. “Validity of Reports of Long-Term Dietary Memories: Data
and a Model.” In Autobiographical Memory and the Validity of Retrospective Reports,
N. Schwarz and S. Sudman, eds., pp. 121-40. Berlin: Springer.
Sudman, S., and N. Bradburn. 1973. “Effects of Time and Memory Factors on Response in
Surveys.” Journal of the American Statistical Association 68:805-15.
Sudman, S., and N.M. Bradburn. 1974. Response Effects in Surveys: A Review and
Synthesis. Chicago: Aldine.
Thompson, C.P., J.J. Skowronski, S.F. Larsen, and A.L. Betz. 1996. Autobiographical
Memory. Mahwah, NJ: Erlbaum.
Tourangeau, Roger, Lance J. Rips, and Kenneth Rasinski. 2000. The Psychology of Survey
Response. Cambridge: Cambridge University Press.
U.S. Bureau of the Census. 1994. The National Crime Victimization Survey(NCVS) Redesign:
Technical Background. NCJ 151172, Washington, D. C.: Bureau of Justice Statistics.
Wagenaar, W.A. 1986. “My Memory: A Study of Autobiographical Memory Over Six Years.”
Cognitive Psychology 18:225-52.
Warner, M., N. Schenker, M.A. Heinen, and L.A. Fingerhut. 2005. “The Effects of Recall on
Reporting Injury and Poisoning Episodes in the National Health Interview Survey.”
Injury Prevention 11:282-87.

5-6