===========================================================================

                   COMPARATIVE STUDY OF ELECTORAL SYSTEMS (CSES)
                                 www.cses.org

                             MODULE 6 (2016-2021)
                          METHODOLOGICAL GUIDELINES

                            - January 25, 2022 -
                              
                              
===========================================================================                                                                                          

Collaborators in the Comparative Study of Electoral Systems shall 
adhere to the following standards of data quality:

a. Mode of interviewing: Interviews should be conducted
face-to-face, unless local circumstances dictate that telephone,
Internet, or mail surveys will produce higher quality data. 
Mixed-mode surveys are acceptable to increase response rates
and/or compensate for undercoverage associated with particular
survey modes. In cases of mode variation as well as in cases of
within-mode variation (e.g. adaption of Internet surveys to 
multiple devices) presentation of questionnaires to respondents
should be as similar as possible. All variation within surveys
shall be documented in detail, and technical information on
survey mode and, where appropriate, device used shall be
identified in the data set for each respondent. National studies
should seek to keep mode changes to a minimum to maximize
comparability of their data sets across modules.

b. Timing of interviewing: We strongly recommend that collaborators
in the Comparative Study of Electoral Systems conduct their
interviews in the weeks following their national election. 
Interviewing should not start later than six months after the
election. Out of concern for data quality, data collection
should be completed in as timely a fashion as possible. In the
event of a runoff election, interviewing shall be conducted 
after the first round election. The date of interview shall be
provided for each respondent.

c. Placement of module in post-election questionnaire: The
questionnaire module should be asked as a single, uninterrupted
block of questions. We leave it to each collaborator to select
an appropriate location for the module in their national survey
instrument. Collaborators should take steps to ensure that
questions asked immediately prior to the questionnaire module
do not contaminate the initial questions in the module.
Collaborators are also free to select an appropriate place in
their survey instrument to ask the turnout, vote choice,
and demographic questions.

d. Population to be sampled: National samples should be drawn from
all age-eligible citizens. No sampling frames with systematic 
undercoverage of significant population groups (such as citizens
without access to the Internet) are acceptable. When 
non-citizens (or other non-eligible respondents) are included
in the sample, a variable should be provided to permit the
identification of those non-eligible respondents. When a
collaborator samples from those persons who appear on voter
registration lists, the collaborator should quantify the 
estimated degree of discrepancy between this population and the
population of all age-eligible citizens. Studies based on
panels or access panels are acceptable if dictated by local
circumstances. In such cases the collaborator should seek to
minimize the time lag between initial sampling and the CSES
survey and quantify the estimated degree of discrepancy to the
population of all age-eligible citizens and provide weights.
Details about initial sampling must be documented.

e. Sampling procedures: We strongly encourage the use of random
samples, with random sampling procedures used at all stages of
the sampling process. Collaborators should provide detailed
documentation of their sampling practices for all stages.

f. Sample Size: We strongly recommend that no fewer than 1,500
age-eligible respondents be interviewed, and under no 
circumstances should fewer than 1,000 age-eligible respondents
be interviewed.

g. Interviewer training: Collaborators should pre-test their survey
instrument and should train interviewers in the administration
of the questionnaire. The Planning Committee will provide each
collaborator with documentation that clarifies the purposes and
objectives of each item and with rules with respect to probing
"don't know" responses.

h. Field practices: Collaborators should make every effort to
ensure a high response rate. Investigators should be diligent in
their effort to reach respondents not interviewed on the initial
contact with the household and should be diligent in their
effort to convert respondents who initially refuse to
participate in the study. Data on the number of contact
attempts, the number of contacts with sample persons, and
special persuasion or conversion efforts undertaken should be
coded for each respondent.

i. Strategies for translation (and back-translation): Each
collaborator should translate the questionnaire module into
their native language(s). To ensure the equivalence of the
translation, collaborators shall perform an independent
re-translation of the questionnaire back into English.
Collaborators engaged in translation of the questionnaire module
into the same language (e.g., Spanish, French, English, German,
and Portuguese) should collaborate on the translation.          

//END OF DOCUMENT