Kensington Research Institute
Services -- Part 2
Implementation
When you are ready to launch your new project, or if your project has already begun, KRI can provide a wide array of services to maximize the quality of the research and evaluation work that you do. Among the many project implementation-related services we offer, KRI staff has had considerable experience with questionnaire design, reviewing questionnaires for thoroughness and selecting or constructing new scale measures, developing coding schema and codebooks, conducting interviewer training sessions, implementing workshops focusing on such topics as enhancing follow-up rates, and performing a variety of data quality assurance tasks. For more information on these topics, please see below.
Questionnaire Design |
Relying upon our many years of experience in questionnaire design,
at KRI, we can help you develop the instrumentation you need in order
to conduct your research and evaluation studies. This might entail
helping you to develop a questionnaire "from scratch" (or just developing
it in its entirety for you). We can also help you to identify specific
modules to add to an existing questionnaire to make it more
comprehensive. For example, perhaps you have a data collection
instrument that you already use, but want to add to it so that you can
capture information on additional topics. We could identify for you any
existing standardized scales that have been created and tested by
others to examine this subject, or we could develop new questionnaire
modules and/or scale measures for you, specific to your study and
research needs.
If, upon completion of the development of this instrument, you decided
that you wanted to pre-test and/or pilot test the newly-designed
questionnaire, KRI could also assist you in designing and implementing
the pre-testing or pilot testing procedures. This could include training
your research team and interviewers to administer the new
questionnaire, designing the pre-test or pilot test, actually conducting the
pre-test or pilot test with whatever persons you wanted to involve,
evaluating the results of the pre-test or pilot test, making
recommendations (as appropriate) for the improvement and/or
modification of the questionnaire, and modifying the data collection
instrument to develop its final form for use "in the field."
Questionnaire / Instrument Review |
Perhaps you have already developed the questionnaire that you plan
to use. Before you begin implementing it "in the field," it is often a good
idea to have an outsider who is an expert in the field of questionnaire
construction review your instrument. This can be an invaluable process
that can identify shortcomings related to the clarity of the questions you
are asking, assess the level of easiness-of-understanding for your
instruments' questions, and discover hidden or unanticipated meanings
that might be inferred by study participants who are asked questions
worded in a certain way. Asking questions that are too difficult or
complex can lead to problems of comprehension, which can only result
in bad or unusable data. Phrasing questions in a way that could lead
them to be interpreted in different ways by different persons is also a
hidden danger in conducting survey-type research. At KRI, our
knowledgeable staff can review your questionnaires to make sure that
these potential pitfalls are identified and remedied before they cause
data-related problems in your study.
Having an outside expert review your questionnaire before you begin
your work "in the field" can also be a good way to identify analysis-related
limitations in the way that study participants are permitted to respond to
various questions. For example, very different things can--and cannot--be
done analytically if people are permitted to respond "never--rarely--
sometimes--usually--always" to a question, versus being asked to respond
to that same question "less than once a year--less than once a month--
less than once a week--less than once a day--daily." Therefore, it is
important to choose a way of responding to questions that best suits
your research needs. Using our broad basis of experience and
knowledge in questionnaire design, we at KRI can help you to make the
decision that is right for you, based on what you are trying to accomplish
with your particular research or evaluation project.
Pre-implementation questionnaire review can also be a valuable process
that can help you make sure that key "pieces" of information or key
topics of interest and importance to your project have not been omitted
from your questionnaire inadvertently. The skilled research staff at KRI can
help with this task, to make it possible for you to get your project off to
the best possible start.
Developing a Coding Schema and a Codebook |
How you decide to allow people to answer the questions you pose to them--that
is, how you decide to code the data that you collect--can have a profound impact
upon the types of research and evaluation questions that you can address and
the types of answers that you can derive from your data. Giving careful
consideration to how you want your questions answered is just as important as
giving careful consideration to how you want those very same questions asked.
KRI can be of great assistance in this process.
In addition, it is usually very helpful to develop a codebook (or a series of
codebooks if multiple data collection instruments are used in a particular project)
anytime you are planning to capture survey-type data. A good codebook not only
lists the questions as they appear on the questionnaire used in the field, but also
such things as previous question wordings (appropriate if more than one version
of a questionnaire has been used on a particular project), response options,
numeric or alpha-numeric codes used for the various response options, special
instructions given to interviewers, special instructions that data entry personnel
should follow, variable names, variable labels, and sometimes other information
as well. Ideally, a completed codebook can be a useful not only to the
interviewers, but also to the persons who conduct the data entry, statistical
analysis, report writing, among others. At KRI, we have had extensive experience
developing and using research codebooks and would be happy to help you
create those needed to complement the work you are doing.
Interviewer Training |
An evaluation/research study can only be as good as the data collected. Part of
the process of collecting good data requires the use of good measures, reliable
and properly-worded questions, and a methodologically-sound questionnaire.
Another key part of the process necessitates the use of properly-trained
interviewers. One thing that we always point out to our clients is that, just
because someone has conducted interviews before does not necessarily mean
that that individual is automatically well-trained to conduct interviews for your
project. Every evaluation and research study is different, with its own
project-specific challenges that must be faced and handled during the interviewing
process. Sometimes, these challenges are the result of the study population
being targeted in your work. Other times, these challenges are the result of the
types of questions you are asking. At still other times, these challenges may be
the result of where your interviews are being conducted (project office versus "in
the field," for example), when your interviews are being held (mornings when
some people may be sleepy versus evenings when people may be distracted by
their day's events), or certain characteristics of the person conducting your
interviews (some people respond differently to men versus women, others may
feel differently about answering questions posed by a younger versus an older
person).
To overcome potential obstacles--ideally, to prevent them from becoming
problems in the first place--it is always wise to train your interviewers before they
begin collecting data for you. KRI can help you with this task. Interviewer training
can take many forms. At its most basic form, there are certain ways that
questions should and should not be asked during an interview. Instructing
interviewers about proper interviewing procedures and providing them with tips
for asking questions effectively is one basic but very element of many interviewer
training sessions.
Oftentimes, it is wise to give interviewers a significant amount of rehearsal time
with their questionnaire(s), so that they can become intimately familiar with the
instrument(s) that they are expected to use. On a project involving the
implementation of a lengthy questionnaire, this may take one or several weeks.
During this rehearsal period, it is important that your interviewers learn how to
answer study participants' queries about the questions being asked of them, that
they can learn how to probe for answers whenever specific responses are not
forthcoming, and so forth.
A good interviewer training process often includes modules focusing on how to
obtain accurate information from people who may be reluctant initially to
disclosing sensitive information, how to look for inconsistent answers given during
an interview (and when logically-inconsistent answers are given, how to "fix" them
so that the inconsistencies are removed), and how to keep study participants
interested in the interview, so that they willingly give thoughtful and accurate
information throughout the entire duration of the interview.
KRI can help you with all of these tasks. If you like, we can conduct full
interviewer trainings for you, taking your staff members from their very first contact
with a particular data collection instrument to their last day of rehearsal and
training. If you prefer, we could also come to your office or project locale and give
your staff a one-day or a two-day workshop about the proper implementation of a
questionnaire, including coverage of the "DOs and DON'Ts" of good interviewing.
One other interviewer training-related service that KRI has provided on several
occasions--and one that we strongly recommend for longer-term research /
evaluation projects--is the booster training. Oftentimes, even though staff
members have undergone extensive and effective interviewer training, over time,
they begin to experience what is known in scientific circles as "interviewer drift"--a
process by which interviewers develop their own individualized interviewing styles,
that, as time passes, take them farther and farther away from optimal interviewing
practices. One effective way of identifying and rectifying this is to provide periodic
booster training sessions. These sessions typically take the form of half-day,
one-day, or two-day trainings, and they can be done one-on-one with individual
interviewers or in group settings with your entire interviewing team in attendance.
During these booster trainings, the rules of good interviewing are reviewed, mock
interviews are undertaken and critiqued, and interview rehearsals are conducted
so that appropriate feedback can be given to the interviewers. Such booster
training sessions are a simple and effective way of minimizing interviewer drift, and
they are an excellent way of enhancing the quality of your project's data
collection.
Another way in which KRI can help you with your interviewing needs is to devise
and implement an interviewer certification process. This is often done when the
person heading up a research/evaluation study wants to make sure that each
interviewer collecting his/her project's data meets a specified minimum standard
of interviewing proficiency. To earn interviewer certification status, after the
completion of the required training regimen, an interviewer ordinarily would have
an interview observed first-hand (ideally) or reviewed on videotape or audio-tape.
The purpose of the observation is for the observer (who would be one of the
project leaders or a KRI staff member assigned to conducting the certification
process) to witness an interview, note the interviewer's strengths and weaknesses,
check the completed data collection form(s) for thoroughness and accuracy, and
make sure that the interview meets the pre-determined minimum proficiency
criteria. Written comments are provided to interviewers after their interviews are
monitored and evaluated, giving them specific feedback about their question
reading skills, clarity, fielding of queries, handling of inconsistencies, and overall
quality of the submitted written document. The written critiques end by indicating
whether the person has earned certification status (indicating that he/she has met
the standards of proficiency established for that study) or whether the interviewer
should submit another interview for review and critique (indicating that the
minimum proficiency standards for certification have not yet been met).
Workshops on Improving Follow-up Rates |
Many research, intervention, and evaluation projects collect data at the time that
people begin participating and then afterwards, at some pre-determined follow-up
point, so that changes occurring in beliefs, attitudes, and/or behaviors between
Time 1 and Time 2 can be measured. It is very important to have high follow-up
rates whenever such Time-1-to-Time-2 changes are to be evaluated. Recent
published evidence by Michael Dennis and colleagues, for example, has shown
that it is difficult to draw conclusions about changes over time if follow-up rates
are below 70%, and that it is far safer (from a scientific standpoint) to do so if
follow-up rates are 80% or greater. Finding people for those follow-up interviews
can be difficult, though, if proper procedures are not put into place when the
study begins, and if interviewers/trackers do not know a variety of ways (some
might call them "tricks of the trade") to locate and keep track of the people they
wish to reinterview.
If you have a plan in place for obtaining your follow-up interviews, KRI can review
your procedures and make recommendations regarding things you might
consider doing to enhance your follow-up rates. If you do not have a plan in
place, we can work with you to develop a strategy that can be used to track your
study participants and maximize your chances of finding them when you need to
reinterview them. We can also conduct a training with your key staff members
regarding this process, including interactive sessions designed to bolster their
effectiveness and increase your follow-up rates.
Data Quality Assurances |
Even when all of the proper training is provided to project personnel, errors
routinely happen in the process of collecting and entering data. Before any
statistical analysis is undertaken, it is always prudent to have someone who is
detail-oriented review the data in search of errors that have occurred.
Sometimes, these errors take the form of a transcription mistake, in which the
person who entered the data from the questionnaire into the computer or data
entry program typed in an incorrect response or number. In other instances,
errors may entail omitting a question or a series of questions, or having a study
participant answer a question or a series of questions that were supposed to be
skipped based on his/her previous responses. Another common data-related
problem is having too many "don't know" or "refused" responses, which more
often than not indicate an insufficiently-motivated interviewer or interviewee rather
than "real" responses indicating that the person truly does not know the answer or
refused to respond to the interviewer's question. Still another data-related
problem occurs when an interviewer inadvertently accepts two logically-
incompatible answers without probing for clarification so that the inconsistency
could be identified.
All of these data-related problems occur easily and, therefore, are commonplace
when quantitative data collection is involved in a research, intervention, or
evaluation study. This is true even when staff members are highly trained and
trying their best to be careful during the data collection process. KRI can help you
with these matters in a number of ways. First, we can conduct spot-checking of
your questionnaires and data files to look for transcription errors. An extra pair
of eyes never hurts! Second, we can review your questionnaires and/or data files
to identify any items that have missing data. Similarly, we can also review your
questionnaires and data files to identify any questions that have too many
unusable answers like "don't know" or "refused," or questionnaire items for which
implausible answers are being given. Not only can KRI bring these types of
problems to your attention, but we could also work with you to conduct briefing
sessions and/or retraining sessions to alert your staff members to the types of
data-related problems that are occurring. Along similar lines, KRI could also
conduct a systematic logical consistency check of your data, to make sure that all
responses given by an individual appear reasonable and, at a minimum, are
logically consistent with one another.