|
Stigma
In Global Context: Mental Health Study Indiana University . Bloomington . Indiana |
|
|
Hungary Sample Design Sampling Plan 1.
Operating method of data collection department
Translation
TÁRKI has got the English version of the actual questionnaire. An
expert will prepare the translation of the questionnaire from English to
Hungarian. Another translator
will make the back-translation into English. An
individual who has not seen the original version of the questionnaire will
prepare the back-translation. All
our questionnaires are carefully designed.
Experts of all kind participate in the process: experienced
interviewers, instructors, and researchers. The
draft questionnaires are always piloted in order to reduce mistakes and
ambivalence. Great emphasis is
laid on trying to prevent mistakes that cannot be corrected later. Sampling:
We
use probability samples that had been selected in multiple stages with
proportional stratification. In
the first stage, localities are chosen, in the second stage, the
respondents are chosen from the localities with simple random sampling. Creating
the sample of localities: The counties
formed the first strata, type of settlements the second. The localities
are chosen from these strata with the help of random sampling. During
this we usually follow the general principle that Creating
the sample of individuals: After
the selection of the localities, the demanded number of respondents is
chosen in every towns and villages. We
use “preselected
addresses” sampling methods for selection of respondents. The
number of respondents in the previously chosen localities is defined in
accordance with the proportion of the population in the given strata –
counties and in the different types of localities (county seats, towns and
villages) within the counties. The
concrete names and addresses are obtained from the Central Registry and
Electoral Office with the help of simple random sampling. First contact:
Knowing
the names and addresses of respondents prior to the actual data collection
procedure request letters are sent to participants in order to inspire
confidence and reduce the proportion of the refusal rate. Instruction of
interviewers:
Written
instructions are prepared for each research for the field workers. Instructions
include the brief description of the research's aim to arouse the interest
of the interviewers, explaining the purpose and structure of
questionnaires; it also gives guidelines rarely occurring interview
situations. Fieldwork and control:
TÁRKI
developed a national network of field workers; it is headed by a full-time
job leader-instructor to whom regional instructors report. The
chief instructor is responsible for all the fieldwork, including the
control of regional instructors. The
regional instructors are responsible for the fieldwork in their region –
one or two counties. Careful
selection is needed to find the best interviewers and they have to be
provided systematic in-service training. Fieldworkers are controlled in many different ways.
The regional instructors
manage the interviewers' work; they are responsible for solving problems
interviews encounter. Apart
from this their task is to carry out 100 percent fill-in control, which
includes handling the possible suspicious cases individually. Besides
these they control all interviewers at random by visiting the participants
(10%). The regional instructors hand
over the questionnaires to our post-fieldworkers working at the Besides of this regional control, a central controller is employed in
TARKI, who has not daily personal connection with the interviewers. After
the results of logical
control of data based on computer, and the reports of post-fieldworkers,
she conducts an additional control of interviewers or regions where the
mistakes or problems seem to be frequent. She
controls personally or with the help of phone, or with letter of
checking. She is responsible
for inspect the controlling work of regional instructors as well. Data processing and
cleaning:
Data
is entered in SPSS Data Entry program. Logical
control and exclusion of extreme values are always built in the program.
This procedure reduces typing errors and other possible mistakes. Cleaning
of data occurs at different stages. Before
compiling the system file a program pre-checks for various types of coding
error. Only after correcting
these errors are the files joined. On
the joined data file a more rigorous cleaning procedure begins: all
demographical data are cleaned and logical uncertainties are checked and
re-checked. The handed down
files are joined and we start the cleaning of the basic data and the
logical controlling. Employees
experienced in searching for mistakes do this part of the operation and
they also understand programming. Weighting
The goodness of fit of the sample is controlled with the common consideration of the main social-demographic features of the respondents. The expected social-demographic frequencies are determined by the latest relevant database of the Central Statistical Office. The sample is weighed by considering the common frequencies of four demographic factors:
Documentation:
Codebooks about every fieldwork are complied on each research. It contains the frequencies and descriptive statistics on each variable as well as information on the exact sampling method and on the comparison of data to current census. All materials related to fieldwork are also included.
|
|
|
1022 E. Third Street, Bloomington, IN 47405 (812) 855-3841
Last updated: 6 September 2005 |
||
