|
Errors reported for Standard Eurobarometer
- as currently distributed by GESIS-ZA -
The following errors in Eurobarometer data sets or documentation have been
detected and/or reported by ICPSR, ZA, or individual users and cannot or
still have not been corrected in currently distributed GESIS-ZA data sets and
documentation. The variables referred to should only be used with caution.
In case of any doubt, please contact your data archive. Complete errata
lists including former data set or codebook editions are
available separately for Standard Eurobarometer
under the respective study profile and for the Central and Eastern
Eurobarometer surveys.
We will always appreciate information about any errors that you find
in Eurobarometer data or documentation.
Please send
mail to the producers.
EUROBAROMETER 5-10 and 16-31A
Variable REGION for Great Britain / United Kingdom:
Should be "EAST MIDLANDS" instead of "NORTH MIDLANDS"!
EUROBAROMETER 12 (ZA study 1037)
Inconsistencies have been reported for the EURO-
BAROMETER 12 "Age" variable with respect to the German
respondents. This variable should be used with caution.
EUROBAROMETER 15 (ZA study 1206)
In variable V29 ("Attitudes if the Common Market had been
scrapped"), data for categories '2' ("Indifferent") and
'0' ("DK, NA") have been collapsed into category '2'.
In variable V15 ("EC Membership - good / bad"), data for
categories '2' ("neither good nor bad") and '0' ("DK, NA")
have been collapsed into category '2'.
EUROBAROMETER 16 (ZA study 1207)
In the case of BELGIUM frequency distribution for V75
(MARITAL STATUS) shows an exceptional high percentage
of respondents LIVING AS MARRIED (10.2 percent compared
to less than 2 percent in former or later Eurobarometer
surveys) and an exceptional low percentage of SINGLEs
(6.1 percent compared to at least over 14 percent in
former or later Eurobarometer surveys). For Belgium
this variable should only be used with caution.
[ZA, January 22, 1998]
Data for Greece in variable REGION (V95) are dubious.
Variable should only be used with caution.
[ZA, January 29, 2004]
EUROBAROMETER 17 (ZA study 1208)
Data for Greece in variable REGION (V250) are dubious.
Variable should only be used with caution.
[ZA, January 29, 2004]
EUROBAROMETER 25 (ZA study 1543)
In EUROBAROMETER 25, dubious frequencies and
inconsistencies have been detected for the "HOUSEHOLD
(FAMILY) INCOME" variables (V296 and V303).
Category '10' in variable V296 has extremely high
frequencies for Ireland and Northern Ireland. These
responses are coded '5' (DK/NA) in the hormonized
variable V303. For unknown reasons responses to
category '10' for Italy and to category '12' for
Luxembourg are also coded to '5' in the harmonized
variable. These variables should only be used with
caution. Please notice also that V296 codebook
category '09' for Denmark ("will not answer") is
actually coded '99' in the data set.
[ZA, September 14, 2000 - Correct data are not available]
EUROBAROMETER 28 (ZA study 1713)
In variable V484 REGION for Great Britain / United Kingdom
data for SCOTLAND and GREATER LONDON are exchanged. Contrary
to the documentation, the cases for Scotland are coded 12 (n=95)
and the cases for Greater London are coded 11 (n=90).
[GESIS-ZA, March 14, 2007]
EUROBAROMETER 29 (ZA study 1714)
Inconsistencies between the "Value Priority" questions
Q.136/Q.138 (V65/V67) and Q.142 (V71) have been reported
for Denmark. The corresponding variables should be used
with caution.
EUROBAROMETER 30 and 31 (ZA studies 1715 and 1750)
In EUROBAROMETER 30, dubious frequencies and
inconsistencies have been detected in the "RELIGION
DENOMINATION" variable (V715) for Belgium and
Germany. Category '9' ("none") shows unusual high
frequencies for Belgium; probably the categories
'9' and '0' (DK/NA) have been merged. For Germany
categories '9' and '0' have probably been exchanged;
usually about 10% of respondents in Germany do not
consider self as belonging to a particular religion.
In general, answers to category '9' ("none") have
erroneously been pre-defined as "missing value"; users
might want to correct this in the SPSS data set.
[ZA, September 14, 2000 - Correct data are not available]
EUROBAROMETER 34.0 (ZA study 1960)
Data for categories 3 and 4 for the Materialism/Post-
Materialism index (V488) have been exchanged. Category 2
should be labeled "Post-Materialist" and category 3 "Mixed"
in the SPSS as well as in the codebook documentation.
Please notice also that all respondents coded as "Missing"
in the basic variables V71 and V72 (Value Priorities) are
coded to "Mixed" in the index variable (V488) which is
not the usual proceeding. It is always strongly recommended
to check the index construction for comparability.
[ZA deliveries after June 19, 2001 are corrected]
Starting with Eurobarometer 34 and up to survey 61, NUTS 1
level data (REGION II) for the NETHERLANDS are not (re-)coded
in accordance with the official nomenclature of territorial
unit statistics. The NUTS 2 level province ZEELAND should be
coded as belonging to NUTS 1 region (landsdel) WEST instead
of SOUTH Netherlands. 
[ZA editions will be corrected from EB 53 onwards, 11.11.05]
EUROBAROMETER 35.0 (ZA study 2031)
Data for VOTING BEHAVIOUR IN LAST NATIONAL ELECTIONS are not
coded correctly for EAST GERMANY. The original codebook and
SPSS documentation was not clear. A correction is possible if
referring to the East German field questionnaire. The following
recode is proposed:
RECODE V237 (51=60) (60=20) (52=30) (62=10) (50=61) (81=50) (61=81).
Other categories are coded correctly. The correct frequency
distribution should be:
<CDU> 320
<SPD> 171
<FDP> 77
<PDS> 60
<DSU> 8
<B-90 / GRUENE> 79
<REPUBLIKANER> 6
[ZA, September 17, 2004]
EUROBAROMETER 37.1 (ZA study 2241)
Original INRA data show identical frequencies for Q32_1
and Q32_2 - V290 and V291 in the ZA version of the data
set. Data of variable Q32_2 (V291) were overrun by Q32_1
(V290) during the processing. Correct data are not
available.
Results to "HEALTH CHECK UP" question Q.21 (V270 to V275)
are probably not correct for Portugal. Categories '1'
(yes) and '2' (no) were exchanged in the Portuguese field
questionnaire. Results are obviously improbable for V272,
V274, and V275.
[ZA, October 23, 2000]
EUROBAROMETER 43.0 and 43.1bis (ZA studies 2636 and 2639)
Portuguese data for INCOME (D.29have been shifted and are
not documented correctly.
The correct categories are as follows:
Answer categories Code
Less than 60000 ESP 1
60001 to 90000 ESP 2
90001 to 120000 ESP 3
120001 to 150000 ESP 4
150001 to 200000 ESP 5
200001 to 250000 ESP 6
250001 to 300000 ESP 7
More than 300000 ESP 8
Refusal 9 (should be recoded to 13)
DK 10 (should be recoded to 14)
[ZA, July 16, 2002; updated April 13, 2004]
EUROBAROMETER 50.0 (ZA study 3085)
In the case of SPAIN variable REGION II (V446 / V448) has not been
recoded correctly from REGION I: 'Baleares'(Code 4 in V433) should
be coded as 'East' (Code 8 in V446 / V448) and 'Cantabria'(Code 6
in V433) should be coded as 'North' (Code 2 in V446 / V448).
[ZA, November 16, 2004]
EUROBAROMETER 54.0 to 57.1
D.29 for ITALY: the income categories 1 to 5 are erroneous. The correct
income categories in accordance with the Italian field questionnaire are:
(1) up to 750.000 L
(2) 750.001-1.000.000 L
(3) 1.000.001-1.500.000 L
(4) 1.500.001-1.750.000 L
(5) 1.750.001-2.000.000 L.
[ZA, July 6, 2005]
EUROBAROMETER 54.2 (ZA study 3388)
The recodes of Q23 into Q23R and Q36 into Q36R are erroneous. Mainly
for Danish respondents missing cases for Q23 and Q36 have been recoded
to valid cases in the respective condensed variable. All missing cases
in the basic variables should be recoded to missing ("NA") in the
derived variables.
[ZA, June 12, 2002]
EUROBAROMETER 54.2 ff.
Starting with Eurobarometer 54.2 (2001)and at least up to Eurobarometer
59 coding for basic British regions in P7_GB and accordingly in all
derived variables (M7_GB, P7R_UK)is erroneous. At least GREATER LONDON
(16) and KENT (18) seem to be exchanged. This variable should only be
used with caution.
[ZA, September 12, 2003]
EUROBAROMETER 55.0 (ZA study 3506)
Data and documentation for Q.15a (NUTS 2 Regions benefiting
from European Union enlargement) are dubious. Data for Q.15a
are collapsed into three of 43 categories (plus 'missing values')
and the respective country specific regions are not documented.
Answer categories for open question Q.15b are not completely
documented. Answer categories (A-M) for Q.65 are not documented.
[ZA, February 14, 2002]
EUROBAROMETER 55.1OVR (ZA study 3508)
For the documentation of answer categories (A-M) for Q.65
(Young Europeans questionnaire) please see corresponding
country specific SHOW CARDS as included in the original
codebook (Q.15 in codebook and data set) in ZA deliveries
after July 16, 2002. Since the correspondence of these
country specific income categories to the SPSS data set
is not clearly documented, this variable should only be
used with caution. This is especially true in the cases
of Ireland and Northern Ireland.
[ZA, February 03, 2003]
EUROBAROMETER 56.3 (ZA study 3635)
Coding errors have been detected for VOTE INTENTION in
variables D4_FRA (France) and D4_NL (Netherlands). For
the time being these variables should not be used.
[ZA, February 20, 2003]
[Corrected in ZA deliveries after March 28. 2005]
EUROBAROMETER 57.1 (ZA study 3639)
Country filters have improperly be used for coding
questions Q.32 and Q.33 (non-EURO / EURO countries).
Variables should be used with caution. An updated
data file will be made available by EORG.
[ZA, October 27, 2003]
EUROBAROMETER surveys 58 and 59
The derived variable of integrated regions for West and
East Germany (7R_GER) is not correct. For German regions
please always refer to the basic (NUTS II) variables
(P7_WGE, P7_EGE) and recode to larger regional units
(NUTS I) if needed.
[ZA, September 12, 2003]
© GESIS Meinhard Moschner
11.11.2005 |
