Annex III: Results of the Cognitive Testing
A general concept question was asked as regards the phrase “in the last five years…” which appears in several questions of the questionnaire. These questions inquire about incidents that happened to the respondent or to a member of his/her household. We inquired whether these questions refer to the current composition of the household (since this may have changed in the last five years). For instance, an incident may have occurred to a member of the household three years before but this particular person no longer lives in the household (e.g. a child who is now an adult and lives on his/her own). Eurostat’s answer was that the term “household” should refer to the current household members, and the people who have left since the incident should not be included in the household. (This remark will also be included in the Instructions’ Manual, which will be prepared for the interviewers).
Section A
Question A2: This question asks about the area in which the respondent lives and gives six different response options (city/town centre, municipal centre, suburb, densely populated area etc.), all of which are very subjective. If we left it to the judgment of the respondent or the interviewer to decide if, for example, the area in question is densely or sparsely populated, there would be many inaccuracies and certainly inconsistencies in the comparison process. For this reason we considered that we ought to have a clearly defined rule for this and that for instance the concept of the Degree of Urbanization (D.O.U.) could be used here since each municipality and community in Cyprus can be classified under one of the three definitions of the D.O.U. (densely populated, intermediate area, and thinly populated area). After consultation with Eurostat, it was agreed to rephrase the question to ask simply for the municipality or community in which the respondent resides, and this will be coded into the four-digit national geographical code in order to determine the D.O.U.
In question A5, the respondent is asked if one (or both) of his/her parents were born in another country. If the answer is “yes” then question A6 asks in which country, however it doesn’t discriminate whether we are talking about the respondent’s father, mother or both. We asked for permission to split question A6 into two parts to read:
(a)In which country was your father born?
(b)In which country was your mother born?
Question A8: The first response option in this question reads “Married or in a registered partnership”. This type of status (registered partnership), however, does not exist in Cyprus like in some other EU countries. After consultation with the experts at Eurostat, we were given the option to omit this response from this particular question.
In question A8, the option “separated” (not yet divorced, i.e. legally still married), does not appear. We decided to include another response option for “separated” and not include the people who fit in this category in the “married” category because there would be problems with the routing of the questions. For example, if the respondent is separated and is classified under “Married” in A8, then s/he will not be asked A9 (whether s/he is currently living with a partner) and the interviewer will go directly to question A10. However, there is the possibility that someone is separated but is living with a partner (other than his/her spouse), in which case A9 should be asked. For this reason, we decided to include a response option (5) for “separated”.
Question A11: As regards the ages of the household members, we will ask their age at their last birthday.
Question A11: We asked if this question can be presented in table format and the coding instructions be placed below each appropriate column (e.g. sex, relation to the respondent). Eurostat’s answer was that this could be done. This will be implemented in the electronic version of the questionnaire, which will be prepared during the next stage of the project.
In question A12, the instructions read “Multiple responses possible” with the reasoning that this could occur for example if someone has two jobs (one as an employee and the other as self-employed). This, however, is not feasible in this case; instead, only one response should be allowed – the person´s major occupation. Otherwise, there will be a problem with the routing of the questions. For example, if the respondent is retired (response number 6) and a full-time employee (1), or if s/he is a student (5) and a part-time employee (2), there will be a problem with the routing. If answer number (6) is selected then the questionnaire will direct you to A14, whereas if number (1) is selected then it will direct you to A13, i.e. the two cannot be chosen simultaneously. Similarly, if one selects number (5), s/he will be directed to A15 whereas if s/he selects number (2), the next question to be asked will be A13; again, these two cannot be selected simultaneously. Therefore, it was agreed to ask for the respondent’s main occupation (as in the Labour Force Survey). Finally, when this question is asked, reference should be made to a specific period (usually this is the week preceding the interview).
Further, regarding the same question, we inquired if we could add two more response options that were missing from this question:
(a)a person who lives off of income from rents for example (a rentier) and
(b)unemployed (a person who has never worked before).
Response option (4) (unemployed) should refer specifically to those who are currently unemployed but who have worked before. Even though both categories of “unemployed” will be summed up in the end, it will be more practical as regards the routing of the questions if we separate the two concepts.
In question A12, response option (7) reads “Permanently disabled”. However, the option “ill for many years” does not appear in any of the responses. There is, of course, the option “Other”, in which such an answer could be recorded. We inquired if we ought to include any who happen to be ill for years (and thus unable to work) in option (7), i.e. “Permanently disabled or ill for many years”. Eurostat’s answer was that “ill for many years” has almost the same meaning as “permanently disabled” and that they should be classified in the same category.
In question A12, if the respondent answers that s/he is on maternity/paternity/parental leave, this implies that s/he still has his/her job (s/he is simply on leave). Thus, this response option should then direct the interviewer to A13 (“What is your occupational title?”) and not A14 (“What was your last occupational title?”).
Question A15 was modified according to the educational system that operates in Cyprus. A few more categories were added:
(a)Never attended school
(b)Did not finish primary school
(c)Technical/professional school
(d)Post secondary education (but not tertiary education)
(e)University (doctorate)
The third category listed above was placed after lower secondary education.
This question will be rephrased to ask for the highest educational level that the respondent has completed, so that only one answer will be given.
For question A16 we suggest that the response category “Don’t wish to answer” be added here, as people tend to be rather sensitive about questions concerning their income. This change has not been implemented in the questionnaire.
Questions A16-A18: These refer to the median income and the lower and upper quartile income. Our question was, for which reference year do we provide figures for these income indicators, and Eurostat’s reply was for the most recent year for which data is available. We provided the data from the 2007 EU-SILC survey (reference year 2006) conducted in Cyprus, and more specifically, the figures are for the “Total disposable household income” (variable HY020). Since the figures were for the reference year 2006, and thus given in Cyprus pounds, they had to be converted to Euro for the purposes of the questionnaire, as our currency now is the Euro (as of 1.1.2008). The conversion (currency exchange) rate used was the one that has been implemented since 1.1.2008.
Section B
Question B3 is asked only if B1 was answered by any one of the responses (1)-(4). Thus, we consider response option (5) in question B3 as redundant. After consultation with the colleagues at Eurostat, we agreed that if B1=5 then B3=5 should be automatically selected as well. After this, the questionnaire will direct the interviewer to question B6.
In question B15(“Which family member or person close to you have you been most worried about being attacked?”) the instructions following this question read “Multiple responses possible”. We consider that there should only be one answer here – to record the one person the respondent is most worried about being physically attacked. For multiple responses to be allowed, we considered that the question needed to be rephrased to read, “Which family member or members have you been worried about being attacked?” (i.e. the word “most” ought to be removed in such a case). After consultation with Eurostat, we rephrased the question so that the respondent can state more than one answers (i.e. more than one persons s/he is worried about being attacked).
The question also arose about the situation where the person the respondent is worried about being attacked lives abroad (e.g. a child who is studying or working abroad). We were advised that these cases of family members (or other loved ones) living abroad, should be recorded. Furthermore, we suggested that there be a sixth response option to this question: “Some other person” (e.g. another relative). In this category, the interviewer could record responses, such as for example a grandmother the respondent is worried about being attacked because she lives in a dangerous neighbourhood. Our suggestions were accepted and these modifications were implemented in the translated questionnaire.
Section C
Questions C1-C6 needed to be clarified as we considered that they all referred to the same concept, but some use a different wording than others. Some use the phrase “had use of” whereas others use the word “had” (i.e. owned). Eurostat’s reply was that the question is meant to include vehicles on leasing, company cars etc, as well as vehicles that are loaned assuming that this is a fairly long-term arrangement. Therefore, we changed the wording in these questions so as to have consistency.
Questions C9, C12, C24 and C27 refer to theft of a motor vehicle or to the possibility of the motor vehicle having been driven away without permission, both in the same question. We presumed that if the vehicle was driven away without permission but was returned (because the person who drove it away did not have the intention to keep it), then this defines a different offence than theft (joyriding) and inquired whether it was correct to include the two in the same question. Eurostat’s reply was that, in the case of joyriding, there is usually damage also to the car (sometimes a complete destruction occurs), not to mention inconvenience/loss of time to the owner, and we were advised to leave the two grouped in the same question.
There seemed to be an inconsistency between question C2 and questions C9 and C12: question C2 refers to cars, vans or trucks. We presumed that any one of the three applies here, and so the parenthesis was removed. On the contrary, questions C9 and C12 refer only to cars and vans. We changed the wording in the latter two questions (to include “trucks”), so as to maintain consistency.
Regarding question C47, we believe that the word “intentionally” was accidentally omitted. We included this word in the translated questionnaire.
Following question C46, the instructions read, “Pick pocketing in the next question”. Since this instruction refers to questions C49-C53, we changed the phrase “in the next question” to read, “in section 10 (questions C54-C58)”.
In several questions in this section (C10, C15, C20, C25, C30, C35, C40, C45, C50, and C55) and a few in section E (E2, E21, E31, E51, E61 and E65), the response options are: (a) this year (2008), (b) last year (2007) and (c) earlier (2003-2006). Since CYSTAT will conduct this survey within the first few months of 2009, we suggested that we adjust the options in the above-mentioned questions as follows: (a) in 2008, (b) in 2007 and (c) earlier (2003-2006). Our suggestion was accepted by Eurostat. (Other questions affected by this modification are D1, D2A and D2B).
Section D
We found question D1 rather unrealistic as it asks if any of the crime types were very similar incidents, where the same thing was done under the same circumstances and by the same people. This is in fact a highly unlikely scenario – for the same offence to happen to the same person under the same circumstances and by the same people.
Immediately preceding questions D1, D2A, D2B and D2C, we find the phrase “during this and last year…” Since the survey will be conducted in 2009, and in order to be consistent with the related questions in section C, we suggested that the phrase be modified to read “during 2008 and 2007”.
In questions D1 and R4, we added two more response options: “I don’t know” and “I don’t remember” and in question R7 we added the response “I don’t know”, after consultation with the colleagues at Eurostat.
About questions D2A, D2B and D2C, we inquired whether the year in which the incident took place should also be asked here (besides the month) and Eurostat´s response was affirmative.
In the case that one of the offences mentioned by the respondent is robbery, both questions D3 and D4 will have to be answered. However, these are similar questions and in fact, the first response options are identical in both of them. We confirmed with Eurostat that if D3=1 then the computer program will automatically select D4=1 (i.e. it will not ask question D4) and it will move directly to the next question for this crime type, which is D8.
Regarding question D3, the following question arose: if the fifth response is selected, (i.e. the particular crime type took place abroad), will the remaining questions in this section be answered, i.e. are we also interested in the crime incidents that took place abroad? Eurostat´s reply was affirmative.
Also in question D3, we found response option (3) (elsewhere in your city or local area) rather vague. We suggested that option (2) should be rephrased to read “Near your own home or in your local area” (i.e. the phrase “or in your local area” was added here) and option (3) should read “Elsewhere in your city” (i.e. the phrase “or in your local area” should be removed). Our suggestion was accepted by Eurostat and the change was adopted.
Response option (6) of question D4 states “In hotel, other commercial accommodation”. It would be helpful if we were given clarifications of what the term “commercial accommodation” includes, so that we can use it in our Instructions’ Manual, which is being prepared for the interviewers.
In question R11, the third response option reads, “punched or slapped” and the fifth response option reads, “hit with fist”. We pointed out that “punched” and “hit with fist” are synonymous terms and after consultation with Eurostat, we moved the word “punched” to response number (5) (along with “hit with fist”) and we left response option (3) to read “slapped”.
In question D9, the second response option refers to a “family member”, whereas the third and fourth options refer to “household members”. (Also, similar questions make reference to the respondent´s household members and not family members). We considered that the most appropriate term here, so as to maintain consistency, would be “household member”.
In question D10, the first response was changed to read “car, van, truck” (the word “truck” was missing) in order for it to be consistent with the relevant questions from the previous sections.
In question D10 (“Which item or items were actually stolen?”), the first response option is “car, van, truck”, the second is “motorcycle, moped, scooter” and the fifth one is “Bicycle”, even though the instructions before this question read: “If D0=2 or 6-7 or 9-10, ask:”. In other words, if D0=1 (theft of car) or D0=4 (theft of motorcycle, scooter, moped), or D0=5 (bicycle theft), then this question must not be asked. This is of course inconsistent, and we pointed this out to the colleagues at Eurostat who confirmed that there was a mistake and that the correct filter for this question should be “If D0=1-2 or 4-7 or 9-10”.
Also, in questions D10, D12, Q30 and Q105, response option (51) reads “Parts of dwelling or other building (e.g. windows, doors)”, whereas options (60)-(63) read “Wall”, “Window”, “Door” and “Other parts of the building” respectively, i.e. there is repetition of responses here. Eurostat´s reply was that responses (61) and (62) should be omitted. This, however, could not be done as the filter for question D13 is, “If D12=61, 62 or 63, ask:”; instead we deleted response category 51 from these questions as it was redundant.
Question D18 asks about the amount of money that the respondent received from the insurance company, however, it is possible that the case is still in the process of investigation by the insurance company and that the money was not yet paid to the victim, even though it is expected that an amount will be paid by the insurance company. Perhaps the question should be rephrased in order to account for such cases as well.