Use of Multi-Mode Methods in Census Data Collection

Transcription

1 Use of Multi-Mode Methods in Census Data Collection Workshop on Population and Housing Censuses for countries of Eastern Europe, Caucasus and Central Asia (Geneva, 2-3 October 2017) Prepared by Diana Beltadze and Ene-Margit Tiit from Statistics Estonia

2 Content Multi-mode methodologies in census Using registers in census; Using internet in census; Using face-to face interview as a part of MMM Advantages of MMM Technology Adoption Lifecycle and respondent s behaviour Quality aspects IT aspects Conclusion

3 MULTI-MODE METHODOLOGIES IN CENSUS

4 Which are multi-mode methodologies (MMM)? Instead of face-to-face interviews the multi-mode methodologies (MMM) use different modes for census data collection, for instance: Using registers Self enumeration via internet or using paper Telephone interviews. Face-to-face interviews can be made using paper questionnaires, laptops or other electronic devices. It is possible to use GPS s to fix the coordinates

5 Different dimensions Using different data sources some data may be taken from registers and not asked from respondents personally, some data may be asked personally. Response method self-enumeration, personal interview and proxy interview are used at the same time. Response technology via internet and using computers, paper questionnaires may be prepared as backup in order to reduce risks.

6 Considerations for using multi-mode methods for data collections Benefits (response rate, coverage, data quality, etc) Factors which might have influence on multi-mode data collection (how to avoid duplication, data comparability issues, respondent burden, complexity of processes, recruitment and workload balance of enumerators, etc) Response integration management( integrating data from all collection systems, etc)

7 Combination of methodologies In the case of MMM s usually different methodologies are combined, for instance: Registers and face-to-face interview; Internet and face-to face interview; Self-enumerating using paper questionnaires and face-toface interview; Telephone and face-to-face interview; Other possible combinations.

8 Why to use MMM in census? Census is always a very big and expensive organization. All countries are looking for the ways how to make censuses more effectively, how to waste less resources, assure higher quality and save the time. With this aim a series of new methodologies have elaborated. But using the first time any new methodology is not an easy and cheap thing.

9 Which MMM is the best? The answer depends on the conditions in the country. If there exist good registers and all people have ID-codes, it is possible to use registers (as one component of MMM); If most of the country is covered with internet and the people have ITabilities on high levele, then use internet; If most of the country is covered with good telephone connection, then the use of telephones is possible. If the people are ready for self-enumeration, but no good internet nor telephone connection exists, use mailing

10 Components of MMM There do not exist ready recipes how to combine different methodologies to get MMM suitable for any country. In the following we introduce the main methodologies, that can be used in MMM, showing their premises and advantages. As a result, each country can find the suitable combination himself. We encourage to use new methodologies also in case, if they are usable only for small part of population to be enumerated.

11 USING REGISTERS IN CENSUS

12 Using registers the most tempting way It seems very tempting to organize census without visiting homes and knocking the doors, using only registers. But this way is achievable only for quite few countries. For this aim the country must have a developed system of registers. All persons, enterprises (including farms) and all houses and dwellings must have identification codes that are used in all registers. But there exist also another way. Otherwise, it is possible to use registers as a part of MMM.

13 Using a register in MMM premises For using a register it is necessary that all residents are identifies using ID-code or some other unique identifier. Before using a register in census the register must be thorough checked: Coverage; Correctness (no mistakes, nor errors, no blanks); The format of data answers to demands of census; Registers are regularly updated; It is also necessary to have the technology for transferring data from register to census data.base.

14 Two ways how to use the registers for census (MMM). 1. Questions not asked. 1. Take the answers to some questions from registers and put the information directly to the census data base. This is possible only in case when the coverage and quality of registers are very good. In this case you save interviewing time, Data entering data checking time.

15 Two ways how to use the registers for census (MMM). 2. Prefilling 2. Use the registers for prefilling questionnaires. If your registers quality is not high enough, you can still use it for prefilling the questionnaires. Then during face-to-face interviews the enumerators can check the information taken from registers. As a result, the census data will have better quality and also you have the basis to improve the register data.

16 USING INTERNET IN CENSUS

17 How to use the internet for census? I Premises for using internet: 1. Coverage of (at least a part of) country with internet connection; 2. Habits and skills of internet usage in population; 3. Existence of computers/ tablets in the usage of population; 4. Existence of open internet points (cafes, libraries); 5. Sufficient volume of internet channels; 6. Positive motivation of population.

18 How to use the internet for census? II 1. Preparation/motivation of population for using internet in census must start early (at least one year before); 2. The census calendar must contain special different time periods for internet and for face-to-face interviews. 3. The safety and security of internet census must be guaranteed; 4. The access to census questionnaires in internet must be checked using safe identifiers.

19 Internet census means self-enumeration Self-enumeration can be done also using paperquestionnaire. Using internet (web-based) questionnaire there are more possibilities to assure the high quality of answers, using On-line checks; Thorough explanations and helps; Checking, if the answering has been finished.

20 Questionnaire for self-enumeration Self-enumeration questionnaires can be sent via internet or by mail usual mail or can be picked up by enumerators. In any case the main issue is the clarity of questions. Some definitions may be complicated then it is important to add to questionnaires explanations. If possible, it is useful to use the menus so that the respondent can choose the adequate answer. Logic of moving in questionnaire must be as simple as possible, especially in the case of internet questionnaire.

21 Web-based solution`s advantages lower cost of the resources due to a smaller number of interviewers; those who refuse to give an interview can fill in an online questionnaire and no interviewer will be sent out; interviewer s subjective error will not be present in online questionnaires; higher response rates (the respondent fills in the questionnaire whenever convenient); questionnaire corresponds entirely to the respondent s profile, i.e. does not include unnecessary questions; possibility to get the answers from respondents who are very mobile.

22 Web-based solution`s disadvantages: Disadvantages: higher development costs; costs to ensure the functioning of system; in case of a mixed method, the method can affect the answer, and that influence requires separate analysis; poorly defined universe that covers the undefined part of the population and does not allow adequate assessment of the involvement probabilities; there is no evaluation of the response activity.

23 USING FACE-TO FACE INTERVIEW AS A PART OF MMM

24 Face-to-face interview (CAPi or PAPI) For face-to-face interview there are several possibilities today: to use computers or some other devices as tablets (CAPI) or paper questionnaires (PAPI). It seems that CAPI is more expensive, but in many cases that is not true. Using computers saves: Human resources in the time of interviewers; Human resources in the time of data entering and checking; Paper (handling, saving and transporting the paper questionnaires). Assures higher quality due to online controls.

25 Face-to-face interview as part of MMM Face-to-face interview is an obligatory part of MMM census, as the all population cannot self-enumerate themselves (problems with health, access, motivation etc) Face-to-face interview has two aims: To enumerate the population not enumerated using other ways; To check the correctness of self-enumeration (only in few cases). In the case when also space coordinates should be defined, participation of enumerators is necessary (this is the only way to enumerate empty dwellings).

26 Advantages: CAWI + CAPI mixed method lower time and resource costs concerning the entering and checking of data; possibility to optimize respondent s work with the help of the program (guidelines, a hierarchical list of answers); possibility to anticipate errors in the course of the interview by using logical checks. Disadvantages: risk of technical problems; quality.

27 ADVANTAGES OF MMM

28 Adventages of MMM Saving human work; Assuring higher quality of results; Possibility to get higher response rate and coverage; Higher flexibility in communication with respondents; Saving resources connected with handling paper questionnaires; Saving time in data-processing; Being more advanced and developed.

29 TECHNOLOGY ADAPTION LIFECYCLE AND RESPONDENT`S BEHAVIOUR

30 Technology adoption lifecycle by G.Moore

31 Respondent s behaviour contact rate non-contact rate cooperation rate refusal rate refusal conversion rate non-response rate (The results of the REL2011 in Estonia refer to a hypothesis that in order to achieve ~30% response rate is possible to get but it demands for deep studies of respondent s behaviour as well as questionnaire evaluation).

32 Solutions Behaviour through learning of respondents: cooperation; simplicity; easy to use; speed; efficiency through feedback; quality; security; support.

33 Goals for the public campaign To introduce the purpose of the census. To explain how to take part of it. To create positive attitude among the publics. To encourage people to complete census form online. To make census nationwide and relevant to everyone. 2-3 October, Workshop 2017in Geneva

34 QUALITY ASPECTS

35 Quality aspects Product quality Relevance Accuracy Timeliness and punctuality Comparability and coherence Accessibility and clarity Process quality Best methods Cost efficiency Low response burden

36 Reasons of under coverage Mobility of people Not willingness to give personal data to authorities Problems with enumerators Special Estonian problem(but not only): living in places that do not coincide with the registered addresses.

37 Lessons of the last PHC 2011 in Estonia We convinced that people are ready to use internet for different procedures, also when it is time-consuming and not very easy. To get good results in large-scale nationwide actions the long-lasting campaign and transparency of actions are necessary. The big problem of contemporary censuses is the mobility of people. In Estonia up to 20% of people do not live in their registered residence. Also the problem is the identity of addresses it was greatly solved in Estonia by new address-standard. You cannot believe all technical solutions for 100% (e.g. existing internet-connection). It is reasonable to have plan B

38 IT ASPECTS

39 Goals of the software Universal software useable in all surveys Improve the coverage of studies Improve quality of raw data Improve efficiency of field work management Reduce data collection cost Increase speed of post-collection data processing

40 Environment For census needs may be created general-purpose systematic software (environment) that will be used in future for surveys, processing, analysing and combining data from different sources. The environment should be thoroughly tested

41 IT related risks Data leakage Attacks against IT solutions Load tolerance is not sufficient Failures of software Data loss

42 Load tolerance tests Test requirements E-respondents application Enumerators application Use of emulating robots

43 Other functionalities Messages Overviews suitable as accounting documents for interviewers salaries and driving compensation Opening filled questionnaires by interviewer to check quality, comments by soft controls Help desk Opening questionnaires by filling CAWI respondent at the time of answering to help if needed (help desk)

44 Other functionalities (2) Assignment sample units with changed addresses automatically to a new interviewer Opening and actuation of defined requests Redivide sample units between Interviewers during the fieldwork Map for supervisor,which shows location of buildings, where sample units locate and statuses 2-3 October,

45 Authentication with ID-card, Mobile ID bank link codes

46 The need of resources is different in the case of different data-collection methodologies.

47 Conclusions When a country wishes to use MMM, it should revise its possibilities and resources in existing registers, having internet coverage, IT habits and skills in population etc. Registers can be used as additional tool for prefilling the questionnaires also in case when they are not excellent. Starting with internet + interviews is possible also in conditions when the expected share of internet users is quite small (e.g. 10%). Nowadays in most countries the population has so high education level taht self-enumeration is not problem.

48