Web surveys and official statistics: The case of the Italian Economic Censuses in Tuscany

Transcription

1 Web surveys and official statistics: The case of the Italian Economic Censuses in Tuscany Luca Faustini, Sabina Giampaolo, Alessandro Valentini 5 th Conference of European Survey Research Association Ljubljana, Slovenia, July 15-19, 2013

2 Summary 1. Aim of the work 2. Short description of the census process 3. Response rates: total and by collection channel 4. Some features of the web respondents 5. Weekly distribution of response rates 6. Directions for further research

3 Aim of the work: Our work focuses on the use of web for the Italian Economic Censuses (Enterprises and Non Profit Institutions) in the territorial area of Tuscany Main goals: To provide an idea of the process quality improvements due to the use of the web To identify the differences between the characteristics of web and paper respondents To evaluate the timing of forms return ( web vs. paper ) Warning: The work is based on provisional data Data refer only to the process 1

4 Focus on the census process: Territorial organization Three levels: National (Nuts 0) Regional (Nuts 2) Provincial (Nuts 3) Regional Office of Census ( in Tuscany) 15 census experts Supervision of 10 Provincial offices (around 190 operators: 300 units to detect for each operator) Training to the network involved Technical assistance 2

5 Focus on the census process List-based surveys: List of units to be detected based on the integration between statistical and administrative sources Postal sending: Mail-out of census forms Multimode data collection: Respondents may chose to return questionnaires in the following ways: Internet Delivery to Nuts 3 data collection centers (UPC) Delivery to Post Offices Enumerators Web based Paper based On line system: immediate supervision of all census operations 3 Two different censuses: Enterprises Non profit units Specific peculiarities

6 Focus on the census process: enterprises Starting list of units to detect: Around 19,200 enterprises (7.4% of Italian total): All the cases with 20 employees and more A sample of around 4% of enterprises with less then 20 employees (stratified by economic activity, classes of employees and Territory, at Nuts 2 level) Robust list of units (derived from ASIA, a well-established Italian statistical register) Innovations: Multipurpose surveys New topic such as competitiveness and decisional processes 4

7 Focus on the census process: Non Profit units Starting list of units to detect: Around 36,000 non profit units (7.6% of the whole national list): All the potential units derived from the statistical and administrative archives low quality of the list (enumerators had to verify through field operations the presence of new units) Innovations: Well structured questionnaires Deep attention on the topics of organization and users of offered services 5

8 Focus on the census process: the timing Preliminary operations Spontaneous restitution Required restitution Census s day Post Offices WEB Provincial data collection centers sending Postal Enumerators 31 st dec th sep st oct th feb 13 1 st sep12

9 A quick overview of the results Front Office data collection

10 Collection channels 8

11 The use of web: a comparison between Tuscany and Italy 9

12 The use of web: Enterprises by class of employees Mean: 74.8% 10

13 The use of web: Enterprises by juridical typology Mean: 74,8% 11 Probably better organized enterprises

14 The use of web: Non Profit institution by juridical typology Aptitude for the use of the web is quite similar for the various typologies! Mean: 53,9% 12

15 Weekly cumulative distribution of total responses 13

16 Web quota of responses by week Faster starting reactions Slower achievement of steady state 14

17 Estimated weekly response rates: web vs. paper - Enterprises Strong intervention of Regional Office Winter holidays Certified Certified Postal reminder 15 Spontaneous restitution First part of required restitution Second part of required restitution

18 Estimated weekly response rates: web vs. paper Non Profit Strong intervention of Regional Office Winter holidays Postal reminder 16 Spontaneous restitution First part of required restitution Second part of required restitution Titolo intervento, nome cognome relatore Luogo, data

19 Directions for further research Preliminary results illustrated are somewhat interesting. They encourage our actions in the following ways: deeper analysis of the web respondents (individual characters coming from micro-data) study of the Tuscany case where the web use has been lower than for Italy geographic analysis of web responses Thanks for your attention! Luca Faustini, Sabina Giampaolo, Alessandro Valentini, 17