This document discusses statistics and democracy in Sweden. It provides statistics on voter turnout in Swedish elections from 1911 to 2010, showing that turnout was historically higher and has declined over time. It also shows that women and those with lower levels of education have lower rates of voter turnout. Additionally, it presents data on the increasing percentage of women who have been elected to political positions in Sweden from 1970 to 2010. In conclusion, it notes some current trends in statistics and questions whether national statistical offices are ready to play a more active role in public debates.
This document summarizes the results of an interview study conducted with 11 non-respondents to government surveys in Sweden. The interviews explored reasons for non-response from a qualitative perspective. Some key findings included:
- Some saw value in contributing if they understood how it could make a meaningful difference, while others did not feel it was important or would lead to anything good.
- Attitudes toward authorities and power influenced responses, with some expressing mistrust while others believed the government wants to do good.
- Experiences and knowledge of statistics varied, with some skeptical due to how questions are asked or outdated data is used to draw conclusions.
- Three types of non-respondents were identified based on
The document discusses international statistics capacity building efforts. Over $2.3 billion was spent on statistics development from 2010-2012, with most funds coming from key providers like multilateral organizations. The purpose of capacity building is to strengthen national statistical institutes and develop the competences of their staff to produce high-quality statistics according to international standards. The document advocates for mutual competence development where competences are built for both provider and recipient institutions. It suggests competence development in areas like intercultural skills, social skills, organizational skills, personal skills, and technical skills can benefit staff and make capacity building efforts more cost-effective.
Kompetanseutfordringer ved restrukturering av svar- og veiledningstjeneste i SSB og hvordan jobbe aktivt med å utvikle medarbeidere mot slutten av arbeidskarrieren
This document discusses trends and challenges facing Nordic statistics and suggests areas for potential Nordic cooperation. It notes that while statistics have historically played an important role, their influence may diminish as more actors share information. Two potential cooperative projects are outlined: 1) Creating scenarios about the future of Nordic statistics looking out to 2033 to understand uncertainties and opportunities. 2) Conducting horizon scanning to identify new emerging issues that need to be addressed. The goal is to help statistics remain relevant by anticipating changes and renewing methods and content.
This document discusses statistics and democracy in Sweden. It provides statistics on voter turnout in Swedish elections from 1911 to 2010, showing that turnout was historically higher and has declined over time. It also shows that women and those with lower levels of education have lower rates of voter turnout. Additionally, it presents data on the increasing percentage of women who have been elected to political positions in Sweden from 1970 to 2010. In conclusion, it notes some current trends in statistics and questions whether national statistical offices are ready to play a more active role in public debates.
This document summarizes the results of an interview study conducted with 11 non-respondents to government surveys in Sweden. The interviews explored reasons for non-response from a qualitative perspective. Some key findings included:
- Some saw value in contributing if they understood how it could make a meaningful difference, while others did not feel it was important or would lead to anything good.
- Attitudes toward authorities and power influenced responses, with some expressing mistrust while others believed the government wants to do good.
- Experiences and knowledge of statistics varied, with some skeptical due to how questions are asked or outdated data is used to draw conclusions.
- Three types of non-respondents were identified based on
The document discusses international statistics capacity building efforts. Over $2.3 billion was spent on statistics development from 2010-2012, with most funds coming from key providers like multilateral organizations. The purpose of capacity building is to strengthen national statistical institutes and develop the competences of their staff to produce high-quality statistics according to international standards. The document advocates for mutual competence development where competences are built for both provider and recipient institutions. It suggests competence development in areas like intercultural skills, social skills, organizational skills, personal skills, and technical skills can benefit staff and make capacity building efforts more cost-effective.
Kompetanseutfordringer ved restrukturering av svar- og veiledningstjeneste i SSB og hvordan jobbe aktivt med å utvikle medarbeidere mot slutten av arbeidskarrieren
This document discusses trends and challenges facing Nordic statistics and suggests areas for potential Nordic cooperation. It notes that while statistics have historically played an important role, their influence may diminish as more actors share information. Two potential cooperative projects are outlined: 1) Creating scenarios about the future of Nordic statistics looking out to 2033 to understand uncertainties and opportunities. 2) Conducting horizon scanning to identify new emerging issues that need to be addressed. The goal is to help statistics remain relevant by anticipating changes and renewing methods and content.
The document discusses the establishment of a Nordic PIAAC Database containing data from the OECD's Survey of Adult Skills (PIAAC) and national registers from five Nordic countries and Estonia. The goals of the Nordic PIAAC Network, which is financing the database, are to create the database, publish a report on skills in the Nordic region, and make the microdata available for other research. The database will contain background questionnaire data, cognitive scores, and register data from national sources and will be housed securely at Statistics Denmark with researchers applying for access. Combining survey and register data provides more information and allows the database to be updated over time.
This document discusses the need to measure uncertainty in register statistics given their inductive nature. It proposes using Confidence Images (CIm), which restrict potential values of population parameters based on multiple sources of information, each with an associated measure of uncertainty. As a proof of concept, the document calculates CIms for the total number of cattle in Sweden using different combinations of register data, survey data, and prior knowledge. The CIm approach fulfills criteria for providing objective, interpretable and comparable measures of uncertainty for register statistics. Further methodological development is still needed, but CIms could become a useful statistical tool for quantifying uncertainty in register-based estimates.
This document summarizes a presentation given by Maija Metsä-Pauri of Statistics Finland about statistical communication and quality in the age of open data. It discusses how user needs are changing with digitalization and social media, requiring a shift from data dissemination to active communication. Statistics Finland has launched new interactive websites and a blog to better engage with readers. It also emphasizes the importance of statisticians telling statistical stories, collaborating with data journalists, and focusing on quality over quantity of information as open data becomes more prevalent.
Statistics Finland is developing web data collection for personal surveys through a set of projects running from 2012 to 2014. The goals are to develop infrastructure for web data collection and support mixed-mode data collection, especially for the Labour Force Survey. Four initial projects have been completed, including a mixed-mode pilot of the Consumer Survey. Current projects include developing Statistics Finland's data collection pages and a 2013 web pilot for the Labour Force Survey to study mode effects. By 2017, an internet questionnaire is planned to be available as an alternative for all personal surveys.
This document discusses non-response and motivation measures in Statistics Finland's EU Labour Force Survey from 1997-2013. It summarizes trends in non-response rates over time, with rates peaking around 2005 and declining since. Various motivation measures are discussed, including lengthening field work, improving contact materials, and interviewer training on the survey's importance. Response rates are shown to vary by region and subgroup, with lower response among those with less education or who are unemployed. Advancing information technology is noted as affecting response, though complete recovery is not expected. The need for quality indicators and better use of register data is emphasized to address non-response issues going forward.
This document summarizes findings from time use surveys conducted in Finland. It finds that:
1) Response rates were higher for telephone interviews (62%) than face-to-face interviews (58%), but total non-response was lower for telephone interviews.
2) There were no significant differences in diary quality or number of activities reported between face-to-face and telephone interviews.
3) A light paper diary pilot survey had a very low response rate of 17.4% and differences compared to full surveys, suggesting interviewer assistance is needed for quality and response.
This document summarizes an experimental study on the use of show cards in telephone surveys. The study tested show cards on alcohol consumption and chronic diseases in three different surveys. It found that show cards had a negative effect on response rates, especially for those with lower education levels. The presence of show cards did not significantly affect response distributions or improve data quality. Additionally, not many respondents reported using the show cards. In conclusion, show cards were found to increase non-response bias without improving data quality or response distributions.
The document discusses the establishment of a Nordic PIAAC Database containing data from the OECD's Survey of Adult Skills (PIAAC) and national registers from five Nordic countries and Estonia. The goals of the Nordic PIAAC Network, which is financing the database, are to create the database, publish a report on skills in the Nordic region, and make the microdata available for other research. The database will contain background questionnaire data, cognitive scores, and register data from national sources and will be housed securely at Statistics Denmark with researchers applying for access. Combining survey and register data provides more information and allows the database to be updated over time.
This document discusses the need to measure uncertainty in register statistics given their inductive nature. It proposes using Confidence Images (CIm), which restrict potential values of population parameters based on multiple sources of information, each with an associated measure of uncertainty. As a proof of concept, the document calculates CIms for the total number of cattle in Sweden using different combinations of register data, survey data, and prior knowledge. The CIm approach fulfills criteria for providing objective, interpretable and comparable measures of uncertainty for register statistics. Further methodological development is still needed, but CIms could become a useful statistical tool for quantifying uncertainty in register-based estimates.
This document summarizes a presentation given by Maija Metsä-Pauri of Statistics Finland about statistical communication and quality in the age of open data. It discusses how user needs are changing with digitalization and social media, requiring a shift from data dissemination to active communication. Statistics Finland has launched new interactive websites and a blog to better engage with readers. It also emphasizes the importance of statisticians telling statistical stories, collaborating with data journalists, and focusing on quality over quantity of information as open data becomes more prevalent.
Statistics Finland is developing web data collection for personal surveys through a set of projects running from 2012 to 2014. The goals are to develop infrastructure for web data collection and support mixed-mode data collection, especially for the Labour Force Survey. Four initial projects have been completed, including a mixed-mode pilot of the Consumer Survey. Current projects include developing Statistics Finland's data collection pages and a 2013 web pilot for the Labour Force Survey to study mode effects. By 2017, an internet questionnaire is planned to be available as an alternative for all personal surveys.
This document discusses non-response and motivation measures in Statistics Finland's EU Labour Force Survey from 1997-2013. It summarizes trends in non-response rates over time, with rates peaking around 2005 and declining since. Various motivation measures are discussed, including lengthening field work, improving contact materials, and interviewer training on the survey's importance. Response rates are shown to vary by region and subgroup, with lower response among those with less education or who are unemployed. Advancing information technology is noted as affecting response, though complete recovery is not expected. The need for quality indicators and better use of register data is emphasized to address non-response issues going forward.
This document summarizes findings from time use surveys conducted in Finland. It finds that:
1) Response rates were higher for telephone interviews (62%) than face-to-face interviews (58%), but total non-response was lower for telephone interviews.
2) There were no significant differences in diary quality or number of activities reported between face-to-face and telephone interviews.
3) A light paper diary pilot survey had a very low response rate of 17.4% and differences compared to full surveys, suggesting interviewer assistance is needed for quality and response.
This document summarizes an experimental study on the use of show cards in telephone surveys. The study tested show cards on alcohol consumption and chronic diseases in three different surveys. It found that show cards had a negative effect on response rates, especially for those with lower education levels. The presence of show cards did not significantly affect response distributions or improve data quality. Additionally, not many respondents reported using the show cards. In conclusion, show cards were found to increase non-response bias without improving data quality or response distributions.
2. Vad är en webbpanel?
• Ett register över personer som förklarat sig
villiga att delta i webbundersökningar
• E-postadressen är känd samt ett antal
andra bakgrundsvariabler
• Ofta representerar panelen allmänheten i
någon mening
• Undersökningar görs på urval ur panelen
3. Varför panel?
• För webbundersökningar behövs ett
register som kan fungera som ”urvalsram”
• E-postadresser till allmänheten finns inte i
befintliga register
• (Notera: longitudinella undersökningar
eller förändringsskattningar är ej primära
syften)
4. Hur rekryteras webbpaneler?
• Baserat på sannolikhetsurval
– Urval med enda syfte att rekrytera till panelen
– Urval med andra syften där fråga om
deltagande i panelen finns med
• Olika former av självrekryterade paneler
– Rekrytering via annons, hemsida eller ”pop-
up”
5. Profilundersökning
• Rekryterade personer genomgår en
profilundersökning där bakgrundsvariabler
registreras
• Urval dras ur panelen för specifika
undersökningar. Respondenten kontaktas med
epost och får en länk till en webbenkät
6. Bakgrund till den svenska
Surveyföreningens engagemang
• Allt vanligare med undersökningar utförda
på webbpaneler
• Stickprovsteorin oftast inte tillämplig
• Svårt att utvärdera kvaliteten (represen-
tativitet, generaliserbarhet) i resultat
• En webbpanelkommitté tillsattes
7. Webbpanelkommittén - Roll att fylla
• Föreslå beskrivningsmått för bedömning av
kvaliteten i resultat från webbpanelundersökningar
• Komplement till ISOs, AAPORs och ESOMARs
riktlinjer
• Målgrupp
– Panelleverantörer, undersökningsköpare, andra
användare av undersökningsresultat
• Öka transparensen i tillvägagångssätt och
metoder
– Viktigt för branschens förtroende och för kundernas
förståelse
10. Mått relaterade till bortfall
• Mått med kopplingar till traditionella undersökningar
• Rekrytering – Rekryteringsandel RA
– Sannolikhetsbaserade urvalsmetoder för rekrytering
• Profilundersökning – Profilandel PA
• Specifik undersökning – Deltagarandel DA
• Kumulativ deltagarandel KA = RA × PA × DA
• Exempel:
– Antag RA=0,3, PA=0,8, DA=0,5
– Kumulativ deltagarandel KA=0,3×0,8 ×0,5=0,12
11. Övriga mått: Två typer
• Mått relaterade till panelen
• Mått relaterade till en specifik undersökning
• Samband mellan mått och kvalitet ofta
komplicerat
12. Mått relaterade till panelen
Uppgiftslämnarbörda
• Belysa hur många undersökningar
panelmedlemmarna konfronteras med (per år)
• Två mått:
– UB1: antal utsända inbjudningar totalt i förhållande till
panelstorlek
– UB2: antal komplett besvarade enkäter i förhållande
till panelstorlek
13. Mått relaterade till panelen
Dominans från vissa
panelmedlemmar
𝐷20 = Andel ifyllda enkäter som de 20
procent mest aktiva medlemmarna står för
16. Conditioning (forts)
Mått:
• Dela in de svarande i grupper efter hur många
undersökningar de tidigare svarat på och jämför
gruppernas svarsmönster
• Beräkna
– AS1 = andel obesvarade frågor (slarv)
– AS2 = mediantiden att besvara enkäten (slarv)
– AS3 = andel med t ex extrempunktsvärde på vald
fråga (professionalisering)
17. Nåbarhetsandel (Absorbtion rate)
• Nåbarhetsandel NA – andelen av alla
inbjudningar som kommer fram (ej returneras)
• Måttet NA mäter den grundläggande kvaliteten
på e-postadresserna
18. Avbrottsandel (Break-off rate)
• Avbrottsandel AA - Andelen enkäter som öppnats
men klassats som ej svar i förhållande till samtliga
öppnade enkäter (besvarade och ej besvarade)
• Är en möjlig indikator på problem med designen av
ett frågeformulär (för långt, för tråkigt)
• Kan säga något om förekomsten av tekniska
problem
• Hög avbrottsandel tyder på att man bör kontrollera
undersökningen
19. Komplement: Verbala beskrivningar
• T ex
– Urval – hur gick urvalet till?
– Viktning – redovisa på ett transparant sätt hur
eventuella vägningar har genomförts
20. Planerad fortsättning
• Publicering av rapporten under hösten 2013
• Inbjudan till undersökningsföretag att presentera
värden på olika mått på Surveyföreningens
hemsida (flera företag är intresserade)
• Uppbyggnad av erfarenheter när det gäller
relationen mellan måttens värden och
resultatens kvalitet