1. The Development of a Successful Household Panel Survey: The HILDA Experience Mark Wooden Project Director, HILDA Survey

2. www.melbourneinstitute.com About HILDA: Introduction  Funded and owned by Australian Government  Multi-purpose survey –Modelled on other household panels – BHPS, SOEP  Survey manager = Melbourne Institute of Applied Economic & Social Research (University of Melbourne)  Fieldwork subcontractor = Roy Morgan Research  Unit record data available (under license)  Want to know more?  Articles in The Economic Record, June 2007 and Australian Economic Review, September 2010  www.melbourneinstitute.com/hilda/

3. www.melbourneinstitute.com About HILDA: Key Design Features  Commenced (in 2001) with national probability sample of households –Area-based clustered / stratified sample design  Annual survey waves  Follow all original hh members and offspring indefinitely  Sample augmented with hh joiners  Interview all “adults” –Face-to-face where possible –CAPI / CATI technology  Refreshment (top-up) sample added in wave 11  Cash incentives paid

4. www.melbourneinstitute.com Survey Instruments  Household Form –Key identifiers / Changing HH membership / HH relationships / Reasons for non-response  Household Questionnaire –Collects hh level data from relevant HH member  Continuing Person Questionnaire –All persons 15+ who have previously been interviewed  New Person Questionnaire –All persons 15+ who have never previously been interviewed  Self-completion Questionnaire –All interview respondents; 16 pp, expanded to 20 from W5

5. www.melbourneinstitute.com What’s In It? HQ / CPQ  Core: –Child care –Housing –Education –Employment status –Job characteristics –Job search –Calendar –Income –Family formation –Partnering & relationships –Living in Australia Disability, Life satisfaction, Spatial mobility, Caring –Tracking –Interview situation  Special “modules”: –W1 (+NPQ) = Personal history –W2 =Wealth –W3 =Retirement –W4 =Youth issues; Private health insurance –W5 = Family formation –W6 = Wealth –W7 = Retirement; Lifestyle –W8 = Family formation; Non- cores. relationships –W9 = Health –W10 = Wealth –W11 = Family formation; Retirement –W12 = Skills & abilities; Non- cores. relationships

6. www.melbourneinstitute.com What’s In It? SCQ  Health and well-being (SF36, Kessler 10, serious health conditions)  Health behaviours (smoking, drinking, exercise, height / weight, diet)  Social capital / relationships (satisfaction with family, social support, community participation, religion)  Neighbourhood characteristics  Life events  Time use  Finances (stressful financial events, savings habits, risk preference, h’hold expend)  Job attributes  Parenting (parenting stress / work family gains and strains)  Attitudes to work / gender roles / marriage  Personality

7. www.melbourneinstitute.com Indicators of Success  We are still going! –Funded renewed until wave 16 –And total funding has increased  Response / attrition rates are good to excellent  Data usage is high  Strong evidence of validity

8. www.melbourneinstitute.com Response in 2001 was good  HH response –In-scope sample = 11,693 –7682 cooperating households = 66% RR  Individual response –W1 individual sample = 15,127 persons –13,969 respondents = 92% RR  Sample reasonably representative, but … –Sydney residents under-represented –People from a NESB under-represented –Males less likely to complete a PQ (but no less likely to be a CSM)

9. www.melbourneinstitute.com Response in 2011 was better

10. www.melbourneinstitute.com Response in 2011 was better

11. www.melbourneinstitute.com Retention is High (Annual Re-interview Rates: HILDA, BHPS & GSOEP)

12. www.melbourneinstitute.com Retention is High (Annual Re-interview Rates: HILDA, BHPS & GSOEP)

13. www.melbourneinstitute.com Fieldwork Outcomes: W1 Adults

14. www.melbourneinstitute.com Tracking Movers 22-23% of all hh’s change address b/w each survey wave  Pre-field office activity –Notifications (1800#, change of address card, email) –Matching to Australia Post –Returns to sender –Move indicator variable  Other household members  Contact information collected at previous ivw  Neighbours  Other community resources  Online White Pages

15. www.melbourneinstitute.com Minimising Refusals MARKETING / RESPONDENT ENGAGEMENT  PAL and brochure, newsletter / Stat report  1800 number PERSISTENCE  2-3 stage fieldwork  NRs re-issued in later waves GOOD PEOPLE  Selection and continuity of interviewers  Training / interviewer engagement RESPONDENT INCENTIVES

16. www.melbourneinstitute.com Data User Numbers ReleaseTotal data ordersNew usersCumulative total 1204 202 2265169373 3279157530 4329176706 5387196902 64011761078 74551991277 84311251402 95001411543 10 (@19 July) 4261321675

17. www.melbourneinstitute.com Publication Count Year Journal articles Books / book chapters Other publications Working papers 20025003 20036228 2004244815 2005243821 20062511923 20073501135 20083802335 20094772735 20105262030 20116403642 2012 / forthcoming5201714 TOTAL37223171261

18. www.melbourneinstitute.com Promoting Data Use  Well-documented, user-friendly data sets  User Manual  Other on-line tools (e.g., PanelWhiz)  Discussion Papers / Technical Papers  User training and panel data analysis courses  Biennial research conference  Membership of CNEF  Presentations to different stakeholders  Annual Statistical Report  Study-specific web site  User email list

19. www.melbourneinstitute.com Research Uses: Key Features  Topic coverage extremely broad

20. www.melbourneinstitute.com Research Topics: Just a Few Examples!  Income and wealth –Poverty dynamics –Distribution of household wealth –Retirement savings  Labour supply / Unemployment –LFP and health –Family policy and couples LS –Impact of child care costs –Forgone earnings of mothers  Employment –Working hours mismatch –Casual employment transitions –Part-time employment and wages –Job insecurity –Responses to long hours –Gender inequity  Marriage and family –Patterns of cohabitation –Children’s living arrangements –Post-separation contact with children –Childlessness –Predictors of marital separation  Subjective well-being –Adaptation to life events –Predictors / correlates of life satisfaction  Mental health and: –welfare reliance –retirement –housing affordability –joblessness –job quality

21. www.melbourneinstitute.com Research Uses: Key Features  Topic coverage extremely broad  Three key types of studies i.Innovative content / questions ii.Unobserved heterogeneity iii.Dynamics of change (and persistence)  Still many cross-sectional analyses

22. www.melbourneinstitute.com Policy Impacts: Examples  Key input into Government’s Pension Review  Annual Wage Reviews  RBA –Household debt and risk –Effect of the superannuation guarantee on household saving  Productivity Commission – Paid Parental Leave report

23. www.melbourneinstitute.com Keys to Success: Response  Expectations of fieldwork agency  Motivated interviewer workforce  Long fieldwork period  Persistence  Cash incentives

24. www.melbourneinstitute.com Keys to Success: Other Ingredients  Champions (and lots of them)  Money (and lots of it)  Imitation  Good people  Many users  Luck

25. www.melbourneinstitute.com Other Issues To Think About (I)  Sample –Population, dwellings, households –Clustered / stratified –Dealing with future immigration  What mode? –Interviewer administered vs self-administered –Single mode vs mixed mode or multi-mode  Respondent burden  How to reach non-English speakers?

26. www.melbourneinstitute.com Other Issues To Think About (II)  Making use of technology –Dependent data / On-line options  How much value adding? –Data cleaning / Weights / Imputation / Derived variables  Confidentiality vs data access  Linkages to admin. data  Scientific stewardship / Stakeholder involvement

27. The Development of a Successful Household Panel Survey: The HILDA Experience Mark Wooden Project Director, HILDA Survey