Next Steps Webinar. 27 June 2017

Transcription

1 Next Steps Webinar 27 June 2017

2 Webinar Outline 12:30-12:50 Introduction to Next Steps with Q&A Lisa Calderwood 12:50-13:10 Age 25 Survey content with Q&A Darina Peycheva 13:10-13:30 Data, documentation and accessing the data with Q&A Sarab Rihal 13:30-13:50 Sample design and weights with Q&A Morag Henderson 13:50-14:00 Q&A General session All 14:00 End

3 Introduction to Next Steps Lisa Calderwood

4 Important study following lives of 16,000 people born in England in Began in 2004 when study members were aged Sampled through schools Annual surveys until 2010 when study members were age 19/20 (sweeps 1-7) Age 25 survey (Sweep 8) took place in 2015/2016 Aim is to continue following study members into future

5 Why is Next Steps important?

6 Timeline of the studies Next Steps follows the lives of around 16,000 people living in England who were born in 1989/

7 Why is Next Steps important? The missing cohort young people born in the early 90s Unique challenges of their generation Only major national longitudinal study focussing on pathways through the teenage years and transitions into adulthood Has already been extensively used by analysts and policy makers across government

8 How has Next Steps made a difference? raising the participation age in formal education and training improving the quality and choice of vocational qualifications setting up the Educational Maintenance Allowance strategy around bullying in schools

9 Recent Research Findings Sexual minorities more at risk of being bullied than heterosexual peers New research using Next Steps data suggests that young people who identify as sexual minorities are at a greater risk of ever experiencing bullying and experiencing bullying frequently during compulsory schooling and in early adulthood.

10 What does the Next Steps data include?

11 Next Steps A study of people born in 1989/ Age 14 Age 15 Age 16 Age 17 Age 18 Age 19 Age 20 Age 25 main respondent Study member Study member Study member Study member Study member Study member Study member Study member Parent response Parents Parents Parents Parents Linked data National Pupil Database National Pupil Database Linkage consents sought sample 15,770 13,539 12,439 11,801 10,430 9,799 8,682 7,707 response rate 74% 86% 92% 92% 89% 87% 90% 51%

12 Topics covered by life stage School years Early Adulthood Family Parental employment Financial circumstances Housing Health Behaviour Educational attainment Educational aspirations Subject choice Views and expectations Parental reports on relationship Bullying Educational transitions Family (partners, children) Employment Income Housing Health Courses and qualifications Sexual identity Political engagement Locus of control

13 Thank you Any questions?

14 The next section Age 25 Survey content with Q&A will begin at 12.50pm

15 Survey Content with Q&A Darina Peycheva

16 Structure of the presentation: Next Steps Age 25 Survey overview Survey context Questionnaire content, and Instrument(s) specific features Initial findings Q&A

17 The Age 25 survey: Survey overview Fieldwork took place between August 2015 and September 2016; Interviewing involved sequential mixed-mode design - including web, telephone and face-to-face; All cohort members who have ever taken part in any of the previous sweeps of the study were contacted; 15,531 cohort members were issued for fieldwork, interviews were completed with 7,707 (representing a 51% response rate); 62% of interviews were achieved via web, 9% via telephone and 29% face-to-face; Across all modes completion of the survey took on average 47 minutes; See Next Steps Age 25 Survey Technical Report and Appendices for full account of the study development and fieldwork procedures.

18 The Age 25 Survey: Context Transition from DfE to CLS from strong focus on education to transitions out of education and into early adult life, and therefore more multi-disciplinary research resource and similarity to the other CLS cohorts (MCS, BCS70, NCDS). Transitions to adulthood time of change in young people s lives: big decisions about careers, families, housing, and impact on health

19 The Age 25 Survey: Research opportunities How to reduce inequalities? How this generation has experienced a severe global recession How to confront problems of mental health How to combat the growth of obesity? How to ensure appropriate skills acquisition? How this generation is getting on in the labour market?

20 The Age 25 Survey: Questionnaire content Data was collected about cohort members housing and family life, employment and economic circumstances, education and job training, physical and emotional health, and identity and participation. A wide range of administrative data linkage consents were collected covering health, education, economics and criminal behaviour. Instrument specific features, and Standardized measures.

21 The Age 25 Survey content covers the following topics: Household relationships Current relationship Previous cohabiting relationships (dating back to September 2006) Children Childcare Non-resident children Non-resident parents Other household members Instrument specific features: an Event History Calendar (EHC)

22 The Age 25 Survey content covers the following topics: Housing Current Previous housing (summary data is collected about the different addresses the study members have lived in since they were 16, if other than the parents home) Instrument specific features: an Event History Calendar (EHC)

23 The Age 25 Survey content covers the following topics: Employment Current activity Current employment Second job Prospective employment (for unemployed) Activity history (dating back to September 2006) Employment details for first job after September 2006 (aged 16) Employment support, work attitudes, and Partner employment. Instrument specific features: a keyword look-up for coding occupation; an Event History Calendar (EHC)

24 The Age 25 Survey content covers the following topics: Finance Current pay/salary main job Pay from second job Income from other jobs Partner s income Benefits Income from other sources Household income Pensions Debt

25 The Age 25 Survey content covers the following topics: Education and job training Job training Education (since previous interview/september 2006) Current education Fees Partner s education Instrument specific features: an Event History Calendar (EHC)

26 The Age 25 Survey content covers the following topics: Health and wellbeing General health Height and weight Exercise Sleep Diet Accidents and injuries Scales: Health module, Long-lasting Health Conditions and Illnesses: Impairments and Disability (ONS, 2015)

27 The Age 25 Survey content covers the following topics: Identity and participation Ethnicity and religion Trust, risk, patience Meritocratic beliefs Adult identity Leisure Politics Social networks and social media participation. Scales: Identity module, Adult Identity Resolution Scale (Côté, 1997)

28 The Age 25 Survey: content covers the following topics: Self-completion module Gender identity Locus of control Overall life satisfaction Mental health, and Self-harm Crime and harassment Drinking and smoking behaviour Drugs, bullying, sexual behaviour, and Pregnancy history. Ensuring comparability of the data in a mixed mode interview adaptations to sensitive questions for completion in the telephone mode Scales: Sensitive module, Locus of control, GHQ-12 and Drinking behavior

29 The Age 25 Survey content covers the following topics: Asking for administrative data linkage consents Domain Data set / information Government Department / Organisation Education National Pupil Database (NPD) Individualised Learner Records (ILR) University participation, attainment and destinations Student Loans Higher education applications and offers Health Hospital Episodes Statistics (HES) NHS Digital Department for Education (DfE) Department for Business, Innovation and Skills (BIS) Higher Education Statistics Agency (HESA) Student Loans Company (SLC) University and College Admissions Service (UCAS) Economic Benefit claims and employment programme Tax credits, NI contributions, income Department for Work and Pensions (DWP) Her Majesty's Revenue and Customs (HMRC) Crime Police National Computer (PNC) Ministry of Justice (MOJ)

30 Data linkage consents obtained by mode Domain Web (%) Telephone (%) F2F (%) Overall (%) Education (NPD, ILR, HESA) Education (UCAS) Education (SLC) Health (NHS) Economic (DWP) Economic (HMRC) Crime (MOJ)

31 The Age 25 Survey: Initial findings Next Steps Age 25 Briefing papers Economic activity and health Mental health Physical activity and quality of life Politics, perceptions and identity How to access the Next Steps Age 25 briefing papers

32 Thank you! Any questions?

33 The next section Data, documentation and accessing the data with Q&A will begin at 1:10pm

34 Data, documentation and accessing the data with Q&A Sarab Rihal

35 Files: Datasets Name Contents Structure Identifier NS8_2015_Main_Interview Modules 1 to 7 Flat NSID NS8_2015_Self_Completion Module 8 Flat NSID NS8_2015_Partnerships Relationship histories Hierarchical NSID, W8RELID NS8_2015_Children Details of children of CM Hierarchical NSID, W8CHID NS8_2015_Household_Members NS8_2015_Activity_History NS8_2015_Benefits NS8_2015_Income_Unfolding_brackets NS8_2015_Benefits_Unfolding_brackets Details of members living in same household as CM Activities and Employment histories Details of individual benefits received Unfolding brackets questions for payments and income Unfolding brackets questions for benefits Hierarchical Hierarchical Hierarchical Flat Hierarchical NSID, W8HHMID NSID, W8HISTID NSID, W8BENID NSID NSID, W8BENID NS8_2015_Derived_variables Derived variables Flat NSID

36 Files: Identifier New case identifier A new case identifier has been introduced called NSID NSID will replace the old serial number surveyid Merge datasets together using NSID Change will apply to all datasets in all sweeps (1 to 8) Redeposit of Sweeps 1 to 7 at same time as deposit of Sweep 8 All datasets that carry the old serial number (surveyid) should normally be deleted within 3 months of receipt of the revised data

37 Files: Identifier New identifier format Format of NSID Identifier Format description Example NSID 8 characters: NS followed by 5 digits followed by a single character NS21140C SURVEYID 5 digits 25533

38 Files: Structure Flat files 1 record per case Cohort member (CM) response uniquely identified by the key identifier NSID Dataset: NS8_2015_Main_Interview Hierarchical files 1 or more records per case CM response uniquely identified by NSID and a secondary index e.g. W8RELID Dataset: NS8_2015_Partnerships

39 Files: Other useful Unfolding brackets If respondents were unable or unwilling to give an exact amount to a relevant income or payment question, unfolding brackets were used to gain an estimate of that amount. Derived variables A set of useful key variables which have been constructed from the questionnaire data. Documented separately in user guide

40 Data: Derived variables (NS8_2015_Derived_Variables) Example 1 Scales: Alcohol Use Disorders Identification Test Consumption (AUDIT C) Variable Name W8AUDIT1 W8AUDIT2 W8AUDIT6 W8DAUDIT Variable label How often has a drink containing alcohol How many drinks containing alcohol has on a typical day of drinking How often had six or more drinks on one occasion in the past year DV: Alcohol Use Disorders Identification Test Consumption (AUDIT-C) scale Alcohol consumption, problems and dependency W8AUDIT1, W8AUDIT2 and W8AUDIT6 scored from 0 to 4 Maximum score of 12 Score of 5+ considered AUDIT-C positive

41 Data: Derived variables Example 2 Income derived variables Variable Name W8DINCW W8DINCB Variable label DV: Continuous weekly income DV: Banded weekly income Use of income banded questions Example 3 Ethnicity 6, 8, 11 and 15 category ethnic groups (2011 Census Office for National Statistics) All cohort members represented

42 Data: Conventions Variable names Prefixed with W8 denoting wave/sweep of the cohort study. CAPI name contained within dataset variable name Length of variable name has not been restricted to a fixed number of characters Multi-coded variables have been given the suffix: 0A, 0B, 0C,..., AA, AB to denote the 1 st, 2 nd, 3 rd,..., 26 th, 27 th response respectively. Examples: W8LOLR0A Effect of longstanding illness: Vision W8LOLR0Z Effect of longstanding illness: Respiratory problems

43 Data: Conventions Variable labels Indication of the question content Common prefixed used for multi-coded variables Derived variables prefixed with DV Value labels Missing values -9 Refused -8 Don t know (or Insufficient information ) -1 Not applicable Order of variables

44 Data: Documentation User Guide Survey background, design and development, questionnaire content, data linkage, response and weights, datasets and data conventions Questionnaire CAI instructions, CAPI names, values, value labels, routing Technical Report Sample design, questionnaire design and implementation, development work, fieldwork, response, coding editing and data preparation Derived Variables User Guide Variable names, variable labels, values, value labels, description of derivation

45 Accessing the data and documentation UK Data Service registration, data and documentation Types of access End User Licence (5545) - Next Steps Age 25 Survey datasets is available under the same series number as previous Next Steps sweeps Secure access - (Administrative, Sensitive) Choice of formats for download (SPSS, STATA etc)

46 Finding Variables There are many Next Steps files and each file contains a huge number of variables. To help locate specific variables you can: Download the questionnaire for each sweep from the UKDS Use UKDS Nesstar catalogue We anticipate that all of the sweeps will be available on CLOSER Discovery (available in 2018) Search for variables from sweeps 1-7 from ilsype Alternatively, download the data (EUL) for each sweep from the UKDS In Stata use the command lookfor which allows you to search variables and labels based on key words

47 Thank you Any questions?

48 The next section Sample design and weights with Q&A will begin at 1.30pm

49 Sample design and weights Morag Henderson

50 Outline The Next Steps sample and design Attrition in Next Steps What do we know about attrition in Next Steps? What can we do about it?

51 The Next Steps sample Respondents were selected to be representative of young people in England in Sample boost at Sweep 4 for some sub-groups Sample were selected from a target population of those in Year 9 (or equivalent) in February 2004 Those who were born between 01/09/89 and 31/8/90 The sample did not include those solely educated at home, boarders and those solely in England for purposes of education

52 The Next Steps sample A two stage probability proportional to size (PPS) sampling procedure with disproportionate stratification Schools were primary sampling units (PSUs) The second stage sampled the pupils within schools Pupils from major minority ethnic groups (Indian; Pakistani; Bangladeshi; Black African; Black Caribbean; and Mixed) were over-sampled at pupil level in order to achieve target issued sample numbers

53 The Next Steps sample Taken together, the school selection probabilities and the pupil selection probabilities ensured that within a deprivation stratum, all pupils within an ethnic group had an equal chance of selection. 15,770 households who took part in Wave 1 and 647 schools

54 Attrition Definition Attrition is the discontinued participation of some individuals in a longitudinal survey for reasons that are unknown and/or beyond the control of the researcher

55 Problems of non-response and attrition Distinction between unit (respondents ) non response and item nonresponse Types of non-response (have separate reasons) for example unitnon-response: Non-contact Refusal Inability Out of scope/ineligible Non-response on increase in all surveys Non-response may not be permanent (in panel survey) Non-response/attrition can have some important implications

56 Productive sample over time Per Cent Sweep 1 Sweep 2 Sweep 3 Sweep 4 Sweep 5 Sweep 6 Sweep 7 Sweep 8

57 Monotone vs. non-monotone response Monotone response: respondents dropped out without coming back. Non-monotone: interrupted response pattern over time. Monotone vs. non-monotone response in Next Steps Number of times productive up to Next Steps wave 8 Patterns Freq. Percent Monotone 8, Non-monotone 2, All waves 5, Total 16, Times productive Freq. Percent One 1, Two 1, Three 1, Four 1, Five 1, Six 1, Seven 3, Eight 5, Total 16,

58 Effects of non-response/attrition Missing data smaller samples fewer transitions incomplete histories low statistical power Biases in results (all surveys) disproportionate to some groups (mobile, disadvantaged, young, men, long working hours) Problem if linked to survey variables or administrative data

59 Non-response bias - implications Ignore the problem: equivalent to assuming that there are no sample bias Use adjustment techniques such as weighting and imputations

60 Design and attrition weights Design weights: adjust the sample composition to take account of over-sampling Attrition weights: adjust the sample composition to take account of the loss of particular type of respondents. Adjustment means giving more importance (weight) to a particular group. Overall weights = Design Wgt x Attrition Wgt

61 Attrition weights construction Logistic response models Dependent variable: binary response outcome (0/1) in wave M. Independent variables: characteristics of respondents in previous waves. Cohort members circumstances in Sweep 1 Cohort members circumstances in Sweep 7 Family circumstances and cohort members (CM) characteristics Highest qualification held by main parent Employment status of main parent Social status NS-SEC of the family Marital status of main parent Cohort member s gender Cohort member s ethnic group Whether cohort member ever identified as having special educational needs Government office region Housing tenure Current activity including education and employment Whether cohort member ever tried cannabis Month of interview Interview mode

62 Next Steps weights Variable name Variable label W8FINWT Weight: W8 Final Weight Overall weights = Design weights x Attrition weights Weights are prefixed by wave number e.g. W1, W2, W3

63 Choosing the correct weight The choice of the weight depends on the analysis you are doing. The weight should correspond to the most recent wave of data you are using For instance, if you are using data from waves 1,2, 5 and 8, the wave 8 weights should be used. svyset [pweight=w8finwt], psu(samppsu) strata (SampStratum)

64 Thank you! Any questions?

65 Thanks to our funders and host institution

66 The next section Q&A General Session will begin at 1.50pm