cite: Thoresen2021
author: Thoresen, S. H., Cocks, E., & Parsons, R.
year: 2021
title: Three year longitudinal study of graduate employment outcomes for Australian apprentices and trainees with and without disabilities
publisher: International journal of disability development and education
uri: https://doi.org/10.1080/1034912X.2019.1699648
pubtype: article
discipline: education

country: Australia
period: 2011-204
maxlength: 36
targeting: explicit
group: disabled
data: experimental survey

design: quasi-experimental
method: quantitative survey (n=489); qualitative semi-structured face-to-face interviews (n=30); annual postal survey, baseline and 2 follow-ups; generalised estimating equation GEE
sample: 489
unit: individual
representativeness: local
causal: 0 # 0 correlation / 1 causal

theory:
limitations: non-representative sample, over-representation of learning disability; limited generalisability through sample LFP bias and attrition bias; small control sample size
observation:
  - intervention: training
    institutional: 0
    structural: 1
    agency: 1
    inequality: disability; income
    type: 1 # 0 vertical / 1 horizontal
    indicator: 0 # 0 absolute / 1 relative
    measures: hours worked
    findings: slightly lower for disabled group initially, increase to no significant difference with non-disabled group at last survey
    channels: significant but small overall increase (3.1 hours to 1 hour difference); fluctuations for non-disability group
    direction: 1 # -1 neg / 0 none / 1 pos
    significance: 2 # 0 nsg / 1 msg / 2 sg
  - intervention: training
    institutional: 0
    structural: 1
    agency: 1
    inequality: disability; income
    type: 1 # 0 vertical / 1 horizontal
    indicator: 0 # 0 absolute / 1 relative
    measures: hourly/weekly income
    findings: wages of disability group substantially lower than non-disability; increases to be non-significant over time; lower for female and disability-pension recipient groups
    channels: strong initial diff means disability group potentially more often initially employed at junior rates or skewed through attrition bias
    direction: 1 # -1 neg / 0 none / 1 pos
    significance: 2 # 0 nsg / 1 msg / 2 sg

notes: Disaggregated results for female participants overall more unequal
annotation: |
  An experimental survey combined with qualitative interviews for participants of a vocational training programme in Australia, looking at the effects on participants' hours worked and incomes.
  It finds, foremost, that initially both the hours worked and the income of people with disabilities are lower on the Australian labour market in general and this reflects in the results for the disability group of participants, which have significantly lower weekly incomes and hours worked than the control group.
  Over time, hours worked increase for the disability group to no longer be significantly different but still lower than for the control group (from 3.1 hours to 1 hour difference per week),
  however there are large fluctuations in the control group.
  Similarly, the wages of the disability group are initially substantially lower than of the control group, 
  which increases to be non-significant though still lower over time, more so for the earnings of female participants and participants which received a disability pension.
  Relevant limitations of the study include the use of a non-representative sample for the national representativeness, 
  and the overall generalisability being low due to an increased labour force participation bias and attrition bias of the surveys, 
  as well as only having access to a small control sample size.
  Thus, findings should be understood as guiding policy directions, while generalisations should be done with care as some of the larger changes may be due to those limitations,
  such as the increased survey response of those with positive wage outcomes.