Can we make environmental citizens A randomised control trial of

Document Sample
Can we make environmental citizens A randomised control trial of Powered By Docstoc
					Can we make environmental citizens?
A randomised control trial of the effects of a school-based
 intervention on the reported behaviour of young people

    Goodwin, M., Greasley, S., John, P., and Richardson, L.

         Institute for Political & Economic Governance

►   Personal responsibility and behaviour change

►   Pro-environmental attitudes and behaviour traced to increasing individual
    contribution (e.g. Defra, 2008)

►   Children and the environment - an oft-neglected aspect: “…little is known
    about the contents of early childhood environmental attitudes and
    behaviors” (Evans et al. 2007, p.636)

►   School-based environmental education programmes: one method used to
    encourage the development of pro-environmental attitudes and behaviour

►   Literature on civic education which shows modest effect
Current state of knowledge

►   Evidence to support proposition that school-based environmental education
    programmes exert positive effect (for reviews see Rickinson, 2001; Zelezny 1999).

►   Example #1: Bogner’s (1998) experiment investigated the effects of a 5-day
    residential programme on 351 secondary school students. Compared with a control
    group, participants reported significant gains in knowledge, attitudes and intended

►   Example #2: Evaluation of a 5-day residential programme on 697 students reported
    a significant difference in attitudes toward wildlife which were also greater than a
    control group. Also suggested long-term attitudinal change (Dettman-Easler &
    Pease 1999)

►   Evaluation of 10 week high-school course based on pre-/post-testing found
    environmental knowledge and attitudes increased but study not based on
    comparison with a control group (Bradley et al. 1999)

►   No need to preach here about RCTs!
Study Design

►   Setting for the study was Vale Royal, Cheshire (Frodsham, Northwich, Winsford)
►   Population of 126,000, 98.8% white
►   In 2004 Vale Royal was ranked 154th most deprived out of 354 English districts
    (fourth highest ranked Cheshire district)
►   Deprivation low
►   Educational performance (GCSE) better than national average
►   Intervention focused on children in Year 3 (aged 7-9 years)

►   27 primary schools opted into the study (1 later withdrew)

►   Randomised into two treatment groups and one control group
      8 schools allocated to ‘long’ treatment
      8 schools allocated to ‘short’ treatment
      11 schools allocated to control group (treatment deferred for one year)

►   Randomisation implemented in Excel using method suggested by Shadish
    et al. (2002, p.313)
►   Long treatment group
      Two in-class environmental education talks and knowledge-based DVD
      Homework environmental workbook and activities
►   Short treatment group
      Two in-class environmental education talks and knowledge-based DVD
►   Control group

► Schools divided into two phases to aid delivery of intervention (Jan-April 2008,
  April-July 2008)
► Duration of programme from baseline-post-survey collection was approx. four
► Deliverer of intervention and surveys was different (former trained instructor,
  latter teacher)
► Outcome measures obtained via pre-/post-surveys:
    1. Survey completed in-class and supervised by teacher
    2. Homework survey completed in collaboration with parents/guardian
Basic features on the study: overview

►   715 pupils
►   Average number per school – 26.5 pupils
►   Classes ranged from 19-38 children
►   Response rate – 63% (completing baseline and post-test in-class survey)
►   Response rate to household survey much lower (26%)
►   We could find no difference in the characteristics of non-responders between
    groups for either survey
►   Quick report on the homework and household survey
►   Methods use regression with clustered standard errors to control for clustering
►  Target variable: reported environmental behaviour 5 point additive scale (0-4):
-Recycle rather than throw things away
-Take part in litter picks
-Walk or cycle to school rather than use car
-Waste less food at home

Measures as children’s response and also adult carers
Regression on classroom score
Table 1: Baseline, post-test means and average change (classroom)

               Baseline       Post-test   Av. Individual   Obs
               mean           mean        change
Control        2.22           2.27        +0.05            176
Long           1.95           2.15        +0.2             127
Short          2.30           2.35        +0.05            123
All            2.16           2.26        +0.09            426

Table 2: Ordered Probit Model of post-test scores (standard errors
  clustered by school)
                       Co-efficient           P-value
Short intervention        0.086                  0.706

Long intervention         -0.129                 0.559
Regression on households
Table 3: Baseline, post-test means, average change (household survey)

Baseline mean   Post-test mean                    Av. Individual    Obs

2.51            2.65                              0.135             74
Long            2.45             2.55             0.091             44
Short           2.49             2.51             0.018             57

Table 4: Ordered Probit regression of post-test scores (standard errors clustered by
                            Co-efficient                  P-value
Short intervention          -0.196                        0.442

Long intervention           -0.161                        0.587

► Lack of change in reported behaviour
► Some change in reported behaviour but not significantly different between
  experimental groups
► Caution from localised nature of the study and small N

Shared By: