Evaluating the impact of urban green space interventions on physical activity and other wellbeing behaviours: developing and implementing improved natural experimental methods

Student thesis: Phd


Evidence suggests that interventions to improve or create new urban green spaces have positive effects on physical activity and other behaviours important for wellbeing. However, the evidence base is scarce, especially in Europe, and often of poor quality. Environmental interventions are rarely amenable to randomisation, so a natural experimental study is the optimal approach to evaluate the causal effects of urban green space interventions. My recent review of twelve natural experimental studies of environmental interventions on physical activity found major methodological weaknesses causing high risk of bias and identified eight recommendations for future research. This PhD aimed to: (1) address these recommendations by developing new methods to improve the internal validity of natural experimental studies; (2) develop and validate a new systematic observation tool for unobtrusively assessing physical activity and two other wellbeing behaviours in urban spaces; (3) implement these new methods in two natural experimental studies of urban green space interventions. Three studies combined to develop and validate MOHAWk (Method for Observing pHysical Activity and Wellbeing): a systematic observation tool for assessing physical activity and two other evidence-based wellbeing behaviours (social interactions and taking notice of the environment) in urban spaces. From 156 hours of observations using six observers across five urban spaces, evidence is provided that MOHAWk is reliable and valid. New natural experimental methods were developed and implemented in two studies of urban green space interventions. Key methodological improvements included a new process for systematically identifying matched comparison sites; appropriate adjustment for confounders to minimise the risk of confounding; publication of study protocols with a priori analyses specified (reporting any deviations); sample size calculations; process measures; and clear reporting in line with standardised checklists. The first natural experimental study assessed the impact of low-cost changes (e.g. tree planting) to four urban amenity green spaces on older adults’ and adults’ wellbeing behaviours in Greater Manchester, UK. There was no evidence that the interventions increased observable wellbeing behaviours or green space use. A nested qualitative study suggested the interventions were not substantial enough to be noticed compared to other recent neighbourhood changes and therefore unlikely to influence behaviour change. The second natural experimental study assessed the impact of new walking infrastructure and green space improvements along an urban canal on canal usage and wellbeing behaviours among adults in Greater Manchester, UK. There was evidence that the intervention significantly increased the total number of people using the canal path compared to the comparison sites at all follow-ups. There was some evidence that the intervention brought about increases in walking and vigorous physical activity, social interactions, and people taking notice of the environment. A process evaluation suggested that there was some displacement of activity, but the intervention also encouraged existing users to use the canal more often. These natural experimental studies provide exemplars of how to use methods with substantially lower risk of bias than previous research. MOHAWk is an unobtrusive and inexpensive outcome measure that will enable more robust natural experimental studies, particularly in Europe where there is a dearth of evidence. More robust natural experimental studies like these are now needed to better inform policy and practice recommendations on the (in)effectiveness of a wider range of urban green space interventions. There is a need for better theory to understand how urban green space interventions bring about their effects, especially focusing on the physical and social contextual factors that influence whether interventions are likely to work or not.
Date of Award31 Dec 2021
Original languageEnglish
Awarding Institution
  • The University of Manchester
SupervisorJamie Anderson (Supervisor), Sarah Cotterill (Supervisor) & David French (Supervisor)


  • Risk of bias
  • Urban green space
  • Systematic observation
  • Physical activity
  • Natural experiment
  • Wellbeing

Cite this