Matching Action and Control Sites

Download PDF here: Matching Actions and Control Sites

The Grippa clips are designed to reduce opportunities for bag theft. Evaluating their impact needs to be measured at two levels:

a) Differences in bag theft levels between action and control sites as a whole, i.e. pooled across pubs, and,

b) Matched pairs of 1 action pub with 1 control pub (a total of 13 pairs – 26 pubs). This accounts for the ‘moving’ pre/post intervention period across action pubs (i.e. the time when the clips are to be fitted), this being inevitable due to the sequential process of implementing the design-based intervention in one pub after another.

The separation of which sites are to be action and which control was already determined through discussion with the bar management team. Here we describe the various analyses required to suitably match pubs within those groups. Specifically we show:

i) The volume of bag theft across pubs, and the distribution within each group (action and control).

ii) The trajectory over time of bag theft for action and control groups, and between the 2 pubs with the largest volume of bag theft within each group.

iii) The number of seating opportunities per pub.

iv) The bag theft rate per year using number of seating opportunities per pub as the denominator.

A set of action and control bars are then matched. These couplets are then assessed for disparities in light of a series of pub management surveys on extant crime prevention measures in the pubs under analysis.