Skip to main content

Table 3 Linear regression models testing message effects on policy support and anti-industry beliefs at Time 2 (n = 3285)

From: Competing with big business: a randomised experiment testing the effects of messages to promote alcohol and sugary drink control policy

  Time 2 outcome measures
  Target policy supporta Anti-industry beliefsb Average non-target policy supportc
  B (95% CI) β p B (95% CI) β p B (95% CI) β p
Message condition
 Control Ref    Ref    Ref   
 Standard pro-policy arguments (Standard) 0.26 (0.06, 0.47) 0.05 .013 0.08 (−0.05, 0.21) 0.03 .255 0.16 (0.01, 0.32) 0.04 .041
 Standard + inoculation 0.31 (0.10, 0.52) 0.06 .003 0.16 (0.03, 0.29) 0.05 .019 0.22 (0.06, 0.38) 0.06 .007
 Standard + narrative 0.32 (0.11, 0.53) 0.07 .002 0.14 (0.01, 0.27) 0.05 .034 0.27 (0.11, 0.43) 0.07 .001
 Standard + inoculation + narrative 0.28 (0.07, 0.49) 0.06 .009 0.07 (−0.06, 0.20) 0.02 .268 0.11 (−0.05, 0.27) 0.03 .181
Covariates
 Health policy assignment
  20% tax on sugary drinks Ref    Ref    Ref   
  Removal of sugary drink sponsorship from sport 0.26 (0.08, 0.45) 0.06 .006 −0.15 (−0.26, −0.03) −0.05 .013 −0.47 (−0.61, −0.33) −0.14 <.001
  Volume-based tax on alcohol 0.21 (0.03, 0.40) 0.05 .025 0.18 (0.06, 0.29) 0.06 .003 0.32 (0.18, 0.47) 0.09 <.001
  Removal of alcohol sponsorship from sport 0.52 (0.34, 0.71) 0.12 <.001 −0.10 (−0.22, 0.01) −0.04 .086 −0.09 (−0.23, 0.05) −0.03 .195
 Age (years)
  18–24 Ref    Ref    Ref   
  25–34 0.28 (−0.04, 0.60) 0.05 .091 0.16 (−0.04, 0.36) 0.05 .121 0.16 (−0.09, 0.40) 0.04 .210
  35–44 0.15 (−0.18, 0.47) 0.03 .371 0.11 (−0.09, 0.31) 0.03 .287 0.04 (−0.21, 0.28) 0.01 .764
  45–54 −0.28 (−0.60, 0.03) −0.05 .080 0.12 (−0.08, 0.32) 0.04 .238 −0.05 (−0.29, 0.19) −0.01 .701
  55–64 −0.15 (−0.45, 0.16) −0.03 .343 0.12 (−0.07, 0.31) 0.04 .211 0.14 (−0.09, 0.38) 0.04 .223
  65 and older −0.13 (−0.43, 0.16) −0.03 .374 0.07 (−0.11, 0.26) 0.03 .433 0.29 (0.06, 0.51) 0.09 .013
 Days elapsed between surveys −0.01 (−0.02, −0.01) −0.06 <.001 −0.01 (−0.01, −0.00) −0.06 .001 −0.01 (−0.01, 0.00) −0.03 .061
  1. B unstandardised regression coefficient; CI confidence interval; β standardised regression coefficient; Ref referent category in linear regression model. Boldfaced results are significant at p < .05
  2. aParticipants’ level of support for their assigned policy at t2 which was recorded on a 7-point scale (1 = strongly oppose to 7 = strongly support)
  3. bFor participants assigned to a sugary drink policy, the anti-industry beliefs measured were that sugary drink companies: “deny that sugary drinks cause obesity”; “only care about making a lot of money”; “try to get young people to drink sugary drinks”. For participants assigned to an alcohol policy, the anti-industry beliefs measured were that alcohol companies: “deny they market their products to young people”; “only care about making a lot of money”; “try to get young people to drink alcohol”. Participants’ level of agreement with their three anti-industry beliefs at t2 were recorded on 7-point scales (1 = strongly disagree to 7 = strongly agree), and subsequently averaged to create this outcome measure
  4. cFor participants assigned to the “20% tax on sugary drinks” policy, the two non-targeted policies were removal of sugary drink sponsorship from sport and health warning labels on sugary drinks. For participants assigned to the “removal of sugary drink sponsorship from sport” policy, the two non-targeted policies were a 20% tax on sugary drinks and health warning labels on sugary drinks. For participants assigned to the “volume based tax on alcohol” policy, the two non-targeted policies were removal of alcohol sponsorship from sport and health warning labels on alcohol containers. For participants assigned to the “removal of alcohol sponsorship from sport” policy, the two non-targeted policies were a volume based tax on alcohol and health warning labels on alcohol containers. Participants’ level of support for their two non-targeted policies at t2 were recorded on 7-point scales (1 = strongly oppose to 7 = strongly support), and subsequently averaged to create this outcome measure