Fig 1.
Tax compliance by treatment and citizenship status.
Panel a shows the average tax compliance ratio for participants in the control, icon and sport treatment group, decomposed by participants’ citizenship status. Number of participants: Ncontrol,non-Aus = 35; Ncontrol,Aus = 25; Nicon,non-Aus = 16; Nicon,Aus = 31; Nsports,non-Aus = 8; Nsports,Aus = 23. Panel b shows the average tax compliance ratio of Australians participants decomposed by country of birth, i.e., born in Australia and born elsewhere. Number of participants: Ncontrol,Aus born = 16; Ncontrol,born elsewhere = 9; Nicon,Aus born = 25; Nicon, born elsewhere = 6; Nsports,Aus born = 17; Nsports,born elsewhere = 8. Error bars represent 95% confidence intervals.
Fig 2.
Effect of video framing on tax compliance.
Panel a shows the ordinary least squares (OLS) regression coefficients of the interaction terms between the treatment variable (icon and sport) and citizenship status (Australian) on participants’ tax compliance ratio. In panel b, using the Australian sample, we show the coefficients of the interaction terms between the treatment variable and a dummy variable which indicates if the participant was not born in Australia. Blue circles represent the point estimates of the unconditional regressions without controlling for other factors. Red squares show the point estimates from regressions that control for participant’s age, gender (female dummy), church attendance, field of study (Business dummy), degree (post graduate dummy), and disposable income. We further control for experimental features such as round fixed effects, and whether the participant was audited or fined in the last round (green diamonds). Error bars represent 95% confidence intervals based on robust standard errors.
Fig 3.
HRV and change in HRV before and during the video triggers (by citizenship status).
Comparison of the average normalised LF/HF ratio (stress indicator) before and during the video framing to determine how participants react to the trigger/treatment. The green bars indicate the average LF/HF ratio before the video, the orange bars indicate the LF/HF ratio during the video, and the purple bars indicate the difference. A negative difference in the purple bars indicates that the LF/HF ratio decreased, due to a relative change in the PNS/SNS ratio, which indicates that participants were relaxed by the treatment. Error bars represent 95% confidence intervals.
Fig 4.
Tax compliance by treatment, citizenship status, and HRV.
The red bars indicate the tax compliance ratios of Australians who experienced a relative increase (darker shades) in the LF/HF ratio above the treatment specific cut-off and who demonstrated a relative decrease (lighter shades) in the LF/HF ratio (below the cut-off), respectively. The blue bars are tax compliance ratios of the non-Australians who were excited (darker shades, ΔLF/HF above treatment specific cut-off) and relaxed (lighter shades, ΔLF/HF below cut-off) by the video.
Fig 5.
Tax compliance by treatment, duration of stay in Australia, and HRV.
Panel a shows the average tax compliance ratio of non-Australian participants by their length of stay (over one year or less) in each treatment group. Error bars represent 95% confidence intervals. In panel b, we further separate the sample into two groups by their physiological reaction to the framing video (ΔLF/HF ratio above or below cut-offs).