Measurement is fundamental to understanding and preventing bullying, but approaches in the field are inconsistent, producing much conflicting evidence. We illustrate this by demonstrating the sensitivity of findings to researcher-led analytical decisions (exposure threshold and type(s) of bullying considered) in a study addressing the following aims: (i) to determine the prevalence of bullying; (ii) to establish the nature and extent of inequalities in bullying exposure between different socio-demographic groups; (iii) to examine the relationship between bullying exposure and internalising symptoms; and (iv) to establish if this relationship varies between socio-demographic groups. Adolescents aged 12–15 (N = 35,825) attending 147 secondary schools in the #BeeWell study completed measures of bullying and internalising mental health difficulties. These data were linked to information on their socio-demographic characteristics (e.g. socio-economic disadvantage). A series of pre-registered analyses were undertaken. With regard to the first aim, the prevalence of bullying victimisation was found to range between 5 and 16%. In relation to the second aim, disparities in exposure to bullying were consistently found among gender and sexual minorities (vs cisgender heterosexual boys), those with special educational needs (vs those without special educational needs), younger students (vs older students), and those from more disadvantaged neighbourhoods (vs those from less disadvantaged neighbourhoods), irrespective of the bullying exposure threshold or type being considered. However, disparities among cisgender heterosexual girls (vs cisgender heterosexual boys) and ethnic minority groups (vs White students) varied by exposure threshold and type of bullying. Pertaining to the third aim, the population attributable fraction for the association between bullying exposure and internalising symptoms was found to range between 6 and 19%, with the odds ratio ranging between 3.55 and 4.20. Finally, in terms of the fourth aim, there was limited evidence that the magnitude of the impact of bullying victimisation varied across socio-demographic subgroups, except that bullying exposure was more strongly associated with internalising symptoms among LGBTQ+ young people and cisgender heterosexual girls (vs cisgender heterosexual boys), and less strongly associated with internalising symptoms among Black students (vs White students). Our findings speak to the importance of developing more consistent measurement practices in bullying research, with consequent implications for prevention and intervention. These implications are contextualised by consideration of study strengths and limitations.