An examination of the construct validity of experimental suicide images among university students

Responses to images depicting staged suicidal behaviour have yielded some insight into cognitive-affective processes related to suicidal thoughts and behaviours (STBs). However, the quality and standardization of images has been inconsistent, and their psychometric properties remain unclear. Research has not systematically investigated whether images accurately represent the construct of suicide. This study validated a set of suicide images using ratings of how much they depicted someone “trying to kill themselves on purpose or who did kill themselves on purpose” (i.e., suicide ratings). Participants were 264 university students (Mage = 20.72) who were predominantly women (79.17%), White (56.44%), and heterosexual (73.86%). Suicide images demonstrated construct validity at image and aggregate levels. Further, suicide images looked more like suicide than pleasant, neutral, and interpersonal violence images, bs ≥ 5.653, ts ≥ 52.505, ps < .001, and suicide images without gore looked more like suicide than suicide images with gore, b = 0.269, t = 7.714, p <.001. Suicide ratings were higher for images depicting high lethality methods (e.g., hanging, firearm), bs ≥ 0.235, ts ≥ 3.316, ps < .001, compared to the grand mean of all methods. Suicide ratings were not associated with STBs after adjusting for multiple comparisons. This study provided detailed norms on commonly used suicide images and examined whether suicide ratings vary based on image features and clinical characteristics. Using valid suicide images could enhance strong conclusions about processes related to suicide. Our results may improve ability to predict STBs with potential downstream implications for clinical assessment and intervention.