This study made an estimate of the effect of psychological and social work factors over time on neck pain. A sample of Norwegian employees (n = 1250) was surveyed on 3 occasions spanning 4 years. Five exposures were studied: quantitative demands, decision control, social climate, empowering leadership, and role conflict. Group-based trajectory models suggested factors changed little over time—it was possible to classify employees by mean levels of exposure into groups exhibiting stable “high”, “middle”, and “low” levels. The exception was decision control, for which a 4-level classification was derived: “high”, “high–middle”, “low–middle”, and “low”. Pain prevalence at the end of the study period was compared across groups. Risk and prognosis were also assessed separately by dividing the sample according to pain status at baseline. For all factors, distinct differences in risk were observed between exposure groups, controlled for pain at baseline, skill level, sex, and age. Statistically significant effects ranged from 0.38 (CI 0.20–0.73, P < 0.01) for high social climate to 3.00 (CI 1.63–5.50, P < 0.01) for high role conflict. The risk of new-onset pain was predicted by all factors. ORs ranged from 0.32 (CI 0.16–0.67, P < 0.01) for high empowering leadership to 2.61 (CI 1.09–6.21, P < 0.05) for high role conflict. Pain persistence was predicted by high role conflict (OR 3.26, CI 1.30–8.18, P < 0.05), high quantitative demands (odds ratio [OR] 3.66, CI 1.58–8.49, P < 0.01), and high–middle decision control (OR 0.45, CI 0.21–0.99, P < 0.05). Future studies should gather information at multiple time points to clarify the effect of prolonged and changing exposure on musculoskeletal pain.