Diurnal type (chronotype) differentiates individuals on an axis between the extremes of evening type to morning type. These diurnal-type preferences are thought to be relatively stable, but follow-up studies are sparse. The study aims were (1) to compare cross-sectional studies of diurnal type preferences between two decades and (2) to analyze the consistency of diurnal-type preferences using a longitudinal dataset. We analyzed a total of 18 087 adult males from four datasets with information on diurnal type and age. Of these, 2144 were available for survival analysis and 567 for analysis of longitudinal diurnal consistency. Diurnal type was assessed by asking the individual to what extent they would rate themselves a morning or an evening person, categorized into four groups. Statistical tests for stability of diurnal type were based on transition matrices and p values obtained using likelihood ratios. Cox regression was used to calculate the relative risk of all-cause mortality in each of the four diurnal type groups. After direct age standardization, 9.5% (95% CI: 9.0–10.1%) of participants in the four datasets were evening types. The cross-sectional data yielded that morning types were less common in the 2000s than two decades earlier. The longitudinal dataset revealed a significant shift from evening type to another type from 1985 to 2008 (p = 0.002). The relative risk of all-cause mortality was 1.3-fold (95% CI: 1.0–1.6; p = 0.05) higher for evening types compared to morning types. At the population level, eveningness appears to have become more prevalent over recent decades. However, on the individual level, the more morningness the chronotype, the more persistent it remains with aging.