Background　It　has　been　suggested　that　colorectal　cancer　be　regarded　as　several　subgroups　defined　according　to　tumor　location　rather　than　as　a　single　entity.　The　current　study　aimed　to　identify　the　most　useful　method　for　grouping　colorectal　cancer　by　tumor　location　according　to　both　baseline　and　survival　characteristics.　Methods　Cases　of　pathologically　confirmed　colorectal　adenocarcinoma　diagnosed　from　2000　to　2012　were　identified　from　the　Surveillance,　Epidemiology,　and　End　Results　database　and　categorized　into　three　groups:　right　colon　cancer　(RCC),　left　colon　cancer　(LCC),　and　rectal　cancer　(ReC).　Adjusted　hazard　ratios　for　known　predictors　of　disease-specific　survival　(DSS)　in　colorectal　cancer　were　obtained　using　a　Cox　proportional　hazards　regression　model.　Results　The　study　included　57847　patients:　43.5%　with　RCC,　37.7%　with　LCC,　and　18.8%　with　ReC.　Compared　with　LCC　and　ReC,　RCC　was　more　likely　to　affect　old　patients　and　women,　and　to　be　at　advanced　stage,　poorly　differentiated　or　un-differentiated,　and　mucinous.　Patients　with　LCC　or　ReC　had　better　DSS　than　those　with　RCC　in　subgroups　including　stage　III　or　IV　disease,　age　＜=　70　years　and　non-mucinous　adenocarcinoma.　Conversely,　patients　with　LCC　or　ReC　had　worse　DSS　than　those　with　RCC　in　subgroups　including　age　＞70　years　and　mucinous　adenocarcinoma.　Conclusions　RCC　differed　from　both　LCC　and　ReC　in　several　clinicopathologic　characteristics　and　in　DSS.　It　seems　reasonable　to　group　colorectal　cancer　into　right-sided　(i.e.,　proximal)　and　left-sided　(i.e.,　distal)　ones.