group by - Access query: TOP 1 with TOTALS is doubling up my SUM -

when run simple query in ms access, single table, doubles sum column.

select top 1      data.item,      reference.[proposed new category] type,      sum(data.[qty sold]) [nbr sold]      data  inner join      reference  on      data.item = reference.[item name]      data.[month date] = [type 1st of month want - aug 2013, type 8/1/2013] group      data.item, reference.[proposed new category] order      sum(data.[qty sold]) desc; 

yes, saw both of these, please don't muss thread referring them unless it's constructively explain why applicable. neither of them addressed situation (one self-join , other 1 joins - have no self joins).

cumulative totals being doubled
access 2010 doubling sum in query

why doubling inherent in query, , how can avoid it?? want top 1 item sum of qty sold descending...sounds simple. i'd avoid doing bunch of typical "access junk", make table queries etc., there way around it?

one thing cause if join many-to-many relation not one-to-many relation intended, in case inner join return more intersection of 2 tables in weak sense of 'intersection'. fact joining on text field makes me doubly suspicious - presumably [item name] isn't primary key of reference? 1 way test hypothesis amend query force one-to-many relation:

select top 1 data.item, r2.[proposed new category] type, sum(data.[qty sold]) [nbr sold] data inner join (     select reference.[item name], first(reference.[proposed new category])      reference     group reference.[item name]) r2 on d2.item = r2.[item name] (((d2.[month date])=[type 1st of month want - aug 2013, type 8/1/2013])) group d2.item, r2.[proposed new category] order sum(d2.[qty sold]) desc; 

if comes expected value, need fix reference remove duplicate entries. i'd recommend adding proper primary key, if there isn't 1 there ('item name' doesn't sound particularly primary key-ish, whereas 'itemcode' or 'itemid' would) , making relationship data against that.