Durbin watson

:confused: :confused: i'm in a big big problem?
i'm a demonestrator in medical statistics department
this week i have to prepare a seminar about assumptions of multiple regression
i cannot understant what Durbin-watson mean? i just feel it!!!!!!
i have many questions in my mind
Is it used when the data included is time series or with any other type of data?
how i correct for presence of autocorrelation?
by looking for an omitted variable or by differencing the data?
What this statment "Test for NEGATIVE autocorrelation: USE 4-d" mean?
I need help
i searched Google to read more and understand but i think i'm just feelling not understanding ?
Again most of the handouts i found illustration examples in economics field which is very difficult to me to understand, i am a doctor:(?:confused: