Dear all,

I have a household survey dataset consisting of two rounds. The dataset provides information on the month and the year of the interview. In round 1, a household could have been interviewed either in 2004 or in 2005 in any month. In round two, a household could have been interviewed either in 2011 or in 2012 in any month.
Additionally, I have monthly weather data for every year (e.g. T2004_1 for average temperature in January 2004).

I would like to generate a variable that captures the following information for every household for every round: average spring (January-March), summer (June-September) and winter (October-November) temperature and precipitation preceding the interview.
So for instance, for a household that was interviewed in round 1 in December 2004 the variables would be generated for the year 2004.

However, a household that was interviewed in June 2004 would draw on both, information from 2003 on winter and summer weather and 2004 for spring weather. I am however not really sure how to generate this variable. Could anyone please help?

Many thanks.