Hello everyone, I am a stats-beginner and I need help to understand if it is possible to compare two datasets.
I am currently looking at the Residential Energy Consumption Surveys (RECS) databases, specifically to the 2015 and 2009 surveys. There have been some major changes in the 2015 survey that prevents me from merging the datasets to directly compare them:

- In 2015 they added many questions and also changed some categorical variables to open-ended responses;
- In 2015 they double checked the self administered questions on energy consumption with the billing data provided by the suppliers to account for the mismeasurement of the householders;
- In 2015 there is less than half of the observations of 2009, making 2015 representative only at a US level and not a state level, differently from 2009;
- In 2015 the used an engineering estimation approach: instead of estimating unknown parameters for each end-use model, they relied on published values for those parameters. Also, the calibration uses a minimum variance estimation approach, whereas in 2009 they use a simple normalization method.

Given these differences, I do not know if it is possible to compare these datasets, or at least merge them (for the matching questions) and analyze a time series. I would be grateful if someone will help me to find out how to handle this problem!