So for example, we have the following polls:
First poll: People who said they like orange juice for breakfast are as follows: Texas (50%), NY (25%), North Carolina (20%), and Florida (5%)
Second poll: People who said they prefer netflix to watch movies are as follows: Texas (10%), NY (70%), (North Carolina (15%), and Florida (5%)
How can we aggregate these polls to say that, for instance, the average person in Texas is likely to be interested in orange juice for breakfast, but likely does not like netflix for movies.
So aggregating data sources to reflect them onto an individual
Please clearify your question more. Do you have the actual data collected from two sources if any one variable from these two surveys is common then one can simply merge the two data sets in SPSS.
There is no definite answer to this question but I am adding simple answer which may be helpful. If you are looking for something fancy then you may Google the following : "data matrix fusion". Below is my answer.
To find out the following: "What is the probability that an average person in Texas is interested in orange juice for breakfast and does not like Netflix for movies?"
Ansewer:
Probability that an average person in Texas interested in orange juice for breakfast = 0.5 (from the data given in the example).
Probability that an average person in Texas does not like Netflix movies =1-0.1=0.9
Therefore, probability that an average person in Texas is interested in orange juice for breakfast and does not like Netflix for movies = 0.5 x 0.9 = 0.45...
You may extend the above to different cases e.g. another case may be, " What is the probability that an average person in Florida is interested in orange juice for breakfast and does not like Netflix for movies?
Once you exhaust all the cases, you have a new transformed data-set which you can use in further downstream analysis.
A good solution is to group these numbers by the US states, and run as much as possible probabilities computations, plot that and try to find out some insights.
Also, Bayesian probability may be used to reveal more information and get a better understanding of the data.
This data can help us predict whether people who watch Netflix are likely to drink orange juice for breakfast and vice versa, its causes, and how it can be related.
After this is done, more specific data ought to be collected to define this behavior causes.
- Badges
- Science topic