# how to pool data from 2 different surveys for two populations

I would like to ask if I can pool the data set from two different surveys, taking into account that the first survey population is 328 staff members of the hospital as (independent variable: six-sigma methodology). The second surveys population is the inpatients in the same hospital, numbering 540 patients as (dependent variable: patient satisfaction). Both surveys consist of five dimensions (define, measure, analyze, improve, control). While the survey of workers with 68 questions and the survey of patients with 36 questions. Both surveys were designed according to five-dimensional Likert scale. My question is: Can I pool the two data-files from both surveys into one dat- file and then calculate the relationship (correlation) or regression between the independent variable (six-sigma methodology) and dependent variable (patient satisfaction)? If, Yes, How I can do this?

• No. Based on your description there is no way to do this. Jun 10, 2017 at 13:20

There is no way to pool the two surveys you have described.

Pooling for comparison is done when two different groups, treatments, etc have had the same measurements made and the two groups are to be compared on those measurements. This might use two sample tests or possibly logistic regression.

Pooling for improved estimation is done when there are two different samples of a population which have had the same measurements made. In this case, the sample size is increased and estimates of rates or means can be made that have smaller standard errors than in either single sample.

Pooling of samples taken across time, eg, annual surveys, is done in order to use multiple samples to measure the pattern of some measurement over time. Again, the measurements must be the same.

Sometimes, different information about the same subjects is taken from different sources and matched into a single set of data for the purpose of finding correlations or differences. This isn't usually referred to as pooling but as matching. The measures can be different but the subjects must be the same.

You appear to have two different groups of subjects, each with its own measurements, taken at about the same time. There is no way to pool the two groups since you do not have the same measures on the two groups.

• Thanks Mr. David. According to the information that I mentioned above, is there any way to calculate the correlation or regression between 2 group's data. Jun 11, 2017 at 16:17
• This would require that both groups be asked the same questions. Jun 11, 2017 at 16:26
• Even though, both surveys have the same 5 domains, also at the same hospitals and in the same period time. Jun 11, 2017 at 16:33
• Yes. Most definitely, even though, and anything else you might add. Jun 11, 2017 at 16:43