Hi Alexander,

Data from ~10 subjects require qualitative methods for analysis and interpretation. The research methodology depends on your research aim. A qualitative study is a good start for the development of a survey when you want to identify the right and best questions to capture the concept of interest, in your case "diffusion" for example. But if you want to use quantitative methods, once the survey/questionnaire is developed, or used in a new population, you need to assess its psychometrics, and a convenience sample of about 100 participants to do that. If your test has good psychometrics, it can be applied on large populations. Maybe the paper below helps:

Nelsona, Earleb, Howard-Grenvillea, Haackc, and Youngd, 2014. "Do innovation measures actually measure innovation? Obliteration, symbolic adoption, and other finicky challenges in tracking innovation diffusion". Research Policy, Volume 43, Issue 6, p. 927–940

Best

The Delphi Forecasting Technique (or 'expert consensus methodology') is a useful qualitative research technique for gathering opinion from experts, including reaching consensus about a topic or likely outcome. This paper neatly describes the application of the method in selecting the most important actions to include in policy guidance.

http://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-016-2822-6

Another suitable approach is the Key Informant Interview.  Although the paper below recommends a minimum of 15, I think of it as information value - the first 10 experts will add far more information to what you have now than experts 11 through 15.

http://pdf.usaid.gov/pdf_docs/pnabs541.pdf

Can I be blunt? This approach doesn't seem like a suitable Master's thesis idea.

Here's what I would do instead.

Follow the approach that RAND used (see link to free paper below) to estimate diffusion of information technology in healthcare. In the absence of any good data in the industry of focus, they basically looked at similar industries, and merely approximated trends in healthcare. Some triangulation with surveys and literature reviews. Much more juicy and defensible.

So to make up an example, suppose you're looking at the potential diffusion of the use of clusters of semi-autonomous micro-drones ('swarms') in civil defense by government purchasers.

Lord knows who can estimate that. But there are analogies you could drive. The underlying innovation came perhaps from hobbyists and gamers, and spread to more serious uses. Government purchasers adopted similar innovations from hobbyists (e.g. citizens band radios) and from gamers (e.g. capable tablet PCs with video cards) at certain knowable rates. Your purchasers are in an area where for economic, trade and political reasons, general technology diffusion across national borders is x, others countries do it a y, and voila, you can make some sensible predictions. Just making it up, but you see what I mean. If you can run your end-results past an informed group of respondents, so much the better...

longwoods.com/files/Diffusion_HIT.pdf