This is the only way I can conduct interventions as I work in a p-12 college (school). It is helpful if you can find a control group and data collected across all groups, not just the group you are working with. That way you have fabulous pre intervention data and data for a control group, ipso facto. Sometimes I use data that relates broadly, e.g an online Math test as it is at least controlled data. Then intervention can be tested and assessed.

If possible it is good to get an observer to come and have a look at how the intervention is being conducted to see if there is a rigour of practice.

Yes, its always better to conduct such studies to make everyone on same understanding. I used this for MBA students as pre-induction course and analysed the results. These helped me to conduct the next sessions.

I recommend this article about blended learning for you

http://www.irrodl.org/index.php/irrodl/article/view/2032/3270

the first question is what are you trying to achieve? because depending on the answer to that question there are several methodologies that you can use.

Thank you very much for all your responses. I can see this is how experts share their knowledge and skills to help others. I planning to work on ways suggested. For Ahmed Mohamed Fahmy Yousef, thank you so much for sharing the article. Well job done.

Haruni, interventions can test a theory of causation.

Let's imagine a theory that the intervention will change distractions, so she can concentrate longer, see more information, build a better conceptual model, which she can recognise in other contexts, enabling her to solve related problems. 

Many things could happen to her that breaks this chain, or she could use other knowledge to solve problems. Measurement of her performance is weak evidence of learning. To confirm our theory, we usually need to ask her how she solved a problem, and why. We get more valid data if we ask open-ended questions than if we offer choices. 

If we can isolate just one link in the chain of causation, we can design a measurement such as error rate during an on-screen task, and look for statistical convergence.

Otherwise, when we are researching holistic learning performance, we need to look for the variety in learners' stories, and allow the learners to make their own choices to continue or exit the intervention. School teachers sometimes use Action Research methodology for developing interventions.

In evaluating commercial training, some people use The Success Case Method, for these reasons.

Brinkerhoff, R. O. (2005). The Success Case Method: A strategic evaluation approach to increasing the value and effect of training. Advances in Developing Human Resources, 7(1), 86-101.

http://deakinprime.com/media/121745/sc-method_ahrd.pdf

Dear, Russel Waldron, thank you so very much. Knowledge is sharing. I acknowledge your participation and sound response.

• Blended Learning with Lessons Learnt  in Higher Education 
• blended learning is a spectrum of learning modes that range from the traditional f2f classrooms to fully online degree programs”.
• Is a blend of tools and/or a blend of philosophies with the goal to have the right theory for the right situation.
• Blended learning with increased support and visible tutor monitoring can improve module retention by motivating learners to complete coursework on time. 
• Blended learning offers several inter-related ways of improving retention.
• Identification and monitoring of ‘at risk’ learners while a course is running and intervening promptly is one way, but improving student learning and a sense of belonging are also potential benefits of blending

http://www.herdsa.org.au/wp-content/uploads/conference/2011/papers/HERDSA_2011_Torrisi-Steele.PDF

https://pantherfile.uwm.edu/simonec/public/Motivation%20retention%20articles/Articles/hughes.pdf