Observation schedule is the best way to find the impact of the factors you have mentioned.

I agree with Chakraborty.. follow ethnographic research procedures..

Before I give my comment I would like to make clear for me what constellation is thought of. I assume you can observe school children (of what age?) who attend school and who attend the Children's club in a parallel mode. The research question then is whether the attendence of the Children's club, where the children are exposed to extra-curricular activities like sports, art and culture (is nothing of those three activities offered in school?) has impact on communcation skills and leadership competences of the children.  At this point, I agree to the two other answers saying that observation is the appropriate approach. Observation is the only way of trying to identify effects. But observation and identification of effects does not necessarily mean that the effects can be measured in a way that the observation and identifcation findings can be objectified in the sense of causal effects. In order to be safe with observation results it would be necessary to have  a control goup, i. e. a group of school children who would not attend the Children's club. On the other hand, the research should also try to control for other variables like gender, may be age, and some social resp. family background variables. If there were no control group because all school children attend the Children's club, the observations which surely can be made should end in terms of prudent hypotheses.

In order to discuss the measurement issue as you put it:

"How can we measure the impact of out-of-school intervention among school children?"

- its important to know if you are doing a few case studies with several villages or if you are doing it at a regional or national level. Using ethnographic research methods could be useful in the former situation but could be "costly at a National level".  Do you have diversity among children and their communities? Are you taking this in account? You can not ignore the context of each school community (economic resources and parents education, among others. You can use pre and after scales to measure interpersonal communication and leadership (try to find from other authors and adapt and validate). Tis article could be useful:

https://www.academia.edu/2051723/Interpersonal_Communication_Scale_ICS_

Let me know if I can be of service.  I have continued my research on interpersonal communication. I have isolated 10 factors to date using SEM.  I am working with a University of Central Lancashire's Social Science PHD candidate and a University of Pittsburgh Masters level student on research in Europe and China.  Macro level testing is an interest of mine.  I am looking for researchers around the globe to look into the portability of my newest index (Campbell Communication Scale) and feed back on a new communication theory (teaching communication) for social sciences (social work).

Best regards,

Jared

Jared,

I'm starting a project with children aged 9-11 and I would like to test your index.  How can we proceed?

Thank you

Yes you may.  I have moved on from that index for a more comprehensive one.  It was tested within the adult population, not sure how it will respond.  The internal validity was marginal at best, I continued the project by moving towards components of communication.  I have not explored the internal/.external construct.  I believe there is merit, let me know what you find.  Good luck!  Let me know if you would like to collaborate, I enjoy working with other researchers.

Best regards,

Jared Campbell

912-512-0052 (US mobile number) 

In my opinion, longitudinal qualitative studies have more value in this area than quantitative studies. I feel this way because qualitative studies can identify the perceived effect as experienced by the subject in their own words. This data can then be compared to social and economic measures already being gathered by other researchers (both governmental and NGOs) to form a more complete picture of the human experience--the more complete the picture, the less vulnerable the data will be to agenda driven misuse by the media and politicians.  I'm aware this is restating some of the opinions above, I just wanted to add a little "why" to the answer in plain language.

Thanks,

A.B.

That has merit.  Open ended questions can help build the conceptualization of the project or study.  You may also want to utilize a mixed methods approach.  Remember not everyone is right-brained. :)

Measuring change for communications will require a control group, or in situ observation in an initial phase of a time series study for the same group. Once any training objectives are introduced into the program, then your observations for a baseline are completed, and you can then "test" the outcomes of the program for your study. Don't confuse attendance, which is an indicator of output, with results derived from exposure to "program", which are outcomes measuring change, not just activity.  Participation in activities is required for results to happen. However if skill attainment is necessary and accumulative as program objectives, again a time series will cover that accumulative gain if measured at regular intervals after exposure. If it is a new program, that approach can also be helpful in seeing discretely what parts were successful and whcih were not (a process evaluation technique).

Leadership is the most overused term in vogue today, so it means anything to anyone who uses it, but clearly defined skills are necessarily structured into the program.

If you wish to track changes for in school performance during the after- school program, a set of associated performance measures can reflect outcomes if they are derived from the after school activities. Sociability measures come to mind.

I concur with other commenters on qualitative vs quantitative -  if measuring is  from scratch, best to start with qualitative approach to define the scope of the inquiry, test the indicators or questions for language and cultural fit, and then use a quantitative approach for measuring results. If the group is smaller than 30 forget the quantitative approach given most (but not all) statistical tests will be of low power and not provide much programmatic support from funders, i.e they won't be impressed by results even if tests show significance.

It is for the first time, I have made this kind of request. Thank you for everyone who could guide me. I will come back after some time.