Once an organization has identified the social capital elements in its programs and specified pathways through which its activities contribute to social capital formation, it can formulate plans to evaluate its social capital impact. Somewhat artificially, we break down this process into steps 4 and 5: Step 4 involves the basic design questions, such as the study's timeframe and the choice of study populations; Step 5 concerns the nuts and bolts of fielding a survey-based evaluation. You can also access a list of social capital survey questions intended to help organizations develop their own survey instruments.

STEP 4: Designing the Evaluation

STEP 5: Conducting an Evaluation

There are a number of concrete issues that an organization fielding an evaluation still needs to resolve. This section considers five key practical questions listed below.

Together with the earlier "Whom to Study?" discussion, the answers to these questions largely determine the overall evaluation cost. Of course, if the resulting cost is too high, an organization may have to scale back, compromising on these four issues until the project is affordable. At that point, the organization will have to ensure that the scaled-down evaluation will still be worthwhile.

INTERVIEW FORMATS (expand)

HOW MANY CASES? (collapse)

Deciding how many interviews to conduct depends both on a number of technical factors and overall budget constraints. Obviously, no other design question has as great a bearing on the overall cost of the evaluation as the number of people surveyed, so organizations have a strong interest in minimizing the number of people surveyed. On the other hand, they do not want the number of respondents to be insufficient to reach a conclusion.

If you are interviewing everyone in the study group or control group, called a census sample, you don't have to worry that your data doesn't match the group as a whole. But if you are interviewing less than 100% of the study group or control group, such as a sample from the community, you must worry about sample size. Basically the size of the sample relative to the size of the study group (or control group) will provide you with a confidence interval (a range in which you are relatively certain that the entire group 's responses would lie if all the population had been surveyed). [This confidence interval, although mathematically easy to predict, is only one of four possible sources of error. Click here for a description of bias from patterns of non-response and nonrandom samples; validity problems; and reliability problems.

For those not interviewing everyone in your study group, click here for more specific guidance on sample size. However, we encourage you to consult a statistician (someone who is well versed in statistics) to advise you on these questions, as they can be complex.

Sample Size Recommendations for In-Person (not paper-based) and Telephone Interviews

If Population is:	We recommend a Sample of:	Approximate Sample Fraction	Confidence Interval
30	30	1	0
60	45	3/4	7.4%
100	65	2/3	7.2
200	98	2	7.1
500	163	1/3	6.3
1000	213	1/5	6.0
1500	313	1/6	5.5
2500	363	1/7	4.7
10,000+	Click here, but 500 individuals is probably sufficient

Note: if after the evaluation you are planning to compare one part of your sample against another part (for example men versus women), please read the section on stratification.

Recommendations for In-Person Paper-Based Surveys
Because the cost of additional cases can be quite small when administering in-person surveys on paper, we suggest that organizations stay with universal samples (in other words trying to interview all of the individuals in the group you are interested in) much longer, perhaps even up to 100 or 200 cases. But, see the discussion on "A Word on Census Samples".

MAKING SURE SAMPLES ARE RANDOM (AND NON-RESPONSE ERROR) (expand)

WHO SHOULD CONDUCT THE EVALUATION? (expand)

HOW TO CONSTRUCT THE QUESTIONNAIRE (expand)