I'm doing a statistical sampling to measure how effectively employees answer inbound telephone calls for a company. I need an expert opinion on how statistically valid my approach is and any possible suggestions on making my statistical estimates more accurate. I have a list of measurable items employees answering the phone are expected to accomplish during the typical phone call (e.g. using a certain type of initial phone greeting, asking for specific information�). My goal to determine how effectively the employees are accomplishing this list of items during their phone conversations by listening to a sample of recorded calls during a one month period. Each item on my list for the employee to accomplish is very binary they either did it or they did not (yes or no) there are no grey areas. Here is how I did a statistical sampling of calls for the company: In one month the company received 2644 inbound calls. I used this survey random sampler calculator http://www.custominsight.com/articles/random-sample-calculator.asp I entered a margin of error of +/- 6% with a 95% confidence level with a total population of 2644 calls. The results showed I needed to listen to a sample of 242 calls to achieve my margin of error and confidence level. Then to ensure I was getting a �true random sample� I went to http://www.random.org/integers/ to get a list of 242 random numbers for my population of 2644 calls for me to listen to and determine if they accomplished my list of measurable items. Then I put each 242 call's measurable results (from my listening and grading each call from my list of items) on a spreadsheet so I could manipulate the data to determine the call effectiveness results in different areas for the company for the month. My question: Is the approach I'm using a statistically accurate way to get a sample to effectively estimate the results of the entire population of calls? Also in the margin of error I chose +/- 6% and confidence level of 95% seem like valid numbers for this type of statistical survey. And is there something I'm missing or doing wrong or any suggestions you might have regarding my approach to ensure a more effective level of statistical sampling accuracy? Can you explain this sampling approach in statistical sampling terms (why this approach is statistically credible) so I can give the explanation to a statistical sophisticated third party? And if not what do I need to change to make it highly statistically credible to fit your explanation? Also I may have some brief follow up questions after I receive your opinion. Thank you Keywords: Statistic, Support, Marketing