Consequently, sample size was determined using the method of “proportional allocation” of the number in each stratum. The size of each stratum was as follows: doctors: 350; nurses: 450; administrative staff: 380; and medical students: 683. The size of each sample was 5% of the people surveyed. We observed a limitation of n/N < 0.10 (Ν is the size of people and n is the size of sample), this is required so that sample taking is considered to be independent and the sample is considered to be random. It also enables us to avoid “finite population correction” [41]. We obtained a final sample of approximately 6%, which is close to the original target of 5% of the total population occupied in the hospital. This meets the aforementioned requirements and additionally gives a value of n = 102 (doctors 21, nurses 23, administrative staff 24, and medical students 34), which is also accepted by the following formula for n-optimum: 22 211 a n N S d z N
where S2: Dispersion, 1-a: level of trust, and d: area of error. For acceptable values of error d departure of estimators from estimated data, as well as trust level 1-a. S 2 factor has been estimated from a first sample [42].