How does the type of data collected and the way in which the data is collected affect the possibility of type 1 and type 2 errors?

Sample size is a direct function of what statisticians refer to as Confidence and Power of a test. 'Confidence' is 1-Prob(Type I error), or the probability of the test rejecting the null hypothesis when it should be rejected. That is, the chances of getting a true positive. 'Power' is 1-Prob(Type II error), or the probability of the test retaining the null hypothesis when it should be retained. That is, the chances of getting a true negative. The levels of confidence and power are arbitrary but generally set at 95% and 80% respectively. As sample size gets smaller so too does confidence and power (as long as your margin of error stays the same).

It is also important that sample data be collected at random from the population. That way each unit in the population has equal chance of being selected for the study. This reduces bias in your results. Bias can also lead to Type I and II errors, but it's harder to quantify if you don't know what the bias is.

So there is no real way of avoiding Type I & II errors (unless you take a census of the whole population). But you can reduce error by randomly selecting a large enough sample.