Also, the randomization of the assignment of participants to the experimental and control groups eliminate many threats to internal validity. Essay Question Three:

Operationalization can be a complex process. However simply put, operationalization is used to turn undefined, immeasurable variables into defined and measurable ones. For example, if a researcher developed a hypothesis that people who listen regularly to political talk radio tend to be less trusting of the federal government, several variables would need to be operationalized. First, the term "regularly" would need to be defined. Does regularly mean two times a week, four times a week, or seven times a week?

The next variable that would need to be operationalized would be "political talk radio." Would this term be limited to certain personalities such as Limbaugh or Beck? Or would it be limited to certain stations, certain sponsors, or specific times of day? The most difficult variable to operationalize in this scenario would be the level of trust, as this is a rather subjective concept. The researcher would need to either develop a list of parameters that measure trust, or use an instrument that has already been proven to be valid and reliable. This is extremely important for researchers who want to contribute unique and valuable information to the scholarly literature, as opposed to merely making hazy predictions with little concrete substance to support them.

Essay Question Four:

The essential feature that distinguishes a probability from a non-probability sample is that for the probability, type every member of the population to be represented has a known chance (greater than zero) of entering the sample and in a non-probability sample there is no such known chance. It is only under a probability sampling procedure that calculated significance levels can yield precise information about the probabilities of errors. Probability samples, in which all people in the population have an equal chance of being selected, are very much like a lottery, in which all numbers in the pool have an equal likelihood of being...

An example of a Non-probability sample is, the "man on the street" interview, in which the interviewer selects people at his own discretion. There is no randomness to this, because the interviewer may inject his own biases into the selection process. Furthermore, an interviewer standing on a street corner or in a shopping mall is not going to get a random flow of subjects, because people demonstrate purpose in their behavior. That is, they are there for a reason, and thus people who are not in that vicinity are excluded from the possible sample.
Non-probability sampling is used in situations where probability sampling techniques are either impractical or unnecessary, for example when sampling frames are unavailable or the population is so widely dispersed that cluster sampling would be too inefficient. Purposive sampling is a form of non-probability sampling where cases are judged as typical of some category of cases of interest to the researcher (They are not selected randomly).

Non-probability sampling does not provide the same technical strengths of statistical inference because it "introduces a level of subjectivity in the selection of the sample and this characteristic provides the challenge of defending the results for the researcher. Except for cluster sampling, the probability sampling procedures require that the entire sampling frame be identifiable. With cluster sampling, all the members of each selected cluster must be identifiable. Frequently, this requirement cannot realistically be met. Since some subjects in the sampling frame will never be selected in a nonprobability sampling process, the resulting sample is most likely to be biased in that it will not represent the population. Furthermore, since inferential statistics are developed with the assumption of a simple random sampling process and a nonprobability sample is a maximum departure from a simple random sample, therefore, inferential statistics will not be able to assess the possible sampling error with any meaningful degree of accuracy. Therefore, results from a nonprobability sample should always be interpreted with caution.