Module 3

Describe the types of conclusions possible when using random sampling and random assignment.

(ideal experiment) There can be a causal conclusion generalized to the whole population. Random sampling can obtain a sample representative of the population (generalized). Random assignment lets us know that any difference between groups are what we are studying (causal).

Describe the types of conclusions possible when using random assignment without random sampling.

(most experiments) Use causal conclusions, but only for the sample. The results cannot be generalized.

• Describe three sources of bias discussed in the Patten text.

- Failure to identify all members of a population - Using samples of convenience (accidental samples)- doing a psych experiment with only students in a psych class. - Volunteerism- (1) when researchers issue a call for volunteers, and (2) after random sample is chosen, those who actually choose to participate are volunteers and may be different from non volunteers because they may be more interested in technology and more successful in their educational pursuits.

Define sample.

A group that is taken out of the population to be tested

• Describe how one might implement simple random sampling.

A simple random sample is one in which every member of a population is given an equal chance of being included in a sample. One might implement simple random sampling by putting names on slips of paper and drawing them from a hat.

Describe how one might implement stratified random sampling.

A strategy to reduce bias. Stratification in conjunction with random sampling. First divide a population into strata (men and women). Researchers usually draw the same percent of participants from each strata. (If there are 600 women and 400 men in population, and sample size is 100, there will be 60 women and 40 men.) Precision can be increased by using multiple strata in selecting a given sample.

Describe how one might implement samples of convenience.

Also known as accidental samples. If a psych professor wants to study the principle of learning theory, but only uses students enrolled in psych class. The sample is biased against all other students. Bias can occur easily in this type.

Describe the importance of sampling criteria.

Defining the criteria increases the likelihood of a study to include a representative sample, which allows for generalizability of the findings from the study to the population. Criteria ensure a heterogeneous or diverse sample

Define sampling criteria.

Eligibility criteria include a list of characteristics or traits essential for membership or eligibility in the sample. Criteria are developed from the research problem, the research purpose, a literature review, the conceptual and operational definitions of the study variables and the study design.

Define sampling error.

Error created by chance (random sampling). Sampling errors are minimized if samples are of adequate size.

What is the primary consideration when judging the adequacy of a sample?

Even if sample size is increased, if all students tested are from the same inaccurate population, the sample will not change in accuracy.

Define exclusion criteria.

Exclusion criteria are those characteristics that disqualify prospective subjects from inclusion in the study. Example: abnormal renal function tests, if the combination of study drugs includes one or more that is nephrotoxic.

Discuss one major advantage of using cluster sampling.

In cluster samples, researchers might get a better response rate if he or she draws a sample of clusters.

Define inclusion criteria.

Inclusion criteria are characteristics that the prospective subjects must have if they are to be included in the study. Example: postmenopausal women between the ages of 45 and 75 years who have been diagnosed with Stage II breast cancer

Discuss the principle of diminishing returns with respect to sample size.

Increasing the size of a sample increases precision, but a larger sample is not necessarily better because of principle of diminishing returns. Researcher A starts with 50 and adds 50. Researcher B starts with 3000 and adds 50. Adding 50 to the population of 3000 will not make a difference because the results will already be overwhelmed by the results of the 3000. At some point, the returns diminish to the point that further increase in sample size is of very little benefit.

Does increasing sample size reduce bias?

No, Small, unbiased samples tend to yield more accurate results than biased samples.

Describe the benefit of using random assignment.

Observed effects can be attributed to the treatment, and we can make causal conclusions based on the study

Define random assignment.

Occurs in experimental settings where subjects are being assigned to various treatments.

Define random sampling.

Occurs when subjects are being selected for a study

• Discuss one major disadvantage of using cluster sampling.

One major disadvantage results from the fact that each cluster tends to be more homogeneous in a variety of ways than the population as a whole. (to avoid this problem, researchers should draw a large number of clusters or stratify on geography. )

Describe how one might implement purposive sampling.

One might implement purposive sampling by purposively selecting individuals who they believe will be good sources for information. For example, researchers observe that members of the academic senate vote on the winning side of controversial issues. Researcher decides to only interview winners to predict outcome of new issue. This is dangerous because professors may change their orientations.

Describe the relationship between quality of sample and quality of inferences made from a sample to the population.

Poor quality of a sample affects the quality of the inferences made from a sample to the population. A poor sample is likely to lead to incorrect inferences. When evaluating a sample, researchers ask if size is adequate and if sample is biased.

What is the symbol for population?

Population- N

Define unbiased sample.

Researchers can select by giving each member of the population an equal chance of being included in the sample.

• Describe how one might implement cluster sampling.

Researchers draw groups of participants instead of drawing individuals. Researcher surveys a sample of members of UMChurches in US. Researcher draws a sample of clusters (congregations) Clusters must be drawn at random.

What is the symbol for sample?

Sample size- n

Discuss the priority of unbiased sampling methods to large sample sizes.

Small, unbiased samples tend to yield more accurate results than biased samples.

Describe snowball sampling.

Snowball sampling can be useful when attempting to locate participants who are difficult to find. A researcher initially finds one individual that has not been tested and is difficult to find. When one person is found, the one participant might put researcher in contact with several other potential participants.

Describe how one might implement systematic sampling.

Systematic sampling is very similar to simple random sampling. Every nth individual is selected. The letter n can be any number. If it is two the researcher would select every second individual. Bias in this sampling style can occur if there is an arrangement in the order.

• Discuss one major advantage of using snowball sampling.

Technique is based on trust. Helps researchers to be able to study special population. Qualitative and quantitative researchers use this.

Discuss the advantage of using stratified random sampling in comparison to simple random sampling or systematic sampling.

The advantage of stratified is its ability to ensure that different subgroups are represented in the correct proportions. Stratified samples account for the percentages of differences in a population, and simple random sampling doesn't account for that. Stratified accounts for sampling error.

Describe the benefit of using random sampling.

The benefit of using random sampling is that each subject in the population is equally likely to be selected and the resulting sample is likely representative of the population. Results are generalizable to the population.

Describe the primary purpose for using stratified random sampling.

The goal in stratification is to obtain a single sample that is representative in terms of the stratification variables.

Define population.

The group in which researchers are ultimately interested, that samples are taken out of.

Describe the types of conclusions possible when using neither random sampling nor random assignment.

Un-ideal (bad) observational study. Can only be used to make correlational statement, not generalizable statements.

• Describe the types of conclusions possible when using random sampling without random assignment.

most observational studies, No causal conclusion, correlation statement generalized to the whole population. Yes can be generalized to the whole population.

Does increasing sample size increase precision?

yes, Increasing sample size increases precision. Results will only vary by a small amount from sample to sample. Researchers should still try to first obtain an unbiased sample and then seek a reasonably large number of participants.