Meyer Et Al. Meyer, Wang, Li, Thomson,

Meyer et al. Meyer, Wang, Li, Thomson, and O'Brien-Pallas (2009) used a convenience sample of nurses and patients from six Canadian hospitals. A convenience sample consists of participants that are drawn from sources using nonprobability sampling methods (Jackson, 2012). In this case hospitals with specific inclusion criteria were selected and case records were used. The study makes no mention of any type of random or stratified sampling outside the inclusion criteria for the hospital type (high volumes of patients in their cardiac units and certain case groups of interest).

In general there are two different types of sampling: random sampling methods and non-random sampling methods such as convenience sampling already discussed above. Random sampling methods allow researchers to extend the findings of the study beyond the sample of participants in the research study because these methods statistically control for differences in the sample and differences in the population from which the sample is drawn (Jackson, 2012).

In a true random sampling method each and every member of the population from which the sample is drawn has an equal chance of being selected for the sample, whereas in nonrandom sampling method this is not true. The study looked at various subject variables and predictors of patient outcomes such as staffing, nurse level predictors, patient level predictors etc. The study is a correlational design as there is no random assignment to the different treatment groups.

The sample size is always an important factor in a research study because sample size is a major influence on the power of the specific statistical test using quantitative studies (the power of statistical test refers to that test's ability to detect a significant effect or difference if indeed such an effect exists; Tabachnick & Fidell, 2012). In the current study the researchers performed a power analysis before actually collecting data that estimated the needed sample sizes to achieve significance and desired effect sizes.

The question of what types of data analyses are possible when conducting quantitative research is actually quite broad. In general, quantitative research is either descriptive or inferential (Tabachnick & Fidell, 2012); however, the types of descriptions and inferences that can be made depend on the research design. Descriptive research (statistics) summarizes certain aspects of the data such as the central tendency, shape of distributions, dispersions, etc. Inferential analyses (statistics) go beyond simple descriptions and describe relationships, associations, conclusions, etc. In the data.

Different methodologies designate the types of conclusions that can be made from the analyses. For example in the current study Meyer et al. (2009) employ a correlational design, therefore they are unable to make causal inferences but can describe the relationships, their relative strengths, and in some cases direction given the analyses. The type of data also influences the type of analyses one can do.

There are a number of different levels of measurement in the current study due to a large number of variables ranging from nominal (e.g., diagnoses or patient employment status), ordinal (e.g., employment status coded as full-time, part-time, or casual), to ratio level variables (e.g., years of nursing work experience). Some of the variables are categorical such as employment status and some variables are continuous variables such as age. Meyer et al.

(2009) used multiple data collection methods that included collecting hospital records, daily unit data, surveys, and patient data forms. In order to ensure that different data sources and collection methods were consistent they calculated the inter-rater reliability for all measures (which they claim was at 90% throughout the study). The use of surveys in the study was extremely important as surveys allow the collection of anonymous data (no identification on the part of the person that takes a survey so they are.