U8D1 Regression Discontinuity and Correlation

U8D1 Regression Discontinuity and Correlation Research Part I: (brief 1-2 paragraphs) Write about the logic of both the quasi-experimental regression discontinuity design and the non-experimental correlational approach and explain their limitations in permitting causal inferences. How are these two designs related?

Regression discontinuity is used when one cannot conduct an experimental study (randomized) and one can only go under the available data. In this case, both treatment and control group must be conducted under available data. What happens in this case is that participants with values of the observed variable above a certain threshold are assigned to the treatment group, and those with values below are assigned to the control group. They are no assigned randomly, but the measure underlying their assignment is observed instead.

The second assumption of the regression discontinuity is that the outcome variable is a continuous and smooth result of the observed variable, particularly near the threshold. (Urban Institute. ). The limitation here is that ultimately one cannot completely test the assumption; since one cannot randomize, and one has to depend on available data this is the best that one can do.

The non-experimental correlational approach is also conducted in a non-randomized manner where a certain behavior is observed along with another set of circumstances that may explain the behavior. The researcher/s observe the pattern and see whether one (the circumstance) shows a correlation with the other (the action). An instance may be whether increase in watching TV is associated with positive increase in aggression.

The limitation with correlation is that a third, underlying and subtle reason may be underpinning the association and, therefore, it is not so much TV that is causing a surge in aggression but rather an underlying third variable.

The two designs are related in that both do not work on randomized methodology and both seem to show demonstrative outcome but, given the data, they are the best assumptions one can make in that situation.

Part II. Briefly summarize the article and identify the variables.

The study assessed predictors of burnout and stress in a sample of 669 child welfare workers. The investigated predictors were personality characteristics and environmental factors. An online survey was conducted with respondents completing the Professional Quality of Life IV (Stamm, 2005) to ascertain compassion fatigue (CF) and burnout symptoms. Predictors of stress included being male, young, Hispanic, living in a rural residence, and not belonging to any religious institution. Also being male and young were associated with high burnout rates, whilst involvement in religious activity were associated with lower burnout rates.

The child worker whose job status was that of professional was more likely to show burnout and compassion fatigue than another not of that status.

The variables here were dependent variable -- burnout and stress, independent variable -- status of the worker as well as personality and environmental characteristics (namely those described above).. Explain how the research exemplifies regression discontinuity or correlation research, and identify the specific design, if appropriate.

This is an example of correlation research. The researchers wished to see whether any specific factors were associated with burnout and stress and whether the occurrence of these variables showed positive association (or negative association) with an increase or decrease of stress / burnout. Analyze the aspects of the research that enabled the researcher to draw conclusions about the results. Include a description of the threats to validity that are particularly relevant to this study, and describe what the researcher did to reduce these threats.

The researchers used a reliable instrument to assess the level of burnout of participants. They also noted applicable and exhaustive demographic data. The threats to validity may have been the fact that participant's stress was related to factors other than work. The researchers, therefore, chose an instrument whose function was to solely measure work-related stress.