Advanced Analysis of Community Health Data Bivariate Analysis Worksheet – Description
The research problem is the foundation of any study. In identifying research problems, researchers articulate areas of concern or investigation. Once identified, research problems drive the research questions and ideally lead to understanding of the research findings.
In this second part of the Scholar Practitioner Project (SPP), you will develop an interpretation of a statistical analysis based on your selected data set, your prepared database, and SPSS calculations. You will make sense of this interpretation and communicate it to users by using tables and/or graphs.
Warner, R. M. (2021). Multiple logistic regression. In Applied statistics II: Multivariable and multivariate techniques (3rd ed., pp. 583–624). Sage Publications.
Banerjee, S., & Panas, R. (2017). Diabetes and cardiorenal syndrome: Understanding the “triple threat.”,
Hellenic Journal of Cardiology, 58(5), 342–347. https://doi.org/10.1016/j.hjc.2017.01.003
IBM. (2017). IBM SPSS complex samples 25,
ftp://public.dhe.ibm.com/software/analytics/spss/documentation/statistics/25.0/en/client/Manuals/IBM_SPSS_Complex_Samples.pdf
UCLA Statistical Consulting Group. (n.d.). FAQ: How do I interpret odds ratios in logistic regression?
https://stats.idre.ucla.edu/other/mult-pkg/faq/gen…
Walden University Academic Skills Center. (n.d.). Course-level statistics,https://academicguides.waldenu.edu/academic-skills…
Statistics Skills
Course Resources for RSCH8210
Course Resources for RSCH8260
Statistics Resources
SPSS/NVivo
SCHOLAR PRACTITIONER PROJECT – UNIVARIATE AND BIVARIATE ANALYSIS
Develop an interpretation for your statistical analysis based on your selected data set, your prepared database, and SPSS calculations. Clearly develop full sentences and paragraph(s) to describe each of the following:
Provide interpretation for descriptive statistical analyses based on your SPSS output. By interpreting the data, you are making sense of the analyzed data and communicating this to your reader. Do not simply repeat the output. Explain what it means. Perform the following tasks for each of your research questions separately:
Summarize the numerical results with descriptive analysis tables or graphs, including your interpretation. Follow APA rules for tables and graphs.
Summarize bivariate inferential analysis, as appropriate.
Summarize the numerical results with inferential analysis tables or graphs, including your interpretation.
From the Discussion for last week: Identify your research question, including explaining the null hypothesis.
Is there a difference between education levels and the probability that a person has diabetes in their life?
Null Hypothesis: There is no significant relationship between education level and if a person has diabetes in their life.
Interpret the Exp (B) coefficients for the model, specifically explaining the odds ratio and how you decided on the reference category for your independent variables.
The college graduates or above group is approximately 66% less likely to have diabetes compared to the base group, which is some college/associate degree. This is significant because the P-value (.004) is 0.05, this group is approximately 26% less likely to have diabetes compared to the base group, which is some college or AA degree.
The 9-11th grade (including 12th grade with no diploma, is not statistically significant because the P-value (.960) > 0.05 and cannot be used to reject, this group is approximately 2% more likely to have diabetes compared to the base group, which is some college or AA degree.
The High school graduate/GED or equivalent group is not statistically significant because the P-value (.181) > 0.05, this group is approximately 41% more likely to have diabetes compared to the base group, which is some college or AA degree.
I decided on the reference variable based on the largest group.
Run diagnostics for the regression model, explaining what was run. Does the model meet all of the assumptions? Be sure and comment on what assumptions were not met and the possible implications. Is there any possible remedy for the assumption violations?
Assumption 1: The outcome variable is binary.
Assumption 2: Based on the survey, we know that they are independent.
Assumption 3: None of the Statistics VIF values using the linear probability model are above ten.
Assumption 4: We assume that independent variables are linearly related to the log odds.
Assumption 5: Some of the groups have less than ten observations, which does not meet the assumptions. To remedy this specific assumption violation would be to merge groups in order to increase the sample size.
Create and explain an interaction term from two variables, as this will be one of your explanatory, independent variables.
The ‘Gender_xEducation_1’ identifies education levels for participants who identify as male. ‘Gender_xEducation_2-7’ identifies education levels for participants who identify as female.
After incorporating the gender variable, we find that there is a significant difference P= (.009) in having diabetes by gender. Females are 64% less likely to have diabetes compared to males.
The college graduate or above group is approximately 80% less likely to have diabetes compared to the base group, which is some college/associate degree. This is significant because the P-value (.004) is 0.05; this group is approximately 14% less likely to have diabetes than the base group, which is some college or AA degree.
The 9-11th grade (including 12th grade with no diploma, is not statistically significant because the P-value (.830) > 0.05; this group is approximately 9% more likely to have diabetes than the base group, which is some college or AA degree.
The High school graduate/GED or equivalent group is not statistically significant because of the P-value (.323) > 0.05. This group is approximately 35% more likely to have diabetes than the base group, which is some college or AA degree.
The interactions below are with gender differences in education groups.
The less than 9th-grade group for females is not statistically significant because the P-value (.439) is > 0.05. This group is approximately 59% less likely to have diabetes than the base group, males with less than 9th-grade education.
The 9-11th grade (including 12th grade with no diploma for females, is not statistically significant because of the P-value (.289) > 0.05. This group is approximately 70% less likely to inject drugs than the base group, males with 9-11th grade education.
The High school graduate/GED or equivalent group for females is not statistically significant because the P-value (.791) > 0.05 such that it cannot be used to reject the claim this group is approximately 14% less to I have diabetes compared to the base group, which is high school graduate/GED or equivalent group for males.
The college graduate or above group for females is approximately 200% more likely to have diabetes than the base group, which is the college graduate or above group for males. This is not statistically significant because the P-value (.156) is > .05.
Explain how the model and results can positively impact social change.
By raising awareness and promoting prevention, the above model, which depicts diabetes and takes gender and educational factors into account, might effectively influence societal change. We can improve people’s comprehension of the risk factors, symptoms, and essential lifestyle adjustments to avoid diabetes through encouraging diabetes education and empowering individuals. For people with diabetes to properly manage their illness, education is essential. People who are knowledgeable about managing their diabetes—which includes taking their prescriptions as prescribed, keeping an eye on their blood sugar levels, and adopting lifestyle changes—gain control over their health and are better able to make wise choices. These initiatives not only help to prevent diabetes but also enhance the care and treatment for people who have already been diagnosed. Better health outcomes, a lighter load on the healthcare system, and improved quality of life for people with diabetes and their communities follow as a consequence of this. Health Equity: emphasizing the importance of diabetes education and awareness can help to advance health equity. By concentrating on communities with limited access to healthcare services and information, such as economically underprivileged or low-income groups, educational initiatives can help close the knowledge and care gaps in diabetes.
References
Banerjee, S., & Panas, R. (2017). Diabetes and cardiorenal syndrome: Understanding the “triple threat.” Hellenic Journal of Cardiology, 58(5), 342–347. https://doi.org/10.1016/j.hjc.2017.01.003
IBM. (2017). IBM SPSS complex samples 25,ftp://public.dhe.ibm.com/software/analytics/spss/documentation/statistics/25.0/en/client/Manuals/IBM_SPSS_Complex_Samples.pdf
UCLA Statistical Consulting Group. (n.d.). FAQ: How do I interpret odds ratios in logistic regression? https://stats.idre.ucla.edu/other/mult-pkg/faq/gen…
Warner, R. M. (2021). Multiple logistic regression. In Applied statistics II: Multivariable and multivariate techniques (3rd ed., pp. 583–624). Sage Publications.
The post Advanced Analysis of Community Health Data Bivariate Analysis Worksheet first appeared on .
