Posted: April 11th, 2015

statistic

For the final project, you will submit a data report. In short, you will tell a story about your analysis. Just stick to a theme throughout; you can, have the same outcome variable and just compare different groupings/predictors for the outcome, such as predictors for an STD. Or have one grouping/predictor, and analyze different outcomes, such as indicators of heart disease.

Below are the specific areas that your report should include; there are many questions in each heading you can answer, but you do not need to nor should you try to answer all the questions under each heading. These headings can be duplicated in your report.

Background

Where did the data come from? Who was surveyed? How were they surveyed? Interviews? Medical records? When were they surveyed? What types of variables were measured? What topic are you focusing on? Why? What is the rationale behind your analyses? What is important about it? Is there epidemiological data that underscores the importance of your analysis?

Descriptive Statistics

Perform some basic descriptive statistics for the data set. How many total participants? Gender/age/education/income/race/ethnicity breakdown? See the example table; you can just put a table in for this section.

Group Comparisons

Compare means in this section; this can be achieved with a t-test or an ANOVA. Three meaningful comparisons should suffice in telling your story.

Associations

Which variables are significantly associated? Which variables are not significantly associated? This can be achieved with a correlation. Three meaningful associations should suffice.

Higher-Order Analyses

This section should be the “punchline” of your story. What variables are predicted by other variables after controlling for your covariates? What are your predictors? What are your outcomes? What are your covariates? This section can be achieved using linear or logistic regression. Three analyses are needed at minimum.

Graphs

Four graphs are needed at minimum to depict your analyses; bar graphs and scatterplots should be appropriately used to visually represent your analyses.

Interpretation/Conclusions

What does your analysis tell you? What can you conclude? How does this relate to previous literature? What explanations do you have for your findings? How else would you have done the study differently? What would you change? Were any of the results surprising to you? Do you think there may have been a methodological limitation to the data collection process?