2 small assignments - due tomorrow

*Simply write out essay response on this sheet*

Should be at least 2 paragraphs each

1. Big Data Analytics

Look at the charts in the IBM presentation on Big Data and Analytics (in Resources for Session 7), and links for Analytics in the StatsPathfinder.

What do these sources tell you about Big Data in terms of the scale of the data inputs, the types of data being used, and how data is cleaned and stored?

What have you learned about the timing of the analysis, and the purposes of the data analytics.   How is this different from what we have been studying in statistics and the hypothesis testing method?    For example, have we been using geo-spatial data, streaming data, or video & image data?   Have we been using samples or populations?   How large have our data sets been?   Have we been solving problems after the fact or as they occur?   Are we describing, prescribing, or doing cognitive work?

Can you name an application of Big Data and Analytics in the public sector or non-profit world?

What are some of the main strengths and weaknesses of this type of analysis -- especially vis-a-vis the traditional research methods?

Response:

2. Can We Inform Policy with Regression?

Can multiple regression be used as a predicting (not a forecasting) technique? Are there principles to be wary of when predicting (e.g., the influence of outliers, predicting outside of the range of given values, or extrapolating)? Explain.

Can you cite an example where regression is used to predict? Find a study on a topic of interest of yours, and cite a regression coefficient, it's t-score or p-value, and how you would interpret a one unit change in that variable. How much would the variable need to change in order to make a big impact on the dependent variable? [You may use an example from last week, and take your interpretation a little further.]

If you have trouble finding an example, look at one of the following:

1) Handcuffing the Cops: Miranda's Harmful Effects on Law Enforcement, http://www.ncpa.org/pub/st218?pg=6. Also look at the regression coefficients in the Appendix on pg. 8. What is the impact of the Miranda laws? What variables are being controlled for? How many more crimes would be cleared without the laws?

2) Homeless in America, Homeless in California, http://urbanpolicy.berkeley.edu/pdf/QRS_REStat01PB.pdf . Look at regression coefficients in Table 3. How can homelessness be decreased by 25%? According to this paper, is homelessness the consequence of individual problems (e.g., mental illness) or structural problems (e.g., lack of housing)?

Response:

3. Lying with Statistics -- What Crosses the Line?

In the lesson notes for this week, there is discussion of what constitutes ethical rules for human subject studies.  What are some of these principles (highlight one) and why we need this?  In "Damned Lies and Statistics", the author states: "The lesson that people count -- that we don't just find statistics but that we create them -- offers both a warning and a promise.   The warning is that we must be wary, that unless we approach statistics with a critical attitude, we run the risk of badly misunderstanding the world around us.  But there is also a promise: that we need not be at the mercy of numbers, that we can learn to think critically about them, and that we can come to appreciate both their strengths and their flaws."   Do you agree?   Do you have examples from this class to illustrate these points?

Response: