write an essay aboout a statistics case

This assignment is designed to get you to read a real news article with statistical content and to be able to understand and critically analyze the details. I have provided a selection of brief articles on various topics that you can choose one of. Select only ONE to write about.

Your report should be between one to two pages of double-spaced text. Write it up like an essay, not just bullet points. The article should provide enough information about the study to assess it reasonably, but you may want to look up the original study if you need more details.

The first part of your paper should summarize the details of the study. Some of this you may have to guess based on common sense, but you may have to refer to the original study, if applicable. The following questions are intended to be a guide for what to write:

1. Study details: summarize information about how the study was conducted. For instance:

a. How was the data collected? Designed experiment, survey, or observational study?

b. In most cases, a sample is chosen to perform an experiment, be surveyed, or observed. If a sample was used, what kind was done? Simple random, clustered, stratified? Why was this particular type of sample done? How many people participated?

c. What were the observational units (e.g. is the data collected on individuals, families, etc.) and what is the population they were trying to describe?

d. If an experiment or observation, how many people were studied, for how long, etc.?

e. Other issues: who conducted the study and funded it, if applicable?

The second part should contain an analysis of any potential or obvious issues you may think are relevant. Again, not all of these will be relevant, but they are here to raise some questions to guide your analysis. Address only the ones you think are relevant.

2. Evaluate the study to see if there are any obvious or potential issues. For instance:

a. Do you expect a particular bias (e.g. nonresponse, self-selection, selection bias) to have occurred due to the sampling method? Why could this be?

b. Does the sample appear to you to be not representative (or is there a danger of it not being representative) for some reason? How might this affect the results?

c. The sample size was too small. This means in absolute size, not relative to the population size (samples are always small relative to the population). A good rule of thumb is that in general, a sample of 1,000 or more is good enough for most things, but if smaller it depends on the situation. If you only have a sample of 20 it is too small to draw any conclusions. Usually the smaller sample sizes are used if there is an experiment and it is expensive to do it for too many people. It also depends on the application. If it is a medical experiment, you probably need more people than if you were doing an experiment that is less important.

d. The study was paid for by a group that may be biased (for instance a politically liberal group running a political survey). Is this a concern?

e. The article was written by a person who may be biased about the study or featured on a news site or newspaper that is known to have a certain political or other bias—for instance Fox News, CNN. Note: I am not saying any article featured on a biased website is bad, just it has to be kept in mind.

f. The study tries to improperly make inferences about a population from the sample where the inference may not be logical (for instance using a sample of only children to say something about adults, or extrapolating overconfidently from a study about mice to infer something about humans). Remember, the population of interest should be similar to your sample.

g. The person writing the article is not knowledgeable enough on the topic

h. Questions asked in the study may be phrased in a certain way to get a certain kind of answer. For instance: asking “Most people think that legalizing marijuana will cause an increase of 5,000 in annual traffic accidents. Do you think it should be legalized?” is trying to get people to answer a certain way.

i. Problems in the survey. For instance, interviewing Spanish-speaking people without having an interviewer who speaks Spanish or having translations available of materials.

j. Does the article itself (usually written by a reporter) distort the study results with bold claims (such as the headline) that are not really true

3. Analyze any graphs or tables presented in the article

a. Is the data presented in a way that distorts it or may mislead the reader?

b. Are the claims of the article supported by this graph or table?