04_Ch03_fall08_pam2100 - PAM 2100 Intro Stats A. Sinan nr...

Info iconThis preview shows pages 1–7. Sign up to view the full content.

View Full Document Right Arrow Icon
PAM 2100 Intro Stats A. Sinan Ünür Producing Data Producing Data Chapter 3
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Chapter 3 PAM 2100 – A. Sinan Ünür Data and Statistical Inference The source of data dictates to a very large extent what conclusions you can reach. Anectodal evidence : Fancy phrase for “not data”. Based on a few haphazard observations, selected most often based on the observer’s bias. Data come from: Others: Government agencies, NGOs, individual scientists, newspapers all collect data through various methods and some, if not most, make the data available. You!
Background image of page 2
Chapter 3 PAM 2100 – A. Sinan Ünür Exploratory Studies Exploratory study: Fancy name for what frequently degenerates in to mining data for patterns and forming theories based on those patterns rather than having a model of the data generating process. Scientific progress depends on an honest effort to find data which contradict our hypotheses. Falsifiability is the defining characteristic of a scientific hypothesis (http://en.wikipedia.org/wiki/Falsifiability). It is impossible to go wrong if you test whether the pattern you already observed in the data is exhibited by the data . Exploratory studies are useful only if patterns you observe are used as a basis to design proper experiments to further study what causes those patterns.
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Chapter 3 PAM 2100 – A. Sinan Ünür Observational Studies Observational studies: No control. E.g. observe study habits of depressed and not depressed individuals but exercise no control over who is assigned to which group. It is impossible to make a statement about causality using observational data: Think obesity and suburban sprawl. The choice of both is determined outside of the researchers control and likely influenced by other variables as well.
Background image of page 4
Chapter 3 PAM 2100 – A. Sinan Ünür Sampling versus Census A census is an attempt to count all observational units within the study’s population. Sampling: Systematic selection of what is a priori believed to be a representative subset of the population. Observational studies most often use sampling: 1. A census is most often not feasible: How do you determine the Time Between Failures for all the light bulbs you produce? 2. A systematic sampling scheme does a better job of avoiding bias. Why?
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
PAM 2100 – A. Sinan Ünür Experiments Deliberately impose treatment(s) to observe responses under different conditions. Experimental units / subjects / participants Treatments / factors / factor levels Does noise affect concentration? Possible design: Assign students to groups. Expose students to low noise, moderate noise and high noise and, most importantly, no noise . Why? Look at how well each group does on some concentration test. If
Background image of page 6
Image of page 7
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 35

04_Ch03_fall08_pam2100 - PAM 2100 Intro Stats A. Sinan nr...

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online