02-01-2012

Why would we use a sample rather than a population? To save time and money

Blindness - Makes for a better study by eliminating the placebo effect.
Placebo - control group receives a fake treatment (sugar pill). A reported 33% believe it worked when they consumed a placebo

Double blind study - When the researcher(s) and the subjects are unaware of who are receiving treatment
Single blind study - When the researcher(s) are aware of who is receiving treatment but the subjects are unaware. (May unintentionally impose bias in the study.)

Similarities between Experimental and Observational Studies:
Predictor Variable (explanatory variable) - Explains the outcome (e.g. medicine makes you feel better, height explains weight)
Response Variable - Changes with respect to explanatory variable (weight in previous example)

Ex: Pill for cold symptoms. The explanatory variable is the pill because it explains the relief from the cold, whereas the diminishing cold symptoms are a response variable of the pill.

Another way to think of this Predictor/Response Variable paradigm is: “This (predictor) causes that (response)”
Using this logic - Do pills cure a cold or do colds cure pills? Does height determine weight or does weight determine height? (if this logic confuses you please see Brad.)

Randomization and control make the experiment

Frequency - A count. "how many times it happens" (e.g. There are 14 men in a class of 42, there are 28 females in a class of 42)

Relative frequency - Converting count to a percentile.(e.g 33% of the class is male, 66% of the class is female)

Distribution - Pattern of variability

Bar graphs are only for categorical variables!


Minitab instructions for frequency/relative frequency
Stat > Tables > Tally Individual Variables

Selecting these options will bring up this menu:

Here you can type in or double click on the variables you are interested in. Additionally you can select "Counts" to report the frequency or "Percents" to report the relative frequency.


Minitab instructions for frequency bar charts
Graph > Barchart > 

Select "Counts of Unique Values" from the drop down menu and "Simple" graph, then select "OK"

Select the variable you are interested in graphing, then select "OK"

Minitab instructions for relative frequency bar charts
Graph > Barchart > 

Select "Counts of Unique Values" from the drop down menu and "Simple" graph, then select "OK"

Select the variable you are interested in graphing, then select "Chart Options"

Select "Show Y as Percent", then select "OK"

Select "OK"