Test 2 Flashcards
Double-Blind Experiments
Neither the patients nor the supervisors know who is receiving a placebo
Block Design
Organize subjects into groups that match the different strata of the population and assign them randomly to different treatments in each block
Lurking Variables (in linking religiosity to longevity)
gender, education, physical mobility, ethnicity, age
Randomized Comparative Experiment
Assign subjects RANDOMLY to a TREATMENT GROUP and a CONTROL GROUP. Treat the subjects in the treatment group but not those in the CONTROL GROUP (or treat the latter with a PLACEBO). Compare the effects and decide whether there is a difference between the two group that is STATISTICALLY SIGNIFICANT.
Measuring
MEASUREMENTS assign numerical values to VARIABLES that are descriptive of certain PROPERTIES of (statistical) individuals. Measurements are often performed by means of INSTRUMENTS that employ specific UNITS. A variable is a VALID measure of a property in it is relevant or appropriate as a representation of that property.
Rates; Counts
___ are frequently more valid or meaningful measures than simple ___.
Predictive Validity
The most useful form of validity
Examples of measurements
length, time, college readiness, highway safety, unemployment, intelligence
Unemployment rate
=number of people unemployed
________________________________
number of people in the labor force
Predictive Validity
Predicting success in college by using SAT scores.
Predicting the weather from current conditions.
Predicting statistical particle behavior by means of quantum descriptions.
Measurement Errors
Measured value = true value + bias + random error
Errors; bias; randomness
___ in measurement can be due to ___ or ___.
Reliable
If the random error is small, the measurement is said to be ___.
Reliability
can be improved by taking averages over several repeated measurements
Bias
can be reduced by using better instruments (atomic clocks instead of mechanical or quartz-based clocks).
Categorical variable
divides a given population into different groups or categories (distribution is commonly displayed by a bar graph or pie chart)
Quantitative variable
variable that takes numerical values (the change over time may be displayed by a line graph)
Distribution; symmetric
a ___ is ___ if the parts to the right and left of the center approximately mirror each other
Distribution skewed to the right
the parts to the rights of the center extend farther than the parts to the left
Distribution skewed to the left
the parts to the left of the center extend farther than the parts to the right
Describing Distributions
the median, quartiles, maximum/minimum, mean/average, variance, and standard deviation
Overall pattern; deviations
in any graph of data, look for an ___ and also for striking ___ from that pattern
Outlier
An ___ in any graph of data is an individual observation that falls outside the overall pattern
mean
= sum of values
____________________
number of values
Sample variance
the sum of a set of numbers divided by n-1 (n=amount of numbers in the set)
Sample standard deviation
finding the sample standard deviation s by taking the square root of the variance