Introductory Statistics For The Behavioral Sciences Presentation - Chapters 1 & 2
Introductory Statistics For The Behavioral Sciences Presentation - Chapters 1 & 2
Summary numbers, or indices, that result from an analysis of data (numbers) pg. 2 All the procedures and tools used to organize and interpret facts, events, and observations that can be expressed numerically pg. 2
Imaginary World
Snap
Crackle
Imaginary World #2
if
Snap
and or
then
Crackle
Pop
Imaginary World #3
Variables
n. an element, feature, or factor that is liable to vary or change adj. not consistent or having a fixed pattern; liable to change
Q: Why is little Sally afraid of dogs? A: What are the variables, and how do they vary together?
(the presence or absence of dogs, Sallys emotional state, Sallys neural structure as it relates to these other events, Sallys history and how it relates to the above elements, etc.) Q: Why does my roommate drink my milk and leave the empty jug in the refrigerator?
IV
DV
Example 1
Octavian is interested in the effect of breathing exercises on time spent on task.
What is the independent variable?
Breathing exercises Time on task
Example 2
Adelpha wants to know if pronunciation can be improved if one practices speaking with marbles in ones mouth.
What is the independent variable?
Clarity of pronunciation Practice with marbles
Design
1. Random Selection
Random = each member of the population has an equal chance of being selected. If you stop here, you cant make causal statements.
Random = each selected participant has an equal chance of being exposed to each condition. This is the only way you can make causal statements. This is an experiment.
2. Random Assignment
Example
Romulus selects the first 50 students that walk into the Bernhard Center to participate in a study on WMU students. The students are asked if they are smokers or nonsmokers and then they fill out a short survey that is supposed to measure their general level of anxiety. Do we have random selection , random assignment , neither , or both ?
So what?
1. No random selection: means that you cannot say anything about the target population (WMU students). 2. No random assignment: means that you cannot make causal statements about the IV on the DV (e.g., you cant say that smoking caused students to have a higher level of anxiety). 3. This experiment is a big failure, no matter how good the statistical analysis is.
Complexity
Variability is brought about by COMPLEXITY, not randomness. Human behavior is very complex (i.e., there are many relevant variables), but because it is orderly, we can predict and influence behavior. Statistics makes this prediction and influence much easier.
Sample
Part of the population
Biased
Summary Measure = Parameter Ex: = pop. mean 2 = pop. variance = pop. standard deviation
Summary Measure = Statistic _ Ex: X = sample mean s2 = sample variance s = sample standard deviation
Notice: Greek symbols are used for parameters, Roman symbols for statistics
Why Sample?
Population
Inference
Sampling
Sample
Inference
Definition: a conclusion reached on the basis of evidence [in our case, sample data] and reasoning [in our case, statistical analysis]
INFER
Sampling Error
Sampling error: the error caused by observing a sample instead of the population.
Represented by the formula:
statistic parameter = sampling error Ex:
The ONLY way to avoid sampling error entirely is to take data on the entire population What can we do to reduce sample error?
1. Random Sampling 2. Large Samples 3. Multiple Samples
Try This
Population: People age 16-65 living in the U. A. E. U. A. E. Male: 3,500,000 What should your Stratified sample look like if your total sample size is 100?
Female: 2,000,000
Heres how I wrote the number of females as I solved it: a fraction of the total population, and put my unknown over my sample size on the other side. Then I solved for X.
Sampling without replacement means that after each person/object is selected, it is not returned to the population and, therefore, cannot be selected again.
In most behavioral studies, this is the only way to sample. Why is that?
Measurement
Definition: the assignment of numbers or labels to objects or events Labels? How is that measurement?
1. Nominal Scales
Naming of objects or events Naming measurement
Naming allows us to categorize and deal with individuals or groups separately Labeling is required to get frequency data, which is type of measurement
Nominal Scales
If we took the last seven digits of each of your phone numbers and averaged them, what could we learn? NOTHING. Be careful, just because its a number doesnt mean its not a name. What if we wrote down the area code of each student in class and found the number of times each area code occurs (frequency)? This would work! Lets try it.
2. Ordinal Scales
Naming objects or events and putting them in order Ranking or rank ordering Example: advisory system
Ordinal Scales
4 3 2 1
Interval Scales
Naming of objects or events and putting them in order using units with equal intervals The distance between 1 and 2 should be exactly the same as the distance between 2 and 3 or 3 and 4.
Interval Scales
In psychology: whether or not Likert scales are interval or ordinal scales is an issue of debate. Why do you think this might be?
Note: We will treat them as interval scales in this course, but only for convenience.
Ratio Scales
Naming of objects or events and putting them in order using units with equal intervals on a scale with a true zero Ask yourself, does zero really mean nothing?
Homework
Prepare for Quiz #2 over Chapters 1 & 2. Read Chapter 3 and be prepared to answer basic questions. Complete Homework assignment #1 and prepare to hand it in at the beginning of the next class.
Check WebCT and/or the course website for the homework assignment and other study materials.