5018 Unit 8 Flashcards
Top-down approach, large to small
Theory-hypothesis-test hypothesis-specific answer
Deductive Research Paradigm
Bottom-up, small to large
data-analysis-generalize
Inductive Research Paradigm
Required in deductive approach to interpret data
Statistics; quantitative data
Required in inductive research paradigm
Qualitative approach
Four types of data
Nominal
Ordinal
Interval
Ratio
Type of data that refers to categories
Nominal
Type of data that refers to order
Ordinal
Type of data where difference between each value is even
Interval
Type of data where difference between each value is even and has a true zero
Ratio
Three measures of central tendency
Mean
Median
Mode
Sum of scores divided by number of scores; most preferred measure of central tendency
Mean
Score that divides distribution exactly in half; gives two groups of equal sizes
Median
Score that has the greatest frequency
Mode
Two types of Mode
Bimodal
Multimodal
Two modes or peaks
Bimodal
More than two modes
Multimodal
Used for nominal scales, discrete variables, or describing shape
Mode
Used for extreme scores, skewed distribution, undetermined values, and open-ended distributions
Median
Three measures of variability
Range
Interquartile range
Standard Deviation
Describes the distribution in terms of distance from the mean or between two scores; how spread out or clustered together scores are in a distribution
Variability
Distance between targets score and smallest score + 1
Range
Criticisms of Range
Crude and unreliable measure of variability
Does not consider all scores in the distribution
Most important measure of variability that measure typical distance from mean and uses all scores in the distribution
Standard deviation
A tool in inferential statistics that measure the likelihood of an event
Probability
Two types of probability
Subjective
Objective
How to express probability
Always positive
Can be in the form of fractions, decimals or percentages
Each individual in the population has an equal chance of being selected; there must be constant probability for each and every selection
Random sampling
The most common occurring shape for population distribution
Normal shaped distributions
Provide incomplete pictures of the population
Samples
The discrepancy, or amount of error between a sample statistic and its corresponding population parameter
Sampling error
Distribution of statistics obtained by selecting all possible samples of a specific size of population
Sampling distribution
For any population, the distribution of sample means will approach a normal distribution as n approaches infinity
Central Limit Theorem
The shape of distribution of sample means will be almost perfectly normal if one of the following conditions is satisfied
- Population from which sample is selected is normal
2. The number of scores (n) in each sample is relatively larger (n>30)
The larger the sample size, the more probable that the sample mean will be close to the population mean
The Law of Large Numbers
Statistical method that uses sample data (statistics) to evaluate a hypothesis (question) about a population parameter
Hypothesis Testing
Basic common inferential procedure of hypothesis testing
z scores, probability, and the distribution of sample means
Purpose of hypothesis testing
Help researchers differentiate between real patterns I data and random patterns in data
Hypothesis testing begins with…
known parameters
The goal of hypothesis testing
determine what happens to the population after the tx is administered
Assumptions for hypothesis tests with z-score
Random sampling
Independent observations
Value of SD is unchanged by the tx
Normal sampling distribution
4 Main Steps of Hypothesis Testing
State hypothesis
Set criteria
Collect data
Make decision
Predicts that IV (tx) will have no effect on the DV
Null hypothesis
Predicts that IV(tx) will have an effect on the DV
Alternative hypothesis
The probability value that is used to define the very unlikely sample outcomes if the null hypothesis is true
Alpha level (level of significance)
Extreme sample values that are very unlikely to be obtained if the null hypothesis is true
Critical region
Purpose of statistic
Determine whether the result of research study (the obtained difference) is more than what would be expected by chance alone
Types of Hypothesis Testing Errors
Type I Error
Type II Error
Reject null hypothesis when it is actually true
False reports in scientific literature
Type I Error
Failing to reject the null hypothesis when it is actually false
Type II Error
A test used to compare two means
Alternative to Z scores
T-test
Occurrence of first event has no effect on the probability of the second event
Independent observations
Advantages of related-samples design (AKA within-subject designs)
- Eliminate the problem of individual differences between subjects
- Greatly reduces sample variance
2 Types of Contaminating Factors
Carryover effects
progressive error
Subject’s response in 2nd tx is altered by lingering aftereffects from the 1st tx
Carryover effects
Subject’s performance changes consistently over time
Progressive error
2 ways to deal with contaminating factors
- Counterbalance the order of tx presentation
2. Use different experimental design f contamination is expected
Assumptions of Related-Samples t-Test
- Observation within each tx condition must be independent
2. Population distribution of difference scores (D values) must be normal
Analysis of Variance (ANOVA)
Tell whether or not there is a significant difference between 3 or more groups
Follow up that would tell you where the difference is located
Multiple Comparison Procedure (MCP)
Statistical technique used to measure and describe relationship between two variables
Correlation
What does correlation measure?
Direction
Form
Degree
Two types of direction
Positive Correlation
Negative Correlation
Positive Correlation
X and Y change together moving in the same direction
Negative Correlation
X and Y change inversely
Describes linear relationship between 2 or more variables
Regression
Ways to distort correlation
Restricted range
Outliers
Measure of the strength of a phenomenon
Effect size
Benefits of Effect Size
- Relatively easy to calculate and interpret for group data
- Can be used to summarize data from many studies with different DV
- Not dependent on sample size
2 Types of Statistics
Descriptive Statistics
Inferential Statistics
Goal of descriptive statistics
Describe properties of the samples you are working with
Measures used in descriptive statistics
Central tendency
Variability
Effect size
Reasons for using descriptive statistics
Complement visual analysis
We already use them
Program evaluation
My open doors for funding
Reason for not using descriptive statistics
May hide trends
Goal of inferential statistics
Use sample data as the basis for answering questions about the population
Reasons for using inferential statistics in ABA
Appropriate for certain types of research
May open doors for funding
Perceived weakness of reliance on visual analysis in ABA
Reasons for not using inferential statistics in ABA
- Don’t tell how likely results are replicated
- Don’t tell the probability of results were due to chance
- The probability is conditional
- Best way to increase chances of significance is to increase n of participants
- Large number of variables that will have very small effects become important
- Limits the reason for doing experiments
- Reduce scientific responsibility
- Emphasize population parameters at the expense of behavior
- Bx is something an individual does, not what a group average does.
- We should be attending to social significance
- Durability of changes
- Number and characteristics of participants that improve in a socially significant manner