Exam 3 Flashcards
Describes the sample (can be known)
Statistic
Describes the population (usually can’t be known)
Parameter
Allows you to make predictions from data; you take data from samples and make generalizations about a population
Inferential Statistics
Simply summaries data in a meaningful way
Descriptive Statistics
Taking statistics from your sample data and using it to say something about a population parameter
Estimating parameters
Assumes no difference or association between the populations from which samples are drawn; to-be-tested hypothesis
Null Hypothesis (H0)
Is accepted (NOT PROVEN) if evidence suggests results obtained would be unlikely if the null hypothesis were true; assumes some difference or association
Alternative Hypothesis (H1)
Independent variable will have an effect on the dependent variable for the population….
Alternative Hypothesis (H1)
What are the three components of writing a hypothesis?
- Specificity
- Clarity
- Testability
Assigning values (numbers) to objects or events based on agreed-upon rules
Measurement
Obtain values from a device (ruler, visipitch, spectrograph)
Instrumental
Trained individual judges quality or occurrence of events (e.g. severity ratings, stuttering frequency)
Observational
T/F: A behavioral test instrument may yield observational data.
TRUE
What is the order of level of measurements from weakest to strongest?
Nominal, ordinal, ration, and interval
Categorical - items are assigned to different, mutually exclusive categories. Ex. religion, gender, ethnicity
Nominal Scale
Ordinal, interval, and ration are all termed….
Continuous (things can be ordered)
You can’t tell if the difference between times is equal from one rank to the next (ex. severity index, size, stimulus complexity, class rank)
Ordinal Data
Constant distance between equal intervals on the scale is maintained (2-1 = 3-2)
Interval Scale
Can identify equivalent ratio values across the scale; One thing IS twice as much as something else; a true zero point exists
Ratio Scale
Indicates how likely it is that a result occurred by chance alone; measures the strength of the evidence against the null hypothesis
Probability value
T/F: Smaller p-value indicates stronger evidence against the null hypothesis.
TRUE
Nondirectional; null hypothesis of no difference (or association); alternative hypothesis - a difference exists
Two-tailed test
Directional; alternative hypothesis - goes in only one direction; null hypothesis, no difference
One-tailed test
T/F: One tailed test tends to ignore evidence in the wrong direction
TRUE
T/F: Two-tailed test is more conservative (safer)
TRUE
Assumes that the population measured is normally distributed (bell-shaped, not skewed, not too peaky or flat)
Parametric Tests
Informally describes the degree to which the distribution is symmetrical
Skewness
Informally, describes the “peakedness” and the shape of the tails of the distribution
Kurtosis
Direction of the relationship
Sign (positive or negative)
Degree of predictability
Strength
y tend to increase when x increases is…..
Positive association
y tends to decrease when x increases
Negative Association
Score on x tells you nothing about what the corresponding score on y is likely to be (no positive or negative trend is visible in the plot of the paired points)
No association
What tell you how much information knowing one variable gives you about the other - whether positive or negative
Strength of Association
When the correlation between two variables is caused by another (third) variable, with which both are correlated
Third Variable Effect
Want to know whether there is a relationship between two things - both continuous
Associations
Want to know whether differences in performance can be found - across groups or tasks
Difference
When are independent sample t-tests used?
When comparing the mean of one population to the mean of another population
When are paired sample t-tests used?
When interested in the difference between two variables for the sample population
What we really want to know
Population Parameters
What we can know
Statistics of Samples
P-value is less than or equal to 0.05 means…
Statistically significant
P-value is greater than 0.05 means…
Not significant
Two groups differ without specifying direction
two-tailed hypothesis
the mean in a sample group will be less than the mean in a comparison group
one-tailed hypothesis
Does positive skew lean towards the left or to the right?
Right
Does negative skew lean towards the left or the right?
Left
Non-parametric version of correlation; does not assume normal distribution; requires ordinal data
Spearman’s Rho
Parametric statistic; requires interval or ratio data; assumes a normal distribution and equal variance for each variable; range -1 to 1
Pearson’s r
What type of question is….
Are scores on tests of phonological awareness predictive of later reading ability?
Association
What type of question is…..
Do subjects perform better on a task following training?
Difference
one independent variable with more than two levels
One-way ANOVA
used to determine which pairs of means are significantly different from each other
Post-hoc Tests
Allows the researcher to manipulate two factors at a time; two independent variables
Two-way ANOVA
more than two independent variables
Three-way ANOVA
