Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
67 Cards in this Set
- Front
- Back
|
What are four types of Data Scales or Levels of Measurement?
|
1) Nominal (least desirable)
2) Ordinal 3) Interval 4) Ratio (best scale) |
|
|
Which type of data scale is this?: Data that involve assigning a name or category to each item. no category is considered better or worse than another category. Categories are mutually exclusive.
|
Nominal Data
|
|
|
What are some examples of the Nominal data scale?
|
positive/negative
yes/no normal/disordered income levels, age groups, other demographic data |
|
|
Which data scale is this?:
Data are ordered or ranked according to some degree of magnitude of difference. It is assumed that the distances between the rankings are NOT equal. |
Ordinal
|
|
|
What are some examples of the Ordinal data scale?
|
Severity ranking
ratings or rankings by subject or others Voice quality ratings levels of linguistic complexity. |
|
|
Which data scale is this?:
Has equal distance between data points. |
Interval
|
|
|
What are some examples of interval data scales?
|
personality, intelligence, or achievement tests
language tests hearing thresholds from audiometric testing |
|
|
What data scale is this?:
Zero point is used as a point of reference. A zero score indeed means that the phenomenon or behavior was absent. |
Ratio scale
|
|
|
What are some examples of ratio data scales?
|
frequency counts
time and distance measures amplitude measures on ABR based voltage latency of response |
|
|
The type of data you have will dictate the type of ___________ ____ you will be able to use to analyze your data.
|
Statistical test
|
|
|
What is a parametric test?
|
Interval and ratio data
Tests: t-tests, ANOVA, Pearson r |
|
|
What is a nonparametric test?
|
Nominal and ordinal data
Tests: Chi square, Wilcoxon rank test, Mann-Whitney U test |
|
|
A) Scores are normally distributed
B) sample size is reasonably large (30) C) Interval or ratio data are used D) There is homogeniety of variance among groups of data E) Subjects have been randomly sampled These are all assumptions of what? |
Parametric tests
|
|
|
Nonparametric tests have sample sizes that are usually _______.
|
Small (6-10 subjects)
|
|
|
Use of parametric tests and non-parametric tests allow the researcher to make inferences about the _____________ from the sample data.
|
Population
|
|
|
What are the most common ways to express data?
|
Standard deviation (most impt) and means.
|
|
|
Sample size is important in that large samples (> 30) usually means that the sample will be normally distributed based on the __________ _____ _______.
|
Central Limit Theorem.
|
|
|
What does the central limit theorem state?
|
That regardless of the shape of the population, the mean of all the samples selected from the population will be normally distributed.
|
|
|
What are two types of error within Inferential Statistics?
|
Systematic error and Sampling error.
|
|
|
What is systematic error?
|
the error that occurs because the measurement device used to collect the data is not calibrated or is unreliable. The error is consistently aplied to all the measurements of the dependent variable; calibration of audiometric equip, physiological equipment, surveys, rating scales, tests, etc.
|
|
|
What is Sampling (or random) error?
|
Error due to the type of people we test. Biased sample of subjects. Researcher may not know that the subject is unique so that's why it is called random error.
|
|
|
Do we have more control over systematic errors or sampling errors?
|
Systematic.
|
|
|
What is a null hypothesis mean?
|
That there is no difference between the groups.
|
|
|
What are two types of alternative hypotheses?
|
Direction and non-directional
|
|
|
What is a directional hypothesis?
|
Greater, stronger, lesser, smaller, weaker, difference/relationship for group A than for the other groups.
|
|
|
What is a non-directional hypothesis?
|
There is a difference between the two groups. States a difference, but isn't specific.
|
|
|
How is the null hypothesis tested?
|
Statistically.
|
|
|
If the calculated value is greater than or equal to the critical value, we ______ the null hypothesis.
|
Reject.
|
|
|
Describe a Type I Error.
|
Risk that you can make a wrong decision in rejecting the null hypothesis when it is true (you say that there are differences between gruops but there really aren't).
|
|
|
Describe a Type II Error.
|
When you accept a false null hypothesis- that is, you say there are no differences when there really are differences between the groups.
|
|
|
What dictates how much risk is allowed for significance?
|
The alpha level
|
|
|
What is a one-tailed test?
|
All of the alpha level is placed in at one end or the other
|
|
|
What is a two-tailed test?
|
the alpha level is split equally between the right and left tails of the distribution (ex- .05 means that .025 is placed at each tail)
|
|
|
Is a one or two tailed test more common and thus, better to use?
|
Two-tailed.
|
|
|
In terms of a one/two-tailed test, what is the critical region?
|
Area associated with the alpha level. Values falling in this region are considered significant because the differences are not due to sampling error alone.
|
|
|
In terms of a one/two-tailed test, what is the critical value?
|
The value associated with the alpha level chosen for a particular statistical test. Each statistical test has its own set of critical values. See t-test of critical values.
|
|
|
In terms of a one/two-tailed test, what is the calculated (obtained) value?
|
The score or value that is obtained from the test statistic that is derived from your data. Value is the one number that is generated at the end of the caluculations of the data using a particulat test. That value either falls in or out of the critical region and exceeds or does not exceed the critical value.
|
|
|
The larger the sample, the ______ the critical value will be to reach significance.
|
smaller
|
|
|
Significance is determined by what?
|
Degrees of freedom.
|
|
|
When would you use an independent t-test?
|
When you want to compare the differences between two groups of subjects that are independent of one another (males/females, children/adults).
|
|
|
What is a restriction for independent t-tests?
|
Need about 10-15 subjects per group. Less than that would require the use of the Mann-Whitney U test.
|
|
|
When would you use a dependent t-test?
|
When you want to compare the difference between two groups that are related (matched) along some variable such as age and/or sex. You could also use this test when comparing the difference between two measures within one independent variable (pre- and post-test scores during treatment for one group of subjects).
|
|
|
What do you use in order to check for the differences between two groups?
|
1) t-test for independent samples
2) t-test for related samples |
|
|
What do you use in order to check for the differences between three or more groups?
|
One or Two-way ANOVA
|
|
|
When might you use a one-way ANOVA?
|
With THREE or more measures of one independent variable for one group of subjects (a group of normal adults are asked to complete a task under 3 different conditions.
|
|
|
When might you use a two-way ANOVA?
|
When you have at least two levels of two independent variables (ex- two groups and 2 conditions) OR (3 types of teaching methods under 3 types of conditions).
|
|
|
The dependent t-test df are exactly ____ of the independent t-test df.
|
Half.
|
|
|
How would you calculate the df for Independent samples?
|
(n-1) + (n-1) = df
|
|
|
How would you calculate the df for dependent samples?
|
n-1
|
|
|
What are the six steps to determine the differences between two groups?
|
1) Independent or Dependent groups?
2) What is your alpha level? 3) Are you going to use a one-tail or two-tailed test? 4) Calculate the t value. 5) Determine degrees of freedom. 6) Compare obtained t value from data to critical values of t-test. |
|
|
How many F-scores does a one-way ANOVA have?
|
One.
|
|
|
How many F-scores does a two-way ANOVA have?
|
Three.
|
|
|
A x B = F-score
What is this an example of? |
Interaction effect.
|
|
|
T or F?: t-test for independent and dependent samples are examples of nonparametric tests?
|
False. They are parametric.
|
|
|
When referring to an ANOVA test (Analysis of Variance), what does factor mean?
|
Independent variable
|
|
|
When referring to an ANOVA test (Analysis of Variance), what do levels mean?
|
subcategories of the independent variable.
|
|
|
T or F?: In analysis of variance, if there are only two levels of one factor, then you don't have to do post hoc analysis.
|
True.
|
|
|
ANOVA tests only tell us that the groups are different. What would we use to figure out which levels of the factor contributed to the differences?
|
Post-hoc analysis.
|
|
|
T or F: you can have dependent t-tests even if the samples aren't equal.
|
False. Samples MUST be equal to do a dependent t-test. Not for independent t-tests.
|
|
|
Are one-tailed tests directional or nondirectional?
|
Directional.
|
|
|
Are two-tailed t-tests directional or non-directional?
|
Nondirectional
|
|
|
Two-tailed tests make it _______ to find differences.
|
more difficult
|
|
|
What is a reason a researcher might have committed a type I error?
|
Set the alpha level too high.
|
|
|
What is a reason a researcher might have committed a type II error?
|
Sample size was weird.
|
|
|
How do you tell if there is significance for a t-test or ANOVA?
|
Significant difference if the calculated value is equal or greater than critical value.
|
|
|
Standard deviations are more important than group means because the show how _________ the group is.
|
Variable
|
|
|
T or F: Post-hoc analysis should always be don for one-way ANOVA's?
|
True.
|
