Top Banner
Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012
13
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

Choosing the Appropriate StatisticsDr. Erin DeversOctober 17, 2012

Page 2: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

Quantitative or Qualitative? Is the data quantitative or qualitative? If qualitative, you have to do the hard

work of coding the material numerically before analyzing it.

Page 3: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

Level of Measurement? Nominal

Descriptive stats: mode Inferential stats: chi-square

Ordinal Descriptive stats: median

Interval/Ratio Descriptive stats: mean and standard

deviation Inferential stats: it depends

Page 4: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

Chi-Square (ϰ2) Indicates whether a distribution of

frequencies is what you would expect If distribution is significantly different

from what is expected, then it suggests there is a true difference present

Examples

Page 5: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

Example: Are more males or females diagnosed with schizophrenia?

Male Female

50 50

Guys Girls

70 30

* p < .001

Page 6: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

Interval/Ratio Data: Did you conduct an experiment? Correlation and Regression t-test, ANOVA

Page 7: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

Correlation Joke

Page 8: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.
Page 9: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

Correlation Coefficient:

Definition: number that represents the relationship between two variables.

Answers the question, how are changes in X related to changes in Y?

Gives you information regarding

Strength of the relationship

Direction of the relationship

Page 10: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

Linear Regression

Basic premise: given previous observations, one can (within some range of error) predict what is likely to occur in futureExamples: GRE scores and performance in grad school Parental smoking and smoking in teens Cholesterol level and risk for heart attack. Maternal depression and childhood

psychopathology Workers level of responsibility and job

satisfaction Beliefs and what one will do within a certain

situation.

Page 11: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

t-tests Provide a way to compare two group

means with each other Paired or Independent (depends on

whether you used a between or within subjects design)

Page 12: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

ANOVA: Analysis of Variance F statistic Takes into account the fact that chances

of getting a difference randomly goes up with multiple comparisons.

F = t2

Comparing means of multiple groups

e.g. What dose of a drug works best for arresting hallucinations in schizophrenia?

Page 13: Choosing the Appropriate Statistics Dr. Erin Devers October 17, 2012.

Questions?