Correlation

1
Correlation
2
Review from statistics

• What is a correlation?
• Different types Pearson, Spearman,
  point-biserial, phi
• Range of possible values
• Information it gives
• Direction of relationship
• Strength of relationship
• Coefficient of determination

3
What if its just due to chance, though?

• Null hypothesis?
• Role of p value
• Larger (farther from zero) test statistic,
  smaller the p
• Magical comparison value alpha
• Relationship between alpha and critical test
  statistic

4
The need for more than one participant

• Why cant a single participant be used?
• Restricted range issues

5
Correlation as a type of research design

• Correlation refers to a type of statistic, to
  find out about a particular relationship between
  two variables
• Also used to refer to a type of research design
• Frequently used interchangeably with
  non-experimental
• (for simplicity, though, use to refer to designs
  where could use a correlation to test hypothesis)
• In contrast to single subject designs we talked
  about no manipulation involved

6
Disadvantages of correlation

• Mantra Correlation does not equal causation
• What does it mean?
• Why does it exist?
• Third variable problem some non-measured
  variable is responsible for the association
  between two other variables
• E.g., birth control and toaster ovens in Taiwan
  in the 70s
• Selection bias similar to third variable
  problem type of person who puts himself or
  herself in a certain situation
• E.g., incidence of respiratory illness in AZ

7
Applying all this

• Bring out worksheet you brought in
• In pairs, complete it, and turn it in
• You can write on it, or you can re-download the
  worksheet at the computers, type your responses,
  and email it to me.

8
Why use correlation at all, then?

• Advantages of correlation?
• Can examine association between things that
  cannot, or should not (think back to ethics!),
  manipulate

9
Putting this into practice

• As a class, will design and conduct correlational
  study
• As individuals, will write up the results of that
  study in APA style

10
How to get started?

• First thing to do figure out a topic
• Should just be one variable
• Should be in the form of a question
• Not a yes/no question
• Should be very broad
• Not yet a hypothesis
• Will then look to previous research to narrow
  down to a hypothesis

11
Types of questions suited to correlation design

• Whats related to sorts of questions
• Whats related to marital satisfaction?
• Whats related to happiness?
• Whats related to a long life-span?

12
To do now

• Think of areas that interest you
• Think back to the areas you examined Thurs, Feb 5
• Think about the research you found on those areas
• Do so more searches, for other research, and on
  topics that you may not have thought of
• Narrow down to one or two general topics you
  might be interested in
• Big commitment well be using the topic we
  decide on for each of the 3 class research
  projects, and youll be writing two papers on it