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;
40 Cards in this Set
- Front
- Back
Non-Experimental Research
|
It's a description of naturally occurring behaviors
Survey Research Observational Research Case Study |
|
Disadvantages of Non-Experimental Research
|
Demand characteristics
Can't do experiment to manipulate/ test everything (maybe due to ethics) Resource intensive (time, money, planning) Lose external validity |
|
Case Study
|
Intensive investigation of a case
Usually only 1 person Can be small group Gives a descriptive account ~Very in depth at multiple levels ~Usually rare clinical/ neuropsychological cases |
|
Drawbacks of Case Studies
|
"n of 1"
Causality issue: was the brain injury the cause of the behavior or something else? Very difficult to reproduce |
|
Observational Research
|
These are naturalistic observations based on naturally occurring behaviors
|
|
Observational Research Based on Ethology
|
Study of naturally occurring behaviors
Ethogram Better in psych to observe to answer specific questions |
|
Ethograms are Used in Observational Research
|
Relatively complete inventory of specific behaviors
|
|
Why is Observational Research Better than Experiments Only?
|
Helps to establish external validity from lab experiment
Also important with ethical/ moral consideration that are in experiments |
|
Important Considerations in Observational Research
|
Coding-operationally define the action or movement
Have inter-observer reliability by having multiple coders for the same observation Frequency of occurrence matters |
|
Controlled/Structured Observations
|
Allow you to control the situation and use carefully arranged conditions
|
|
Participant Observations
|
Researcher makes themselves part of the action
|
|
Reactivity in Observational Research
|
Knowing someone is watching you changes you
Unobtrusive Observations Unobtrusive Measures |
|
Unobtrusive Observation
|
Researcher does not get in the way. Participants don't know they are being watched
|
|
Unobtrusive Measures
|
Indirect observations of behaviors that are a result of specific behaviors
ie: GPA, arrest record, mortality rate |
|
Marketing Research
|
An example of observational research
observe products in use observe license plates socio-econoimic status decision time to make a purchase |
|
Relational Research
|
Attempt to determine how variables are related to one another
Does not involve manipulation 2 types: ~Contingency Table ~Correlational Research |
|
Contingency Table Research
|
Tabular representation of all combinations of categories of 2 variables
2 rows X 2 columns (at least) |
|
Nominal Research
|
Individuals can only be in one cell of contingency table
Analysis: kai2 (squared) examines a relationship if one is present to determine if the 2 variables are independent of one another or not |
|
Correlational Research
|
Research between 2 or more measures
The correlation tells us: Magnitude (degree) of relation Direction of relation |
|
Magnitude of Relation
|
Indicated by the size of the absolute value
The distance form zero (negative or positive) The stronger it is in either distance from zero, the stronger the relationship between variables |
|
Direction of Relation
|
Indicated by sign
Either a positive or negative correlation |
|
Negative correlation
|
If one goes up, the other goes down (opposite)
|
|
Positive correlation
|
If one goes up, the other goes up (same)
|
|
Relation vs Relationship
|
Relationship:
connection, association, or involvement between between PEOPLE Relation: Existing connection/ significant association between or among things The stronger the relation, the closer to 1 or -1 the correlation is BUT this does NOT apply causation |
|
Pearson's Correlation Coefficient
|
Pearson's r
-1, -.8, -.4, 0, .4, .8, 1 If all measures are on a straight line to the right, only 3 std dev. away from 0, they are all considered positive 1 If all measures are on a straight line to the left, only 3 std dev. away from 0, they are all considered negative 1 IMPORTANT: If there are scores greater than 3 std deviations away from the mean they are outliers, and must be removed |
|
Restricted Range (Trunkated Range)
|
This kind of range can change correlation if it's restricted
Even if the data collected is not linear, it can still have relation, so be mindful of that |
|
Survey Research
|
Most common type of research
Easy to do badly |
|
Types of Administration (Survey Research)
|
Self-administered questionnaire
OR Investigator-administered questionnaire (phone or in person) Usually computer aided ~CATI ~CACI Answers can determine the next question that you get |
|
Choosing a Sampling Strategy (Survey Research)
|
Census
Sampling Frame Random Sampling Convenient Sampling Snowball Sampling |
|
Census
|
Asking everyone in a population to take the survey
EG: US Census |
|
Sampling Frame
|
How do you identify the population from which you want to survey?
List everyone in the population to pull sample from |
|
Random Sampling
|
Take a stratified random sample to survey
Every member in the population has a chance to be sampled (This is NOT simply random sampling) |
|
Convenient Sampling
|
Convenient participants are surveyed
EG: the cosmo magazine surveys (dumb girls), Fox News polls (right sided), or CNN News Poll (left sided) |
|
Snowball Sampling
|
Sample who receives the survey can and is encouraged to pass on the survey to ppl they know
EG: find one family with a child with downs... they tell you what other families have kids with downs so you can question them, ect.. |
|
Nonresponse vs Response in Survey Research
|
Volunteer problem
Those who answer may not represent the entire population Volunteers-- more motivated, care more, more patient, or have more free time. Might be more educated so that's why they care to respond |
|
What is Ideal for Survey Research
|
Want to be able to generalize and only a true random sample can allow us to do so
Nonresponse biases may ruin sample and overall conclusion |
|
Polling vs Voting
|
The amount of ppl who vote as compared to those who take polls are vastly different
Remember FDR vs Alf Landon example: Literary Digest Poll predicted Alf, sent out 10 mil surveys, got 2 mill back Gallop predicted FDR, sent out only 50,000 surveys FDR won This happened because the Lit Digest may have sampled more ppl BUT Gallop surveyed from a more broad sample of fewer ppl |
|
Survey Design
|
Fixed Alternative
Open-Ended |
|
Fixed Alternative Survey Design
|
Yes/No
Multiple Choice Likert Scale These are reliable, but not as powerful due to restiction |
|
Open-Ended Survey Design
|
No fixed answer
Difficult to score... How do you code? May not be properly answered More rich data |