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;
15 Cards in this Set
- Front
- Back
|
What are the steps in choosing individuals to study from a population?
|
1. Define the population
2. Define the sampling frame ---> The people you can actually grab 3. Choose the sampling method 4. Choose the target sample size ---> Practical costs vs. statistical power benefits |
|
|
Reasons studies might not give valid conclusions are called
|
- Threats to validity
|
|
|
Internal Validity
|
Does the data support the conclusion?
|
|
|
External Validity
|
Do the findings generalize to the broader situation?
Ex) An experiment is run properly and analyzed to give a reasonable conclusion about mice, but it might not be reasonable to infer about humans |
|
|
The whole goal of experimental design and entire purpose of research is to...
|
- To do studies that teach us what's going on in the real world or how to change the real world according to our desires (this is why the choice of which research subjects to sample is critical)
|
|
|
Types of Sampling:
|
1. Simple Random
2. Systematic 3. Stratified 4. Cluster 5. Convenience Sampling |
|
|
Simple Random Sampling
|
- Each person with unique chance of being selected
- Does not depends on traits - Does not depend on anyone else - Only determined by a random event |
|
|
Systematic Sampling
|
- Usually based on some kind of order or alternating pattern
- Could be affected by the order people show up for assignment - Randomize the starting position Ex) Counting one, two, one, two |
|
|
Stratified Sampling
|
- Changing the sampling procedure based on traits of the participants
Ex) Changing the probability of selection for people of a particular type |
|
|
Why stratify?
|
- To get the same proportion of each subgroup in sample as the population
- To intentionally over-sample minority groups so that the study has enough statistical power to draw conclusions about them - To intentionally over-sample highly variable groups for statistical power - To find ways to save money |
|
|
Cluster Sampling
|
- Select people in groups rather than individually
- Members of the same group are not statistically independent ---> If members of a group are very similar to each other, the greater the reduction in degrees of freedom (DF = number of clusters instead of number of people) |
|
|
Convenience Sampling
|
- No procedural requirements at all
- Not random - Maybe clusters - Maybe stratified - Likely to be biased |
|
|
Recoding
|
- Dummy Coding = A process where you take names and assign them numbers to make them available for mathematical analysis
---> Nominal (Categorical) to Quantitative - Transforming quantitative to categorical ---> Use diagnostic thresholds Ex) Blood test whether high or low in glucose, choose "normal" or "not normal" |
|
|
What type of recoding is your BMI?
|
- Quantitative to ordinal
|
|
|
What type of recoding is:
Psychiatric interview --> Survey points --> Diagnosis |
- Quantitative to nominal
|
