1.1 Review of Statistics: The Basics

Question 1

What is science?

Answer 1

Any field of study that uses the scientific methods to evaluate theories
Scientific method-
1. Generate theory
2. Collect relevant data using appropriate research methods
3. Evaluate data with statistics
4. Evaluate theory: accept, modify, or change theory

Question 2

What is a theory?

Answer 2

an explanation for some aspect of the natural world that is examined by the scientific method and repeatedly tested and confirmed or disconfirmed.

Question 3

Good theories must be:

Answer 3

1. Quantifiable- must be something we can gather data to evaluate
2. Falsifiable- must be testable. Must be able to disprove, doesn’t mean you have to.
3. Operationally defined- a clear and specific description of concept or variable in terms of procedures or actions used to measure it
4. Predictive and explanatory- needs to actually predict real situations or behaviors. Must include explanation.

Question 4

Types of studies that test Theories

Answer 4

1. Experimental Study -typically conducted in a lab setting. We control all the conditions except one or more that we will manipulate. We have direct control over the manipulation (condition of the IV)
   Anova, ancova, manova

2.Correlation study-often associated with survey data. Correlate difference variable to determine if associations exist. We can define IV and DV like in exp. study but can’t manipulate them.
Multiple regression, factor analysis, other forms of regression

3.Observational study-can lead to correlation analysis or other descriptive analysis.

Question 5

Independent Variable in experiments

Answer 5

the manipulation or factor that you purposely change in order to see what effect it has on the DV scores.

Question 6

independent variable in correlational studies

Answer 6

the variable that you want to study for its effect on a dependent variable. (cannot manipulate IV)

Question 7

Dependent variable

Answer 7

identified measure that is affected by the independent variable

Question 8

Four types of measurement scales

Answer 8

1. Nominal- variables are often referred to as categorical variables
   no scaling (eye color)
2. Ordinal- similar to nominal scale, but the categories can be ranked (academic rank)
   Distance between categories is not of equal intervals
3. Interval - like ordinal scale, but the distance between categories must have equal increments
   (celsius temp scale)
4. Ratio - similar to interval scale, but must have a true zero point (kelvin)

Question 9

Continuous variables

Answer 9

can have an infinite number of values between any two values
(time, weight, distance in meters)

Question 10

Discrete variables

Answer 10

have a countable numbers of values between any two values (number of mistakes made on task)
Discrete variables are often treated as continues for regression analysis or as categorial for ANOVA

Question 11

Categorical variables

Answer 11

contains a finite number categories or distinct groups
can be nominal or ordinal
(gender, ethnicity)

Question 12

Population

Answer 12

includes all subjects that are to be studied

Question 13

sample

Answer 13

includes some of the possible subjects that are to be studied
By studying the sample, drawn from a population, it is hoped to draw valid conclusions about the larger group (population)

Question 14

generalization

Answer 14

Data is gathered on the sample and inferences are made about the larger population

Question 15

Types of sampling (2)

Answer 15

1. Random Sampling
   2.non- random sampling

Question 16

random sampling

Answer 16

methods of selecting participants such that each sample has an equal probability of being chosen from the larger population

Question 17

non random sampling (3)

Answer 17

1.convince sample-Nonrandom selection of subjects, which is what most of us do
2.matches sample -Is done on demographic variables such as age, gender, education, health status
3.Representative sample-People are carefully chosen to represent larger groups of people (populations)

Question 18

danger with sampling

Answer 18

The sample may not represent the population of interest, causing inaccurate results and conclusions
Samples may be too small or too large
(Appropriate sample sizes are best determined by power analysis)

Question 19

parameters

Answer 19

values associated with populations
pop mean (µ)

Question 20

statistics

Answer 20

values based on samples
sample mean (X)

Question 21

Types of inferential (frequentist) Stats

Answer 21

Parameters statistics
- Count data
- Chi-square
- Logit analysis

Correlational
-Correlations
-Regression
-Factor analysis
-SEM and MLM

Mean differences
-Z test for population-based comparisons
-T test: one group; two independent groups; two dependent groups
-Analysis of variance (ANOVA)
-Two way; factorial (2x2); repeat measures
-Analysis of covariance (ANCOVA)
-ANOVA with Covariates
-SEM and MLM