Data Analysis Exam 1

Sample basic requirements

Representative of population, randomly selected

What is a naturally occurring discrepancy in statistics?

Sampling error: sometimes a sample doesn’t accurately depict a population, margin of error

Descriptive Statistics

Seek to describe important characteristics of data sets. Ex: what scores? High or low? Close tg?

(Simplify and summarize data)

Inferential Statistics

Procedures for inferences about population based on sample characteristics

(Study samples and generalize about populations)

Two basic types of research

Correlational designs: if two variables are associated in change (sometimes can’t manipulate/unethical)

Experimental: manipulate one variable, random assignment, causation

Non-experimental: quasi-independent variable: not truly independent (no manipulation) but still what causes change (ex: based on pre-existing groups like gender, or before/after studies)

Variables must be clearly defined. Two kinds of definitions?

Constructs: internal characteristics, can’t be directly observed

Operational Definitions: identifies and defines terms, links concepts w/ very defined concrete observations

Discrete vs. continuous

Discrete: no gaps

Continuous: gaps

Nominal data

Categories

Ordinal data

Number rank in order, with space between not equivalent (ex: 1st, 2nd, 3rd)

Interval Data

Numerical amounts w/ no true zero (ex: temp)

Ratio data

Numerical amounts w/ true zero

Real limits

The numbers included in intervals to make sure continuous variables have no gaps

Where does summation go in PEMDAS?

Before addition/subtraction

Histogram x and y and what type of data

X (frequency) vs. y (dependent variable)

Interval and ratio, grouped designs

Bar graphs and type of data

Have gaps, nominal/ordinal data

Describing distributions

Symmetry (or skew), modality

Measures of central tendency + best for…

Mean, median, mode

Mean: interval/ratio data

Median: ordinal, interval, ratio

Mode: all

Mean vs. median

Non-resistant vs. resistant to outliers

Median disadvantages

Can’t “just calculate” it and differs greatly sample to sample (so varies from population)

Finding median

n+1/2

When removing/adding data, when will the mean remain the same?

Adding/removing the same number as the mean or all scores are the same

Variability and very important rule for exam

Range, standard deviation, variance. Variance/SD can never be negative, only 0 if all scores are the same

Range disadvantage

Non-resistant

Variance advantage

Most accurate measure of variability

Standard deviation interpretation

“On average, the scores in (sample) deviate from the mean by (SD)

To be an outlier…

3 SD away from mean

SD/Variance with constants

Add/subtract: no change

Mult./divide: do same thing/square same thing

Biased statistic

Consistently over/underestimating population parameter

Standardizing distributions

Turning all values into z-scores

Observations when “standardizing” a distribution

Standard deviation: 1

Mean: 0

Shape remains

Percentages with normal distribution

68, 95, 99