E1: Descriptive Statistics

Question 1

Descriptive Statistics

Answer 1

Summarize raw data or simplify our understanding of the data

Question 2

What to keep in mind with descriptive stats?

Answer 2

1. What do we want to know about our data?
2. What is the level of measurement?
3. What patterns can be seen in the data?

Question 3

Central Tendency

Answer 3

average or typical case in the given data

Question 4

Dispersion

Answer 4

How spread out or concentrated the cases are around one or more values

Question 5

Nominal Central Tendency

Answer 5

Mode

Question 6

Mode

Answer 6

Most common value of variable (can have more than 1)

Question 7

Nominal Dispersion

Answer 7

evaluated using percentages.

Question 8

Ordinal Central Tendency

Answer 8

Mode or Median

Question 9

Median

Answer 9

Value of variable at 50% mark.

Question 10

Ordinal Dispersion

Answer 10

Determined by how values are distributed

Question 11

Interval Central Tendency

Answer 11

Mathematical Mean or Median

Question 12

Mathematical Mean

Answer 12

Mean = sum of individual cases / total # of cases

Question 13

How to determine mean or median for interval data?

Answer 13

MEAN is vulnerable to outliers & skewness

Question 14

Skewness

Answer 14

data trends on 1 side of mean rather than even distribution.

Question 15

Nonskew

Answer 15

even distribution on either side of mean

Question 16

Positive Skew

Answer 16

Data slopes TOWARD 0

Question 17

Negative Skew

Answer 17

Data slopes AWAY FROM o

Question 18

Standard Deviation

Answer 18

Average dispersion point of how far each data point is from the mean

Question 19

Normal Distribution

Answer 19

Roughly equal # cases fall on both sides of mean. (bell curve)

Question 20

How to determine Standard Deviation?

Answer 20

1. determine Mean
2. Calculate individual deviations (data - mean)
3. Square each deviation
4. Sum of squares
5. Calculate Variance (divide by (n-1)
6. Calculate square root

Question 21

Index

Answer 21

Creating new variable from a variety of different sources of information

Question 22

Validity

Answer 22

Does conceptual definition match data you are measuring?

Question 23

Reliability

Answer 23

Is how we are measuring consistent? Will all researchers come to same measurement?

Question 24

Types of Validity

Answer 24

1. Face Validity
2. Construct Validity
3. External Validity

Question 25

Face Validity

Answer 25

Does it make sense at first glance? Does it really measure what we want?

Question 26

Construct Validity

Answer 26

Is the variable functioning as we expect?

Question 27

External Validity

Answer 27

Is the way the data was collected allow it to be generalized to the broader population?

Question 28

How can external validity be ensured in some surveys?

Answer 28

oversampling

Question 29

Types of Reliability

Answer 29

1. Split-half 2. Test-retest 3. Cronbach's Alpha 4. Intercoder

Question 30

Split-Half Reliability

Answer 30

Half of sample get asked 1 version of question, other half get other version.

Question 31

Test-Retest Reliability

Answer 31

Repeated testing should show consistency

Question 32

Cronbach's Alpha Reliability

Answer 32

Mathematical evaluation for interval data

Question 33

Intercoder Reliability

Answer 33

For subjective data. Researchers trained to code data a certain way. Reliable if different researchers code same data the same way.

Question 34

Additive Index

Answer 34

Easiest index to create. Variables that are measured same & of equal importance are added together to create new variable.

Question 35

Importance of recoding data`

Answer 35

1. Simplify/consolidate values to create ordinal scale 2. Fix codes so that values across a set of variables match 3. Create dummy variables to transform nominal data into something else. 4. Visual binning to consolidate interval data into equal sized groups

Question 36

Why run a frequency table before recoding?

Answer 36

To decide how to recode. To know what to expect after collapsing data.

Question 37

Mean-Centered Data

Answer 37

creating a scale for variables where 0=mean

Question 38

Why create Mean-Centered Data?

Answer 38

Converts confusing data points into more familiar data that is easier to read.