Principles of Measurement and Descriptive Statistics

Define Data

• info collected during systematic observations

• result of measurement

Define variable

factors collected, measured, manipulated, observed

Independent variable

• intervention

• what’s being manipulated in the study

Dependent variable

• outcome of interest

Confounding variable

variable other than independent variable that can affect the outcome

Population

all members of a certain group

Sample

portion of population

1. random sample

2. representative sample

3. convenience sample

Define random sample

every member of the pop has an EQUAL chance of being selected

Define representative sample

researchers match the larger pop on specific characteristic

Define convenience sample

participants selected based on proximity, ease of access, or willingness to participate

Parameter

describes characteristics of population

Statistic

characteristics of a sample

Qualitative data

• non numeric or categorical → groups

  • sex, race, ethnicity

2 types of qualitative data

1. nominal

2. ordinal

Quantitative data

• numeric → measured

  • age, BP, temp

2 types of quantitative data

1. interval

2. ratio

Discrete data

• fixed # of values → whole #s only

  • # of children, # of live births

Continuous data

• infinite range of values

  • height, weight, temp

Qualitative: Nominal data

• classified into groups w/ exclusive categories with no natural order

• BINARY data (falls into one of 2 categories)

Qualitative Data: Ordinal

• ranked in a specific order or categories that have a rank

  • amount of pain (mild-severe)

Quantitative Data: Interval

• continuous data that does not have a true/meaningful zero point

• measured by standard units

• comparable but cannot be multiplied or divided

  • ex: temp in F or C

Quantitative Data: Ratio

• continuous data measured in units where a true zero represents total absence of that unit

• can be multiplied and divided

  • ex: BP, HR

define Mean

average value

define median

middle value

define mode

most common value

Range

interval b/w lowest and highest values within a data group

Percentile

percent of individuals or % of data falling BELOW a particular value

Interquartile Range (IQR)

range b/w score that marks the 25th percentile and 75th percentile

• Q1

• Q2 = median

• Q3

How to calculate IQR

Q3 - Q1

Positive skew is skewed to the

right

Negative skew is skewed to the

left

How to calculate outliers

1.5 x IQR

• if # falls more than 1.5 x IQR above Q3 or below Q1

Standard Deviations

• 99.7% b/w ± 3 SD

• 95.4% b/w ± 2 SD

• 68.3% b/w ± 1 SD

When the data is not normally distributed, what should be used to describe the central tendency?

median

T/F: the mean is sensitive to outliers

true

T/F: the mean works well when the distribution is symmetric with NO outliers

true

Pie charts

used w/ categorical data

Line graphs

shows how values change over time

Histogram

• representation of distribution of data

• continuous/discrete

What is an epi curve?

type of histogram that displays the # of new cases of a disease over time

What is the best way to display the distribution of quantitative variables?

histogram

Box Plots

• quantitative data, sometimes ordinal

• easily identifies outliers

Scatter Plots

• used to graph 2 continuous variables

• easily identifies outliers

What graphs would a reader be able to visualize outliers?

• box plot

• scatter plot