variable classification scheme (test 1)

measurement

the process of collecting and recording observations about the variable of interst

data

the observations that are collected and recorded

statistics

summarizing, organizing, presenting, analyzing, and interpreting data

• descriptive

• inferential

descriptive

summarizing, organizing, and presenting data

• reducing a large quantity of data into a manageable (and useable) form

inferential

analytic

• drawing conclusions about a population based on information contained in a sample

variable classifications schemes

1. qualitative data vs quantitative data

2. levels of measurement

3. conceptual roles of variables in research

types of variables in statistics

1. qualitative

2. quantitative

   • discrete

   • continuous

• the methods used to summarize/organize data depends on the type of variable

• variables fall into one of 2 broad categories

qualitative data

• Describe the quality/condition of an individual person

  • “Meaningful information collected in words”

  • Non-numerical format: cannot order/measure

  • Describes things that do not inherently have any numeric value and cannot be ranked

• Examples:

  • Gender

  • Marital status

  • Geographic region

• Collection:

  • Interview transcripts

  • Notes in medical records

quantitative data

• Data collected as numerical or countable info

  • Measures of values or counts 

  • Describe things that inherently have numeric value and can be ranked

  • “anything that is countable in nature”

• Examples

  • Age

  • Weight

  • Blood Pressure

quantitative: discrete

• a few possible values and defined as “counts”.

  • Characterized by gaps or interruptions in the values in the assume

• Examples

  • Number of Emergency Dept. visits (count variable)

  • Number of students in a class

quantitative: continuous

• take on any value (infinite number of values) within a given range

  • has a logical order with values that continuously increase (or decrease) by the same amount

• Examples

  • Age

  • Height

  • Height

  • Blood Pressure

  • Heart Rate

  • Serum Drug Concentration

levels of measurment

• nominal (named levels)

• ordinal (named + ordered levels)

• interval (named + ordered + proportionate intervals between levels)

• ratio (named + ordered + proportionate intervals between levels + can accommodate absolute 0)

each levels builds upon the previous one, offering increasing precision and mathematical possibilities

nominal (no set order)

• data that can be placed into different named categories but with no particular order

  • Categorical data in nature

  • Describes the characteristics of groups with no rank or orders

• Examples: sex, ethnicity, eye color, marital status

• Summary:

  • can be stored as words or text, or given a numerical code

  • E.g., Male (=0), female ( =1)

  • Summarize: frequency or percentage: Male 40%,  female 60%

ordinal (ordered, ranked)

• finite number of well-defined categories with ordering (ranked)

  • Categorical data in nature

  • Meaningful categories and with inherent order or rank

  • Distance between each of the responses is unclear

• Examples: rank, satisfaction, response to treatment (excellent, good, fair, poor), income level

• Summary:

  • have meaning orders but interval between scales may not be equal

  • How satisfied are you with your providers?

  • Very satisfied  Satisfied              Unsatisfied Very unsatisfied

                                   1 2 3 4

  • Summarize ordinal data: frequency or percentage

interval (ordered, =)

• ranked data that contains a meaningful measure of the distance between categories

  • When the scale of ranked data represents meaningful differences between numbers, but still lacks a defined and meaningful zero point

• Examples: Body temperature measurements in Celsius, score on an IQ test

  • If a patient’s body temperature changes from 98.2 F to 102.2 F then it has increased 4 F and this 4 F difference is the same as a change from 98.9F to 102.9 F

  • Interval data are considered to be continuous and may be compared using simple mathematical operations, such as addition (+) and subtraction (-)

ratio (ordered, =)

• Interval scale with a true zero

  • There is a defined and meaningful zero point that denotes “none” of the property being measured

• Example: value of a drug concentration in the blood of zero means that there is no drug in the blood

  • Other examples: Temperature in Kelvin, height of a person, drug concentrations, most laboratory test values

• are continuous and these data can be mathematically manipulated in various ways to yield description of the data, including addition (+), subtraction (-), ratio calculation

  • A patient who weights 105 pounds weighs twice as much as a patient who weights 105 pounds

  • This comparison (210/105) is not possible with data from other levels of measurement

independent variable

• Predictor variable, explanatory variable, exposure variable, or X variable

• is changed (manipulated) by the researcher in order to determine whether it has an effect on a particular outcome; variable that is hypothesized to explain an observed clinical phenomenon

  • What we expect to influence the dependent variable

  • Factors that may influence the outcome

  • Can be set to a desired level (treatment dose in an experimental design) or observed as they occur in a population (characteristics of different groups)

dependent variable

• Response variable, criterion variable, outcome variable, or Y variable

• is the presumed effect, outcome, or response in a study; it is the variable that is to be explained or predicted by independent variable

  • Altered by changes in the independent variable

  • What happens as a result of the independent variable

  • Variable we are predicting (i.e. a disease or outcome of interest)

control variable

• Other explanatory factors that are related to the dependent variable are called control variables

• hold external conditions constant so that the effect of the independent variable may be measured more precisely

• Example

  • A new beta-blocker vs placebo for management of resistant hypertension

  • There may be other factors such as age, race, and physical activity that may influence the blood pressure