Ch. 4 Summarizing Data Collected in a Sample Flashcards
Sample
subset of individuals selected from the population at random
should be representative of the population
number of individuals in the sample is less than the number of individuals in the population
First step in analysis
characteristic of interest must be summarized using appropriate techniques
Types of variables (outcomes, endpoints, characteristics, etc.)
Dichotomous, ordinal, categorical, continuous
Dichotomous response
has only two possible responses
Ordinal response
more than two possible responses and the responses are ordered, i.e. symptom severity - minimal, moderate, severe
Categorical response
more than two possible responses and the responses are not ordered, i.e. race/ethinicity, blood type, marital status
Continuous Variables (measurement or quantitative variables)
can theoretically take on an infinite number of values between a fixed min and max
Clustered or repeated measures
multiple measurements taken on the same person
Statistics
numerical summary measures computed on samples
Parameters
numerical summary measures computed on populations
Descriptive Statistics for Dichotomous Variables
responses described as “success” or “failure”
Frequency distribution tables (proportions of each response in the sample)
relative frequencies -
Descriptive Statistics for Ordinal and Categorical Variables
Cumulative frequency - number of subjects at a certain level or below
cumulative relative frequency - useful for summarizing ordinal variables and indicate the percent of patients at particular level or below
Population
collection of all individuals about whom we wish to make generalizations
Measurement/Quantitative variables
can take on an unlimited number of responses between a theoretical minimum and theoretical maximum value
Descriptive Stats for Continuous Variables - Measures of Center
Sample Size - larger usually better but at a certain point, there is not more benefit from more subjects
Sample Mean - Sum all values and divide by sample size, gives sense of the middle of the data (Xbar), report one more decimal place than the statistics measured
Sample Median - middle value in the ordered data set, separates top and bottom 50%s, if even number of data points, median is mean of two middle points, unaffected by extreme values
When there are no extreme values, the mean and median are close in value
Mode - most frequent value - given in addition to the mean and median and not instead of