Descriptive Statistics
Mode
Modality: # modes found in a population Amodal --0 Unimodal--1 Bimodal--2 Multimodal >2
Interval/Ratio
Mode, median, mean Range, Minimum-Maximum, Standard Deviation, Variance Skew, Kurtosis
Standard Deviation
Most commonly reported variability measure Based on normal "bell curve" Determines the number of values in a data set within a specific interval in a normal distribution Summarizes avg amount values deviate from the mean Always reported with a mean ± SD- reported in units so cannot cross over b/t units
To describe the frequency of the single variable pneumonia in elderly adults age 65-75, which of the following could be utilized 1. descriptive statistics 2. inferential statistics 3. univariate analysis 4. bivariate analysis
1 and 3
Describing risk used with crosstabs data
Absolute risk (AR): proportion of people who experienced an undesirable outcome in an group (experimental or control) Absolute Risk Reduction (ARR): subtract absolute risk for experimental group from absolute risk for control group (Proportion of people who may be spared a negative outcome by treatment/intervention)
Not Normally Distributed
Asymmetric or "Skewed" Peak of the data(mean) not at center of distribution "Tale of two Tails" mean is closest to the tail Positively skewed= Mean > median and mode Longer tail to the right Income—many have small/middle but few have high Negatively skewed= Mean < median and mode Longer tail to the left Age of people with chronic illness, fewer younger/middle aged, more old "Tailedness"
Median
Center of the data set "location not calculation" "Point in distribution that divides the scores in half" (Polit & Beck, 2014, p 218) 50 scores: Median is halfway between 25th & 26th 51 scores: Median is 26th
Descriptive Statistics
Collection and presentation of data to explain characteristics of variables in a study sample Estimate parameters of population
Normal Distribution
Data represented by distinctive "Bell" shaped curve When the two halves folded over, superimposed are the same, then mean=median=mode "tails" never touch base ...distribution is theoretical..
Bivariate
Describing relationship between 2 variables: Crosstabs Correlations Risk: Absolute Risk Absolute Risk Reduction Odds Ratio
Range
Difference between maximum and minimum values in a data set Sample specific, not a stable measure of variability Only based on 2 data points, a change in either will change the range
Frequency distributions are an effective way to present inferential statistics
False-descriptive
Univariate
Frequency distribution—how many times do these values occur? Central tendency—how can we summarize these values overall? Variability—How spread out, or variable are these data?
Crosstabs (Contingency table)
Frequency of 2 variables are cross tabulated Matrix of 4 measures
Nominal
Frequency, percentage Mode
Ordinal
Frequency, percentage Mode, median Range, Minimum-Maximum
Mean
Mathematical "average" Total values added together/total # values Most commonly used measure of central tendency Greatly affected by outliers Larger the sample size, less impact from extreme values Careful attention when appraising data Mean best if no extremes Median best if extremes
Measures of Central Tendency "Typical-ness"
Overall summary One number to represent many Mode Median Mean
Stages in Data Analysis
Prepare data for analysis Describe sample Test reliability of measurement methods Conduct exploratory analysis Conduct confirmatory analysis guided by hypotheses questions, or objectives Conduct post hoc analyses
Measures of variability
Range Standard deviation
Frequency Distribution
Systematic way to display how often values occur, starting with lowest Count vs Percentage (%age needs at least 20 sample size) Appropriate for nominal, categorical Not appropriate for Interval/Ratio
Correlation Coefficients
To what degree are two variables related to each other? Direction: Positive: as one variable increases, the other increases (2 variables move in same direction—hours studying & test grade) Negative: as one variable increases, the other decreases (2 variables move in opposite direction—hours of exercise & weight in kg) Magnitude -1------0------+1 -1=perfect negative, 0= no relationship, +1= perfect positive Pearson's r (need interval or ratio data)
Categories in grouped data must be mutually exclusive
True
Percentages are often used to describe characteristics of samples
True
Inferential Statistics
Used to make inferences about the population based on sample data Make inferences about population
For each descriptive statistic
What is the variable for each of the descriptive statistics? Was the unit of analysis and reported statistic appropriate for measurement level?
Odds Ratio
proportion with adverse reaction vs those without it. Simply speaking—Estimated odds of having negative outcome in those go get intervention vs those who don't get intervention *Used in EBP *Used frequently when discussing clinical significant, and application to practice