class 13 Flashcards
purpose of statistics in research
-describe
-explain
-make predictions
-examine causality
-generalize to entire population
-explore and test theories
clinical use of statistics
-analyzing research done by nursing staff
-reading or critiquing published reserach
-examining outcomes of nursing practice
-examine administrative data
-demonstrate a problem or need
-developing administrative reports
what is the purpose of descriptive stats?
describe and summarize data (e.g. averages, frequencies, etc)
what is the purpose of inferential stats?
test hypothesis and/or answer questions, objectives
-allow inference from a sample to a population
4 types of descriptive stats used to summarize data
-frequency distribution
-central tendency
-variability
-bivariate descriptive statistics
what is a frequency distribution?
-data values arranged from highest to lowest, with count of occurrences of each value
-allows to easily see highest and lowest values, and where data are clustered
data can be:
-symmetric vs skewed
-grouped vs ungrouped
normal frequency distribution
middle value is the mean and distributions go up or down following the standard deviation (+1,+2,+3 etc)
-the curve is even on both sides of the mean
kurotsis in frequency distributions
the degree to which the data clusters around the peak or the tails of the distribution
leptokurtic frequency distribution
data clusters around the mean (large peak with steep curve)
platykurtic frequency distribution
data is distributed evenly across the graph (curve has very little peak and low degree of curve)
grouped frequency distribution
data collected is in “groups” i.e. 10-20,21-30, 31-40 etc
positive skewed frequency distribution
the data is “ more negative” aka the mean is a lower number than the median
peak is close to the y-axis
negative skewed frequency distribution
data is “more positive” aka mean is a higher number than the median
peak is further from the y-axis
ungrouped data in frequency distributions
data collected is only 1 number not a range for each answer
i.e. how many kids do you have 1,2,3,4
what is central tendency
indicate what is “typical” (average) of the data
-three indexes: -mean, median, and mode
what is mode
it is a part of central tendency
the value or score that occurs the most frequently
i.e.1846444857 mode is 4
what is median
it is a part of central tendency
the point that divides the scores in half (i.e. marks 50th percentile)
what is mean
it is a part of central tendency
sum of values divided by number of values being summed
what is variability
-measures of dispersion: the way data are spread within a set
2 common types: -range -standard deviation
what is range
it is a part of variability
difference between highest & lowest scores in a data set
-calculated by subtracting lowest score from the highest score
what is standard deviation
it is a part of variability
-summarizes average amount of deviation from the mean
-calculated using every value in data set
-more stable and widely used than range to measure variability
what are bivariate descriptive statistics
-describe relationships between two variables
-text lists various types:
-cross-tabulations
-correlations
-descriptions of risk
what is correlation
it is a part of bivariate descriptive statistics
-to what extent are two varaibles related to one another
-described using correlation coefficients :-direction -intensity
-can be positive or negative, strong or weak
what is a positive correlation?
values move in the same direction i.e. increase in temp correlates to an increase in ice cream sales
what is a negative correlation
values move in opposite directions i.e. increase in AC sales correlates to decrease in sweating
what is a correlation coefficient
measures strength of correlation between to variables
range from -1.00 to +1.00
-sign (- vs +) indicates direction
-absolute value (i.e. value regardless of sign) indicates intensity
-several types:
-pearson’s r
-spearman’s rho
-Kendall’s tau
what is pearson’s r
-part of bivariate descriptive statistics correlation coefficients
-usd to measure relationships in interval or ratio-level data
-used when data are assumed to be normally distributed
-reported as r=x
what is spearman’s rho
-part of bivariate descriptive statistics correlation coefficients
-range from -1-+1
-used to measure relationships in ordinal level data
-may be used when data are not normally distributed
-reported as p=x
what is a risk estimate
-it is a part of bivariate descriptive statistics
-range from 0-infinity
-1=no relationship
-value <1=risk factor
-value >1 = protective factor
-several types: Odds ratio (OR) Relative risk (RR)
what is odds ratio
-it is part of risk estimate
-used to measure the odds of having been exposed to a condition if having developed an outcome of interest
-E.g. taking a sample of women who have postpartum depression and calculating their likelihood of having smoked during pregnancy
-reported as OR=x
what is relative risk ratio
-used to measure the odds of developing an outcome of interest after having been exposed to a condition
-reported as RR=x
-e.g. in a study of the risk of smoking on PPD, taking a sample of pregnant women then examining them 6m after delivery to see if they developed postpartum depression