Statistics in Allied Health Flashcards
What is the purpose of statistics?
- control sources of variation; detect outliers
- analysis of data
- interpret the statistical and practical significance of results
- in making scientifically sound decisions and communicating them
What are descriptive statistics?
- describe what was observed in the sample numerically or graphically
What are some common numerical descriptors?
Mean and standard deviation for continuous data types (i.e. age)
Frequency and percentages are useful for describing what data?
categorical data i.e. gender
What is inferential statistics?
uses patterns in the sample to draw inferences about the population represented
What are some examples of inferential statistics?
- Hypothesis testing: yes/no questions about the data
- Estimation: estimating numerical characteristics of data
- Correlation: describing associations within the data
- Regression analysis: modelling relationships with the data
What can inferential stats be used for?
- forecasting, prediction and estimation of unobserved values either in or associated with population studies
What is a sample as compared to a population?
- sample is a subset of the population
- population is a group of people with a common trait
What is one of the most important factors of the sample?
It’s representation on behalf of the population as we use the sample to infer about the population.
What are the two types of sample selections?
- probabilistic: everyone has a similar chance of being selected
- non-probabilistic: not everyone has the same chance of being selected
What are some probabilistic sampling methods?
- simple/stratified random sampling
- systematic random sampling
What are some non-probabilistic sampling methods?
- convenience sampling
- snowball sampling
- purposive sampling
What are the 4 types of variables?
Nominal, ordinal, interval and ratio
What is nominal data?
numbers given to a variable have no significance
What is an example of nominal data?
0 = male, 1 = female
What is a dichotomous variable?
a variable (nominal) with two possibilities
What is ordinal data?
the order of the number holds significance
Give an example of ordinal data
- pain scale
- likert scale
What is interval data?
continuous variable, the value of 0 does not indicate the absence of a quality
Give an example of interval data
temperature
What is ratio data?
continuous variable where the value of 0 does indicate the absence of a quality
What is an example of ratio data?
weight
When is a continuous variable classified as discrete?
if it is restricted to a fixed number of values
What is raw data?
value the same unit as it was measured in
What is scaled data?
units of measurement given a relative value that makes it comparable to other values in the general population
What is the criterion variable/outcome variable?
The presumed effect in a study
What is the predictor variable/covaraite
The presumed cause in a experimental study (potentially associated with the outcome variable)
What is relative frequency?
The frequency in a subgroup relative to the total number
What are mutually exclusive variables?
Variables that don’t overlap
What is a example of a mutually exclusive variable?
age
What data does bar graphs show?
crude data not relative data
What data does pie graphs show?
can show both crude and relative data
What do scatter plots show?
correlation between two continuous variables simultaneously
What do line graphs show?
time trends where the x axis shows the unit of time and the y axis displays the values of the variable being plotted
What do histograms show?
differences in frequencies or percentages among categories of continuous variables
In a histogram, what is the 1) width of the bars 2) height of the bars proportional to?
1) width of the category
2) frequency of percentage of that category
What does the histogram present?
distribution of data, gives an estimate of the probability distribution of continuous variable. Shows how skewed/shifted data is
In a histogram, if the data is left tailed what does it mean?
negative skew
In a histogram, if the data is right tailed what does it mean?
positive skew
What does a box plot body represent?
the first to third quartile
How are outliers plotted in a box plot?
separately as positions on the chart
What is the line in the box plot body?
median of data set
What are the whiskers in a blot plot represent?
bottom whisker - Quartile 1 to lowest non-outlier
top whisker - Quartile 3 to largest non-outlier
What are some measures used in quantitative research?
- counts i.e. number of patients
- proportion: relative frequency, must be divided by the total number in the group
- rates: used to involve or imply a relationship
- measures of central tendency: measures that best represents the data
What are some measures of central tendency?
- mean
- median
- mode
How is the mean calculated?
add all observations and divide by number of observatios
What is the mean effected by?
In repeated measures but is easily effected by extreme values
What is the median
The middle value i.e. 50% of data above and below the median
Is the median effected by extreme values?
no
What is the mode?
most commonly occurring value
Is the mode effected by extremes?
no
What are measures of variability?
Inform of the spread of the ata
What are some measures of variability?
standard deviation, range and interquartile range
What is variance?
average of the squared differences of each of the observations from their mean
What does a low and high standard deviation mean?
- low: little spread of data
- high: large spread of data
What must the sum of deviations around the mean equal?
0
What is the normal distribution?
a curve that is bell-shaped and symmetrical around the mean
What is the area under a bell-curve equal to?
1
In normal distribution, what values are equal?
mean, median and mode
How many values lie 1 standard deviation away from the mean?
68%
How many values lie 1.96 standard deviations away from the mean?
95%
How many values lie 2.58 standard deviations away from the mean?
99%
What is the confidence interval?
limits within which the true population probably lies. Gives a range of values that may reasonably contain the true population parameter
What are the parameters of the confidence interval?
lowest value (lower confidence limit) and highest value (upper confidence limit)
What is the 95% confidence interval?
the range of scores or values in which it is 95% confident that the true population mean lies
What is a null hypothesis?
A is similar to B or A is not different to B
What is alternative hypothesis?
A is different from B or A is larger/smaller than B (there is a relationship between variables)
What is a type 1 error?
when we reject the null hypothesis when it’s true
What is a type 2 error?
when we accept the null hypothesis when it’s not true
What is the power?
the probability of rejecting the null hypothesis when it’s false
What is the P value?
significance of comparison
What does a P value smaller than 0.05 mean?
researcher is confident enough the reject the null hypothesis and accept alternative hypothesis
What can a P value tell us?
observed difference between compared measures could have been obtained by chance alone
How can nominal data be analysed?
non-parametric statistic techniques i.e. counting individuals in groups
What are some weaknesses of ordinal scales?
may be subjective - terms need to be clearly defined
How can ordinal data be analysed?
non-parametric statistic techniques
What are the continuous variables?
interval and ratio data
How can interval and ratio data be analysed?
parametric statistical techniques
What is an accurate measure?
One which close to true population value
What is a precise measure?
one which yields close to the same value with many repetitions
Where does systematic error arise from?
measuring, collecting, analyzing and interpreting data
What does a CI of 90% mean?
90% sure the patient’ true score lies within a determined range
What does a P values of >0.05 indicate?
results were chance findings
What do percentiles indicate?
how many lie above and below the score
What are the quartiles?
1st quartile: 25th percentile
2nd quartile: 50th percentile
3rd quartile: 75th percentile
What are the deciles?
First Decile: 10th percentile
Second decile: 20th percentile etc.
