Exam 2: Quantitative Data Collection and Analysis Flashcards
descriptive statistics
- summarize or describe data
- reduce large sets of data into meaningful pieces
inferential statistics
- draw conclusions about a population based on data from sample
- looks at differences between groups
correalational statistics
- strength of relationship between two variables
what is quantitative data analysis
- process of organizing data to draw conclusions
how is data in descriptive statistics displayed?
- through frequency distribution
what is frequency distribution in descriptive statistics?
- a table or figure that shows number of data observations that fall into specific intervals
- histograms, pie charts, line charts, table
What is a measure of central tendency?
- it describes the center point of data using a single value
- mean, median, mode
what is a measure of variability/ measure of data dispersion
- describes how far a data set has strayed away from the mean
- standard deviation
mean
average of adding all values and dividing
median
middle score
mode
most frequently occurring
Normal distribution of central tendency
- smooth bell shape curve
- mean, median mode are same value
skewness
- negative is skewed left
- positive is skewed right
range
difference between highest and lowest score (11 - 4 = 7)
variance
reflects variation of distribution within set of scores; used to calculate standard deviation
standard deviation
spread the data of the mean; square root of variance
- most commonly used measure of variance
empirical rule
if distribution scores follow a bell-shaped curve we would expect 68% in 1 SD, 95% within 2 SD, 99.7% within 3 SD.
Descriptive statistics is split into what 2 categories?
- central tendency (mean, median, mode)
- variability (standard deviation, range)
one tailed hypothesis test
- directional test
- research suspect intervention is better, sets up hypotheses to reject null hypothesis
- mean 1 greater or less than mean 2
two tailed test
- non-directional test
- consider either possibility
- null hypothesis rejected if result fails at either tail
- mean 1 not equal to mean 2
what is probability (p-value)
- significance level ( denoted by alpha)
what does low probability (p-value) indicate? less than .05
- lower chance to find effect due to chance
- probability of found different due to chance is low
- reject null hypothesis
Type I error
- error created by rejection null hypothesis when it is true
- test that shows patient has disease but really doesn’t
Type II error
- accepting null hypothesis when its false
- blood test failing to test for disease when disease is present
Student t-test (independent t-test)
- used when there is no relationship between two groups
paired t -test
- used when groups are same
- compares pre and post test of same individuals
analysis of variance (ANOVA)
- test difference between two means
- computes multiple t-test at one time and tells you if there is difference between groups
- compares within group difference between group difference
post-hoc analysis
- after ANOVA, this tells us which groups differ from one another on which specific variables
- tukey scheffe and duncan adjustments test
parametric data
samples that assume normal distribution (bell curve)
- have homogenous variance
- continuous data i.e. weight
non-parametric data
- samples that do not have homogenous variance
- does not follow normal distribution
- nominal (chi square), ordinal scale (mann-whitney U)
Chi-square test
- non parametric statistic
- analyzes frequencies or proportions
- sum of differences rather than expectations
- one tailed test
- coin toss
correlation coefficient
- strength of relationship bt two variables
- no cause - effect can be inferred
- ranges from +1.0 to -1.0; 0 shows no relationship
Know degrees of correlation: strong positive, strong negative, weak positive, weak negative, etc..
visualize it!
correlational statistics relationship test: person r test
used with parametric data
correlational statistics relationship test: spearman rank-order correlation coefficeint
- used with non-parametric data
regression analysis - predictive values
- as one variable increases predict increase in other variables
simple regression
examine relationship bt dependent and independent variable
multiple regression
used when there is more than one independent variable
factor analysis
- identification of factors that compose a construct
- group items together to create a factor
i. e. sensory profile
considering how to choose the right statistic
- consider research question
- how many groups compared
- simplest way to understand data
Ratio
- continuous
- numeric
- can take any value
- height, weight, age
Interval
- discrete
- countable
- ordered numeric; no smooth value to value transition
- # of students, # of strokes
Ordinal
- Categorical
- Natural order
- Education Level, socioeconomic status
Nominal
- Categorical
- No natural order
- Gender, Ethinicity
