Module 2 Flashcards
What is a normal frequency distribution
Mean. median and mode are all in the same
Curve is symmetric around the mean not skewed to one side
Negatively Skewed distribution
Mode occurs at the peak
Mean and median pulled to the left
tail to the left
Positively skewed distribution
Mode at center of curve
Mean and median are pulled to the right towards more positive numbers
tail to he right
Curtosis
leptokurtic points
mesokurtic - normal
platykurtic - flat peak
What does an increase in standard deviation do to a distribution or curve
it flattens the curve out or pulls the curve outward
Where does 68% of data fall in normal distribution
within + or - 1 standard deviation of the mean
Where do 95% of scores occur within a normal distribution
+ or - 2 standard deviations from the mean
Where do 99% of data fall within a normal distribution
within + or - 3 standard deviations from the mean
What is a z score and how do you calculate it?
compares the mean of a sample with the population mean
Also a way to identify outliers if they are 3 or greater standard deviations
Z = (individual score - mean) / SD
How do you perform a hypothesis test
- Develop a null and research hypothesis
- Choose a level of significance
- Determine which statistical test is appropriate
- Run analysis to obtain test statistic and p value
- reject or fail to reject the null hypothesis
- Make a conclusion
Hypothesis
statements that predict the relationship between variables
they are testable and translate the research question into a prediction
Null hypothesis
There is no relationship
Research hypothesis can be two things
directional or non-directional
directional vs non-directional
you think something is going to go in one direction
one tailed test
non-directional - you don’t know which direction so you would run a two tailed test
Level of significance
How much error you are okay with
percent of time you say there will be a statistical signifcance but there in fact is not
p = 0.05
5% time you make a type I error
can raise or lower if you want to be conservative or not
Type I error
probability of concluding that there is a difference in groups when there is not
Type II error
probability of concluding that there is no difference between groups or no relationship between variable when there truly is
Probability of accepting a null hypothesis when it is not true
What does which statistical test you choose depend on
number of variables
levels of measurement (nominal, ordinal, interval, ratio)
assumptions of statistical test
P value
probability that the difference, or one larger, could arise by chance
Probability of seeing this outcome by just chance
to reject null hypothesis your p value needs to be lower that the statistical significance score you set 0.05
If calculated p value is bigger than level of significance
fail to reject null hypothesis or ACCEPT
calculated p value is smaller than level of significance you set
reject the null hypothesis and accept the research hypothesis
T test for independent groups when would you use
When you are comparing the means of two independent groups
Groups have to be unrelated to each other
T test for independent groups assumptions
Independence
Normality
Homogeneity of variances
dependent variables must be interval or ratio (comparing means)
How do you determine homogeneity of variances?
By doing a Levene’s test for equality of variances
