Lecture 10, 11: Analyzing Data Flashcards
To look for the effects of the IV we examine the ______ for the DV
means
*if the independent variable had an effect on the DV the means of the experimental conditions should differ
The problem with only observing the means of experimental conditions is what? and why?
whats a solution
means can differ while the IV had no effect
confounding variables may be present
error variance = random influences of variables that remain unidentified in a study
estimate the means difference when the IV has no effect –> if the variance observed after IV is larger than that of the estimated variance without the IV then difference can be attributed to the IV
What are the three stat tools researchers use to estimate the probability that differences is due to error variance to help decide if a difference in means was a result of a real effect of the IV?
significance testing: determine the probability that the difference between the means is due to error variance
effect sizes: examine the size of the difference to see if its noteworthy
confidence intervals: judge the difference between means relative to the precision of the data
Significance testing/ null hypothesis testing
used to determine if the observed effect is “real”/due to the IV or due to error variance
What does the null hypothesis assume?
there is no effect in the population
the IV had no effect on the DV
what does it mean to reject the null hypothesis?
the IV did have a statistically significant effect
the difference between means of the experimental group was larger than expected given the amount of error variance in the data
If you fail to reject the null hypothesis you are claiming that…?
the IV had no effect on the DV
the difference between means reflects nothing more than the influence or error variance
What does a p value of 1 indicate?
the difference between the means is exactly what one would expect based on the amount of error variance
Type I error
researcher rejects the null hypothesis that is true and concludes IV has had an effect on the EV when it has not
Type II error
Researcher fails to reject the null hypothesis when it is false. Concludes the IV did not have an effect when it did
How do researchers try to reduce the likelihood of making a type I error?
they try to design experiments with high power
power = a studies ability to detect any effect of the IV that occurs
studies with low power may fail to detect the IVs effect on the DV
T tests are used to
you report them in the following format:
test the difference between two means
italicized lowercase t(df)=t-value, p<0.05, 0,01 or p=n.s.
what is a t value ? what is a p value? how do they relate to one another?
t value is a comparison of differences between two groups to an index of error variance
a p value is the probability that the differences between means in the population is zero
as t value increases p value tends to decrease / there is association with lower p values and larger t values –> this makes sense because the larger the difference between the means the more likely this variance is not due to error variance
what is a way of preventing a type I error ? what is its downside?
Bonferroni adjustment = desired alpha level is divided by the number of tests run and then only consider something significant if it has a p value less than this new value
while it reduces a type I error it increases the chance of a type II error
What is an ANOVA?
analysis of variance between more than two conditions means
it is based on the F-test
describe the F-test
its a ratio of the between group variance to within group error variance
if the IV has no effect the between group and within group variance will be about the same and F roughly = 1.00
the larger the variance between group as compared to within group variance the larger the F value and the more likely differences detected are due to the independent variable
how do F value and P value relate
the larger the F value the lower the P value tends to be
what is the format for reporting an F test/ ANOVA results
upercade italisized F(df,df error)= f-stat, p<0.05, 0.01 or p= n.s.
what is a post hoc test? why/when are they used?
they are used after significant findings from an Ftest are revealed.This is because a significant anova finding will reveal that differences in means occured but will not inform where they occured. Thus post hoc test inform researchers where the differences in means occured / which conditions means differed
What information does an ANOVA provide in a factorial design with more than two IV ?
the error variance
the main effect of IV A
the main effect of IV B
the interaction between IV A and IV B
simple main effect
the effect of one IV at a particular level of another IV
Paired t-test
used for two groups
within-subjects ANOVA
experiements that involve more than 2 conditions including factorial designs
MANOVA
tests for differences between means of two or more conditions on two or more DV simultaneously
the DV measured all tap into the same concept
they help control type I error
the calculated value of F is based on the ratio of between groups variance to _____ variance
error
