unit 4 exam Flashcards
what is a longitudinal design?
design where participants are assessed at two or more points in time
what 4 things does a longitudinal design wish to examine?
1.what happens over time
2.what precedes what
3.what follows what
4.mediation
what is a prospective longitudinal design?
follows participants into the future
what did the NY high risk study of schizophrenia study do?
-tracked children from ages 5-30 who had parents who either had schizophrenia, a mood disorder, or neither and saw how they developed psychologically
what did the NY high risk study of schizophrenia find?
-children who had parents with schizophrenia were more likely to develop schizophrenia
what is a retrospective longitudinal design?
retrieves information about participant’s past
what did walker’s study of childhood home movies do?
looked at adults home movies to see if they had motor issues as children
what is the major methodological issue with prospective longitudinal design?
attrition: participants dropping out of the study
when there is attrition (participant dropout), what is important to do?
test whether drop-outs differ from participants who do not drop out in order to see if there is a systematic bias
what is the major methodological issue with retrospective longitudinal design?
may not be able to find retrospective data for all individuals you want to include in study
Being expensive and time consuming are disadvantages of which type of longitudinal design?
prospective design
are retrospective designs cheaper and quicker than prospective designs?
yes
The information you want probably not being available is a disadvantage of which type of longitudinal design?
retrospective design
can you measure exactly what you want with prospective design?
yes
True or false: longitudinal designs can prove causation.
false
According to longitudinal design, if A is hypothesized to contribute to B, then___ at time 1 should predict ____ at time 2.
A, B
If A does not precede B, you can rule out the possibility that _________.
A causes B
Just because A measured at time 1 predicts B at time 2 does not mean that A causes B. What are the 2 alternative explanations for this?
1.a third variable (C) causes both A and B
2.B causes A and then if B is stable enough, A will predict B again
what should you do if you want to rule out alternative explanations?
measure A and B at both time 1 and time 2 and control for B at time 1
when should you conduct a t-test?
when examining a single variable and there are only two levels of the variable
when should you conduct an analysis of variance?
when examining single variable and there are three or more levels of the variable
If there is one group with multiple times/multiple measures per group, which kind of measurement should be used?
paired sample (dependent ) t test
If there are two or more groups with multiple times/multiple measures per group, which kind of measurement should be used?
repeated measures analysis of variance
what is effect size?
measure of magnitude of an effect such as the difference in mean scores or association between two variables
when comparing two means, effect size if measured using what? what is the formula
cohen’s d
d = (mean 1-mean 2)/SD
When there is a bigger effect size, there is ____ overlap.
less
In which type of hypothesis will there be a difference but no prediction of the direction of the effect?
two-tailed hypothesis
In which type of hypothesis will there be a difference and a prediction of the direction of the effect?
1-tailed hypothesis
what does null hypothesis significance testing test?
the probability of obtaining the results you obtained if the null hypothesis is true
what is statistical significance influenced by?
-sample size
-effect size
what are the possible errors of null hypothesis significance testing?
-type 1 error
-type 2 error
what is a type 1 error?
reject the null hypothesis when the null hypothesis is true
what is a type 2 error?
fail to reject the null hypothesis when the null hypothesis is not true
type 1 errors and type 2 errors have what kind of relationship?
inverse
is a correlation of 0.50 with a sample size of 10 statistically significant?
no
is a correlation of 0.10 with a sample of 500
statistically significant?
yes
what are the criticisms of null hypothesis significance testing?
-the cutoffs are unreasonable (p = 0.049 vs p = 0.051)
-statistical significance says nothing about the magnitude of effect
what are the proposed solutions/alternatives to null hypothesis significance testing?
-confidence intervals
The more confident you want to be, the _____ must be the range/interval
larger
the smaller the range/interval, the ____ confident you will be.
less
