Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
26 Cards in this Set
- Front
- Back
|
Z-score:
|
sample-mean / SD (+ if directional)
|
|
|
S squared (sub 1, sub 2)
|
Estimated variance of populations for indepedent t-test. S pooled is teh pooled estimate of population variance.
|
|
|
S squared (sub m1, sub m2)
--S squared (sub difference) |
Variance of the distributio of means for populations. -- Total variance of the distribution of differences between means.
|
|
|
t-tests
|
Compare our sample mean to the DISTRIBUTION OF MEANS of the population. Effect sizes look at how extreme the mean sis compared to the estimated population distribution.
|
|
|
Five Experimental Characteristics:
|
1) IV on DV 2) Two IV levels (control – a comparison) 3) Random assignment 4) Control procedures for causal relationships 5) Controls to reduce internal validity threats.
|
|
|
One way vs three way anova
|
1 vs 7 effects (no vs four interactions)
|
|
|
ANOVA: df:
SS within: |
k (# levels) - 1
SS within: how much people are different from others in their own condition. unsys var = indiv diff + error |
|
|
ITS
--Pretest-posttest |
One group tested repeatedly. --No control; commonly used.
|
|
|
T-Distribution versus Z-Distribution.
|
T : Fatter tails, extremes more likely especially with smaller sample sizes. **As sample size approaches infinity, t-distribution matches z-distribution; harder to reject the null.
|
|
|
One-way ANOVA
|
(with 3 or more levels) and 1 Dv, right? It has 1 DV, but can have any number of levels. A one-way ANOVA just has one factor (e.g. ethnicity) but this can have any number of levels.
|
|
|
--Chi-square test for independence has __ DV
--Finally, are RM ANOVA and Interrupted Time Series tests pretty much the only ones measured on numerous occasions? |
1
--It depends on the circumstances. You also could use a dependent measures t-test if measuring performance before and after a given manipulation. The t-tests don't work with 3+ groups however, so then you would need a RM ANOVA or ITS. |
|
|
--1) Step-wise multiple regression compares means (vs. scores)?
|
The means for 2 groups?
It compares whatever DV you want to examine. This can be means, scores, IQs, foot sizes, etc. It can also examine any number of groups, which are simply coded as levels of factors. |
|
|
A) Correlation compares means?
B) Does reliability necessarily measure scores? |
Again, correlations compare whatever DV you want to look at. (Jena: YES)
B)In general reliability tests look at how "replicable" results are. NOT EXACTLY - IT'S LESS ABOUT WHAT IT MEASURES, AND MORE IMPORTANT THAT YOU KNOW THAT RELIABILITY IS A WAY OF ASSESSING WHETHER SOMETHING IS A GOOD TEST OR QUESTIONNAIRE. SO I GUESS, YES, IT'S GENERALLY MEASURING SCORES, BUT IT'S MORE ABOUT ITS PURPOSE |
|
|
For #3 in the advanced procedures homework, (a. A study examining the effect of time of day on speed and performance on a complex task, in which a group does the task both at night and in the morning; Answer: Dep means t-test):
Aren't speed and performance two dv/levels, so you'd use a 2x2 anova? |
WELL, BUT REMEMBER THAT YOU'RE ASSESSING THE SAME PERSON TWICE, SO IT'D BE A DEP MEANS T-TEST SO THAT YOU CAN TAKE INTO ACCOUNT THE WITHIN PERSON VARIABILITY
|
|
|
Step-wise multiple regression compares means? 2 groups? I'M
|
YES - YOU'RE LOOKING AT MEANS. AND THERE'S NOT A SPECIFIC NUMBER OF GROUPS NECESSARILY, JUST THAT YOU'RE LOOKING AT MORE THAN ONE FACTOR.
|
|
|
For the same effect size, a dependent means t-test has more power with half the sample size as that of an independent means t-test
|
YOU'D HAVE MORE POWER WITH A DEP MEANS T-TEST BC IF YOU'RE TESTING THE SAME PEOPLE TWICE, YOUR CHANGE FROM ONE MEASUREMENT TO THE NEXT WON'T BE AS GREAT BC PEOPLE ARE PRETTY SIMILAR WITHIN THEMSELVES, BUT PERSON A MIGHT BE VERY DIFFERENT FROM PERSON B (AS YOU'D BE LOOKING AT WITH AN INDEP MEANS T-TEST, SO YOU'D HAVE GREATER VARIABILITY AND THEREFORE MORE ERROR ASSOCIATED WITH AN INDEP MEANS T-TEST).
|
|
|
A) LSD Test: Comparison distribution
B) Ind Means t-test: The comparison distribution |
F
B) a distribution of differences between means. With equal N, these two estimates are averaged (the pooled estimate of population variance) |
|
|
Effect size (d=m/s) vs. power
|
not influenced by a large N.
|
|
|
F is positively skewed...
|
variances can only be positive!! But can be less than 1. Does NOT indicate direction!! (Use Fisher’s LSD)
|
|
|
ANOVA vs RMANOVA
|
(between df, within df) However, for a RMANOVA, (Within df, Error df)
|
|
|
Dep means vs One-way anova:
|
1) Less than 3 means. Ex. Sleep or no sleep on driving performance. (All have one dv, even rmanova or dep means)
|
|
|
The resulting data set has over 1600 data points for each mom-to-be. What statistical analysis would you recommend to help the researcher make these data more manageable, if the researcher is interested in measuring fluctuation over time?
|
- Exploratory Factor Analysis
- Confirmatory Factor Analysis - SEM NOT ITS |
|
|
Three types of fat: peanut oil, corn oil, and lard. The scientist wishes to determine whether the amount of fat absorbed depends on the type of fat used.
|
[Three conditions]
- one-way anova (not mult reg or correlation) |
|
|
What analyses would you recommend that he conduct on the class data to determine if there were any problems with the exam itself? ((The problem is most likely with the scale, multiple regression may come later)) **Quality of test
|
20. Reliability
10. Cronbach’s Alpha 21. Item Response Theory 23. Item-total correlation Not factor analy or equation modeling. |
|
|
Using the pilot data what analysis can she run to determine if her survey is a good one?
|
- reliability
- item-total correlation - cronbach’s alpha - item response theory - **exploratory factor analysis (NOT mult. reg or confirmatory factor analy) |
|
|
In addition to the word recall task, the intrepid cognitive psychologist wants to determine if level of cognitive activity, such as reading books, working crossword puzzles, or attending lectures, changes the relationship between word recall task performance and chronological age. She believes that knowing a person’s level of cognitive stimulation will be more informative in predicting their word recall performance beyond knowing their chronological age. What statistical test can the researcher use to determine if her hypothesis is correct?
|
The effect of cognitive activity (levels 1, 2, 3) and age on word recall.
(2 IVs, 1 DV) - ****HIERARCHAL MULTIPLE REGRESSION |
