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;
18 Cards in this Set
- Front
- Back
|
Why do t test instead of z test |
t uses sample variance so it is used when population variance is unknown |
|
|
Why t has more variability than z |
because it uses sample variance which changes from one sample to another therefore increasing variability, z uses population variance which stays constant across samples |
|
|
Plain English for confidence interval |
I am 95% confident that birds will spend an average of between 2 and 3 minutes on the plain side of the box |
|
|
Sample size influence on Independent Measures t |
size of samples influences the denominator of the t statistic because it influences ESE, as sample size increases the value of t also increases and so does the likelihood of rejecting the null hypothesis
|
|
|
Variability influences on Independent Measures t |
variability influences the denominator in the t statistic because it influences ESE, as variability increases the value of t decreases and so does the likelihood of rejecting null hypothesis |
|
|
Homogeneity of Variance |
variances are equal for the two populations from which the samples are obtained, even when the sample variances differ **if this assumption is violated the t statistic can cause misleading conclusions for a hypothesis test |
|
|
Advantage of repeated measures design |
>reduces or eliminates problems caused by individual differences >individual differences increase variance so eliminating this factor allows researcher to focus on the effect of the variable being studied |
|
|
When to use repeated measures |
>when it's difficult to find many subjects who qualify >when responses are across time or developmental questions >when there are large individual differences |
|
|
Carryover effect |
When the 1st treatment has a lingering effect that is observed in the 2nd treatment round |
|
|
Progressive error |
the subjects performance or response changes over time |
|
|
Precision vs. confidence |
>there is an inverse relationship between the two >more confidence=larger interval=less precise >less confidence=smaller interval=more precise |
|
|
Relationships with confidence interval width |
>larger sample=smaller ESE=smaller interval >smaller sample=larger ESE= larger interval >higher confidence %= bigger interval >bigger variability=bigger ESE=bigger interval |
|
|
How to tell significance based on confidence interval |
>if confidence interval includes zero then the effect is not significant >if effect is significant all values in the interval will be on the same side of zero and null hypothesis will be rejected |
|
|
Why is F-ratio expected to be 1.00 when null hypothesis is true |
When there is no treatment effect the numerator and denominator of F-ratio both measure the same sources of variance so ratio=1 |
|
|
Similarities between F-ratio and t-statistic |
>compare actual mean differences between sample means with differences to be expected if null hypothesis is true >if numerator is significantly bigger than denominator we conclude there is a significant difference |
|
|
Why use ANOVA instead of multiple t tests |
each t test involves a risk of Type I error so more tests means more risk of error and ANOVA allows us to simultaneously test all samples with only one fixed alpha level |
|
|
Post hoc tests |
>used after we have determined a significant difference exists >not used if you fail to reject null hypothesis >tells us exactly which treatment effects are significantly different |
|
|
3 things that matter with two-factor ANOVA |
1) is there a main effect for variable A 2) is there a main effect for variable B 3) is there an interaction effect between variable A and B |
