Term
| Large mean difference, large SEM, you get a ____ t statistic. |
|
Definition
| small While a large mean difference does exist, we are relatively unsure about what the true population mean is, so we can’t be too sure if it is different from a value |
|
|
Term
| Small mean difference, small SEM, you get a ____ t statistic. |
|
Definition
Large t-statistic
The means are not very different, and we aren’t very sure about what the true value of the population mean is, so we can’t be very certain about whether the mean is equal to or different from a value |
|
|
Term
|
Definition
probability of obtaining our test statistic (t hat) given that (vertical bar) the null hypothesis (Ho) is true.
or probability of getting a test statistic as extreme or more extreme than what we had. |
|
|
Term
|
Definition
| distribution we'd use if null hypothesis is true, always centered on zero |
|
|
Term
Does the test statistic fall within the central 95/99/90% of the distribution?
If yes,
If no, |
|
Definition
if yes, they are the same
if no, they are different |
|
|
Term
| Large mean difference, small SEM, you get a ____ test statistic |
|
Definition
|
|
Term
| Small mean difference, large SEM, you get a _____ test statistic. |
|
Definition
|
|
Term
| T distribution is always centered on ___, because: |
|
Definition
| 0, because if the Null Hypothesis is true, the mean would be 0. |
|
|
Term
|
Definition
Ho: u= theorized value
Ha: u not = theorized value |
|
|
Term
|
Definition
true state of affairs is the Ho is true, we find a difference
ex: person is innocent and we find them guilty |
|
|
Term
| alpha is our accepted type __ error rate |
|
Definition
|
|
Term
| If (u) is not equal to our Ho, the t statistic is >/The error of 0.05 is distributed how? |
|
Definition
the t statistic is either much larger or much lower than the mean
the 0.05 is split on both tails, with 0.025 on each side |
|
|
Term
| If u>x and Ha is accepted, the test statistic is >/ |
|
Definition
test statistic is > the mean.
0.05 is all on the top tail. |
|
|
Term
tails: critical/noncritical regions
central: critical/noncritical regions |
|
Definition
tails: critical
central: non-critical |
|
|
Term
| If the calculated t statistic is less than or equal to -2.06, OR greater than or equal to 2.06, it falls in the ____ and we will reject the null hypothesis |
|
Definition
|
|
Term
| If the calculated t statistic is greater than -2.06 AND less than 2.06, it falls in the ____ and we will fail to reject the null hypothesis |
|
Definition
|
|
Term
| Two reasons for a two mean comparison: |
|
Definition
test for equality
test for magnitude of difference |
|
|
Term
| how do you differentiate sample means from two populations? |
|
Definition
|
|
Term
| if the test statistic falls in the outer tails, we say it is part/not part of the distribution |
|
Definition
|
|
Term
| In a two mean scenario, if the t hat comes out positive, it means pop1 or pop2 is larger? |
|
Definition
pop1
because part of the equation is xbar1-xbar2 |
|
|
Term
| What is the point of an ANOVA? |
|
Definition
| to see whether or not at least one sample population mean is different |
|
|
Term
What is the simplest form of ANOVA?
What is the difference between this and more complex form? |
|
Definition
One way ANOVA
Used to test 3+ treatments
This has only 1 variation source (factor) |
|
|
Term
|
Definition
| mean of all the means of the treatments |
|
|
Term
| what is the treatment effect? |
|
Definition
| difference between that treatment mean and the grand mean |
|
|
Term
| what is the residual error? |
|
Definition
| difference between the individual measurement and the mean of the sample population |
|
|
Term
| Draw a diagram of treatments A,B,and C with a grand mean of 5, and the null hypothesis is true |
|
Definition
| I would probably draw A B C on the x axis and numbers on the Y. I'd make a dotted line at 5 all the way across to show the grand mean. Then I'd have all three treatment means on the dotted line. |
|
|
Term
| Draw a diagram of treatments A,B,and C with a grand mean of 5. The treatment means are all different but the within variation is small. Would we most likely reject or fail to reject the null? |
|
Definition
I'd have the points on both sides of the grand mean with small error bars.
We'd most likely reject the null. |
|
|
Term
| Draw a diagram of treatments A,B,and C with a grand mean of 5. The treatment means are all different but the within variation is large. Would we most likely reject or fail to reject the null? |
|
Definition
| Id have the means on both sides of the grand mean, with large error bars. We'd most likely fail to reject the null because the variation overlaps so much that they wouldn't truly differ. |
|
|
Term
Not sure if we have to know this buuuuttt.....
When inferences only apply to the treatments you studied in your 1 way model it is called _____ |
|
Definition
| A fixed-effects model or model 1 |
|
|
