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;
65 Cards in this Set
- Front
- Back
T or F? Chi Square Tests are parametric tests.
|
False, Chi Square tests are nonparametric tests.
|
|
Chi Square Tests do not assume: (2 things)
|
1. interval level of measurement
2. Data are normally distributed |
|
What are the 2 types of Chi Square tests?
|
1-way and 2-way
|
|
What is the purpose of a 1-way Chi Square test?
|
To compare an observed distribution with a hypothesized one. "Goodness-of-fit test"
|
|
The Chi Square "Goodness-of-fit test" allows you to make this comparison by comparing the _____ frequencies in your sample to the _____ frequencies that you would expect based on the population percentages.
|
Observed, Expected
|
|
What is the purpose of a 2-way Chi Square test?
|
To see whether two variables are independent of each other or are associated.
|
|
No relationship in a 2-way Chi Square test looks like what percentages in the table?
|
80% 80%
20% 20% |
|
The 2-way Chi Square test of independence tells you what?
|
Whether there is an association (relationship) between type of therapy and outcome.
|
|
In a 2-way Chi Square test of independence, you could say: did one therapy work better than the other? If not, then type of therapy and outcome are?
|
Independent
|
|
Sit. where you can use a Chi Square Tests 1 of 5:
Both types of Chi Square tests are based on the difference between _____ frequencies and _____ frequencies. |
Observed, Expected
|
|
In a 2-way Chi Square Test, what type of frequencies are expected if there is no relationship between the variables?
|
Expected frequencies
|
|
Sit. where you can use a Chi Square Tests 2 of 5:
Chi Square Test are used with data in the form of ____ (not scores). |
Counts
|
|
Chi Square Tests can be any one of the following types of frequency data? (3 types)
|
Frequency data (f)
Proportions (f/N) Percentages (proportion X 100) |
|
What do you do with proportions and percentages in order to see whether the observed frequencies differ from the expected frequencies?
|
Convert them to frequency data
|
|
Frequency data conversions:
|
.5 of 100 participants = 50
35% of 100 participants = 35/100 x 100 = 35 |
|
Sit. where you can use a Chi Square Tests 3 of 5:
The IV in a 1-way and the 2 IV's in a 2-way Chi Square Test must be ____ categories. |
Discrete
|
|
In Chi Square Tests, the data must be _____ and the ______ must not overlap.
|
Categorical, categories
|
|
Sit. where you can use a Chi Square Tests 4 of 5:
Can have a 1-way or a 2-way design. Each IV can have ___ or more levels (i.e., categories) |
2
Ex. Religion (Prot., Cath., Jew., Other) by Ethnicity (Afr-Am., Asian, Latino, Other) |
|
Sit. where you can use a Chi Square Tests 5 of 5:
A participant can only be in ___ cell of the table. |
1. A participant can not be in more than one group, he can only be in the 2-way table only once.
|
|
Chi Square Goodness-of-Fit Test (1-way) is appropriate when there is one _____ IV with __ or more levels and the DV is a _____.
|
Nominal, 2, Count
|
|
The purpose of a Chi Square Goodness-of-Fit Test is to compare an _______ distribution to a ______ distribution or some previously specified distribution.
|
Observed, Theoretical
|
|
The test statistic that is compared with the table value is the _____ _____ _____ and it provides a measure of how much the observed and expected frequencies ______.
|
Chi Squared Statistic, Differ
|
|
In a Chi Squared Test, when we compute X-squared and if it is larger than the critical value with (____)df, we _____ the Ho that there is no difference between the observed and expected frequencies.
|
(k - 1)df; reject the Ho
|
|
Rejecting a Chi Squared Test Ho, means that the _____ data did not come from a ______ with the hypothesized distribution.
|
Sample; Population
|
|
Chi Squared Test Example:
|
See p.3 of Chi Squared Test handout
|
|
When you use the Chi Squared Test, you must have expected frequency for each and every category of:
|
5 for df > or = 2
10 for df = 1 |
|
If you don't have at least this many, you can do one of two things:
|
1. Collapse your groups together to increase the expected frequencies (have few categories). Do this only if the categories created make sense.
2. Take a larger sample. |
|
In a Chi Squared Test, when there is only 1 df, ____ correction for ____ should be made.
|
Yates; continuity
|
|
For a Chi Squared Test of Independence (2-way), you must have ___ _______ IV's, each with __ or more levels, and a DV that is a _______ variable.
|
2, Nominal; 2; Frequency
|
|
What is a table called with 2 nominal IV's with frequencies in the cells?
|
Contingency Tablle
|
|
What is the purpose of a Chi Squared Test of Independence?
|
To test whether the 2 IV's are independent of one another.
|
|
In a Chi Squared Test of Independence, if outcome is independent of type of therapy, you conclude:
|
Outcome is related to type of therapy.
|
|
Hypothesis testing: in a Chi Squared Test of Independence, generally in the Ho variables A and B are _______ in the population and H1 variables A and B are _______ in the population.
|
Independent; Related
|
|
To calculate the value of X-squared, you use the same formula as in 1-way. Expected frequencies are calculated differently.
|
See Handout #1 p.5 for E sub rc equation
|
|
E sub rc equation notation:
|
E sub rc = expected frequency for the cell in row r and cell c
f sub r = frequency for row r f sub c = frequency for column c N = total # of observations |
|
The expected frequencies are the cell frequencies that would be observed, given the row and column frequencies, if variables A and B were ______.
|
Independent (i.e., not related)
|
|
What are you really calculating with a Chi Squared Test of Independence?
|
How much the observed frequencies differ from the frequencies that would be expected if the 2 IV's were independent.
|
|
Conceptually, what is important to notice with a Chi Squared Test of Independence?
|
If the 2 IV's are independent, approximately the same percentage of participants receiving each type of therapy are able to get on the plane.
|
|
With a Chi Squared Test of Independence if there weren't equal n's in the 2 therapy groups, it would be more useful for purposes of interpretation to look at _____ rather than _____.
|
Percentages; Frequencies
|
|
When can you not use the Chi Squared Test if Independence?
|
If you find any expected frequency < 5 with df > or = 2, or, < 10 with df = 1
|
|
If you have a 2 x 2 contingency table, use ____ ____ Test.
|
Fisher's Exact Test
|
|
When do you need to use Collapsing Levels of a Variable?
|
If it is a larger table, you can use it to increase sample size.
Ex. Can collapse occupation #1 with other occupations if it makes sense to do so. |
|
Parametric tests test hypotheses about _____.
|
Parameters
|
|
Parametric tests involve making the three following assumptions and differ from nonparametric tests assumptions in what way?
|
1. Interval level data
2. Normally distributed 3. Equal variances on the DV in all groups Nonparametric tests do not assume making any of these assumptions. |
|
The two nonparametric tests that we have already covered are 1. ____ and 2. _____, and they are used to examine whether there is a ______ or _____ between two variables.
|
Spearman rank correlation coefficient,
Chi Square Test of Independence; Relationship, Association |
|
An alternative to the independent groups t test that tests whether the central tendency is the same in two groups is:
|
Wilcoxon t test for Two Independent Groups (Mann-Whitney U)
|
|
How does the data in the Wilcoxon t test for Two Independent Groups differ from the independent groups t test?
|
Independent groups t test data was the scores, Wilcoxon t test for Two Independent Groups (Mann-Whitney U) data are ranks.
|
|
Steps for using the Wilcoxon t test for Two Independent Groups (Mann-Whitney U) test:
|
1. Convert scores to ranks
2. Write the scores out in order 3. Rank from lowest to highest 4. The computer tells you whether the groups are significantly different in CT and gives you average rank for each group. 5. If the groups are different in CT, the average ranks will be different. |
|
When to use the Wilcoxon t test for Two Independent Groups (Mann-Whitney U)?
|
If the assumptions for the t test are not met, i.e., data are not interval level, but ordinal, data are not normally distributed; or variances are unequal.
|
|
If the t test assumptions are met, why is it better to use a t test than the Wilcoxon t test for Two Independent Groups (Mann-Whitney U)?
|
The t test has more power.
|
|
There is another test to determine whether two tests have the same median. It is called the _____ _____ and is based on the number of observations ______ and ______ the median.
|
Median Test; above and below
|
|
What is the test called that is analogous to the dependent samples t test?
|
Wilcoxon t test for Dependent Samples
|
|
When do you use the Wilcoxon t test for Dependent Samples?
|
It is for comparing the means under 2 treatments. Use when the same participant is observed in 2 treatment conditions (or when you have matched pairs of participants)
|
|
In the Wilcoxon t test for Dependent Samples, how do you conclude that the 2 populations are different?
|
When the sum of the positive ranks is not equal to the sum ofthe negative ranks.
|
|
Steps to compute the Wilcoxon t test for Dependent Samples:
|
1. Compute the differences, say pre and post
2. Rank order the differences without regard to sign (remember to average ranks of the same scores) 3. Attach sign of the difference 4. If no difference between the distributions in the two populations (i.e., pre and post), then the sum of the positive ranks should equal the sum of the negative ranks. 5. Conclude there is a difference when, for example, the sum of the positive ranks is much greater than the sum of the negative ranks. |
|
What test should you use to test for paired (correlated) dichotomous variables?
|
The McNemar Test
Ex. Success/failure on 2 tests; is there a difference between test 1 and test 2? Participants take both tests. |
|
What is the test that is a nonpaframetric alternative to 1-way ANOVA?
|
Kruskal-Wallis 1-way ANOVA by Ranks
|
|
What does the Kruskal-Wallis 1-way ANOVA by Ranks assume?
|
All data are ordinal (can also be interval).
|
|
Steps to run the Kruskal-Wallis 1-way ANOVA by Ranks test:
|
1. All scores are rank ordered regardless of group beginning with the lowest score.
2. Then the ranks are summed for each group. |
|
In the Kruskal-Wallis 1-way ANOVA by Ranks test, if the results are significant, we conclude what?
|
That the groups differ in their average scores on the DV.
|
|
You can use the Kruskal-Wallis 1-way ANOVA by Ranks test when: (3 things)
|
1. There is one nominal IV with 2 or more levels
2. The participant is in only one group. 3. The number of people in each group is at least 6. |
|
When we get a significant result in the Kruskal-Wallis 1-way ANOVA by Ranks test, we need to use what to see which specific pairs of groups were significantly different on the DV?
|
Pairwise Post Hoc Comparisons to apply the multiple comparison procedure
|
|
Pairwise Post Hoc Comparisons applies the multiple comparison procedure to:
|
Mean Ranks
(Tukey's HSD procedure compares all possible pairs of means.) |
|
In Pairwise Post Hoc Comparisons applies the multiple comparison procedure comparing mean ranks, what do we need to have?
|
Equal n's in the groups
|
|
What is the equation for the mean rank for one group?
|
The sum of the ranks for the people in that group/the number of people in that group
|