Flashcards
Lecture 8
Flashcards do conjunto (41)
- 1Finish these sentences... In a sense, A single factor ANOVA is an extension of ... A repeated measures ANOVA is an extension of...1. t-tests
2. Paired samples t-tests - 2RECAP: What does ANOVA allow you to do?Analyse more than two groups at once
- 3Recap: what are the defining characteristics of an independent groups design?Each participant in involved in only one level of the IV.
- 4RECAP: What are the two sources of variability within groups?Error Term
Individual differense + experimental error - 5RECAP: What are the three sources of variability between groups?Treatment + Individual differences + Experimental error
- 6Where does the F ratio come from?Between group variability
______________________
Within group varaibility - 7if the F ratio is strong, will the F ratio be bigger or smaller?Bigger
- 8What do we have to make clear in the null?Any difference we see is due to sampling error
- 9What does the sums of squares represent?The variability in the data
- 10What are we asking about variability? What does ANOVA attempt to do?Determine how much variability comes from individual differences, experimental error, and/or the IV
- 11What do post hoc tests do?Break everything down into possible pairs (pairwise comparisons) and determine if there's a significant difference anywhere among the pairs.
- 12When is ANOVA suitable?This procedure is suitable if your experiment is essentially exploratory (i.e., you might have some idea of what to expect, but you basically want to explore the data generally).
Example: the literature shows mixed results or there isn't a large bank of research, therefore your aim reflects this exploratory nature and you want to make sure you cover all bases. - 13What approach do you use if you have very clear, set ideas about what is going to happen or established hypotheses?You use planned comparisons (also called planned contrasts).
(A priori tests) - 14What are a priori tests?Means beforehand.
Tests that you decide before the experiment. Once you have decided to use these a priori tests, you are limited to their use only. - 15What are planned comparisons?Planned comparisons are specific hypothesis tests involving specific sub-groups of your experiment, and are carried out INSTEAD of overall ANOVA followed by post-hoc tests
- 16What's the condition with panned comparisons?They are carried out INSTEAD of ANOVA and post-hoc tests.
Can't do both! - 17Provide three examples of planned contrasts (1)Contrast 1: The Method of Loci group (4) will recall more words than the other three groups overall (1, 2, & 3).
*Doesn't compare MOL with each, compares MOL with the average of the other three conditions. - 18Provide three examples of planned contrasts (2)Contrast 2: The Method of Loci (4) group will recall more words than the imagery group (3).
- 19Provide three examples of planned contrasts (3)Contrast 3: The two cognitive groups (3 & 4) will perform better than the drug group (2).
- 20What's an advantage of planned comparisons over ANOVA followed by post-hoc tests?Planned comparisons allow greater flexibility, more specificity, and more complex comparisons. For example, rather than just making pairwise comparisons, the researcher can compare one group with the average of a number of groups.
- 21What are the advantages of planned comparisons over post-hoc tests:1. For the same comparison, you are more likely to find a significant difference with a planned comparison than with an equivalent post-hoc test (i.e. they are statistically more powerful).
2. Can design complex comparisons (that include combinations of groups). - 22What is the advantage of post-hoc tests over planned comparisons?Let you compare everything with everything else and therefore obtain unexpected results if they're there (planned comparisons permit a smaller number of focussed comparisons).
- 23What are the rules for planned comparisons?Rules for planned comparisons:
1. You do them instead of an overall ANOVA (not as well as).
2. The maximum number of planned comparisons you can have is the number of groups - 1.
3. Your contrast coefficients must sum to zero (has to do with the way you set up comparisons). - 24Why are you restricted to a maximum number of planned comparisons (which is the numbers of groups - 1)?Related to orthogonality (which means independence). Planned comparisons should, in theory, tap into independent non-overlapping information.
They should, statistically, derive independent information. a simple way of achieving this is to limit the number of comparisons carried out. - 25How do you 'set up' planned comparisons?Using contrast coefficients.
- 26What are contrast coefficients?Numerical indicators that tell the program what's being compared with what else?
- 27How do contrast coefficients work?One needs to be positive and one needs to be negative so they sum to zero.
- 28Describe the steps to 'setup' contrast coefficients1. Write down the names of the levels of the IV as columns (the order in which you do it is the order in which you coded these in your data file as numbers L:56.26)
2. Planned contrast are in rows - 29What are the rules for setting up planned contrasts1. Must sum to zero
2. If a group is not involved in a contrast it gets a coefficient of 0 as it's not part of the comparison
3. The order of negatives and positive are arbitrary. the value is arbitrary e.g., could use 1 or 2.
4. When setting up combinations, the groups in the combination are assigned the same value. - 30How do you set up planned comparisons in SPSS?1. Analyse
2. compare means
3. One-way ANOVA
4. Add dependent and independent variable
5. Contrasts
6. Each numeral and add (as indicated in the table)
7. Next (repeat) 'Contrast 2 of 2'
8. For the final one, you press continue not NEXT. - 31How can you check that you are entering the coefficient correctly in SPSS?Coefficent Total will = 0.000
- 32Why do we need to check the homogeneity of variance when conducting planned comparisons?So we read the correct result
- 33How do you read planned comparisons?In the same way you read t-tests
- 34How are t statistics and f ratios related?If you take the square of a t score t, you'll get the equivalent F ratio as they are mathematically related to each other.
- 35How do you know which condition is superior when reading the results of a t-test, planned comparison etc?Reading means and looking at the graphs
Example: comparison one where MOL is significantly better than the other three methods combined. - 36What determines whether a t-score is negative or positiveThe sign that was given to the contrast coefficients, the sign of t-score is irrelevant.
- 37Do all programs report planned comparisons as a t-score?No, some report them as an F ratio as the two are related.
- 38How would you write up the results of the planned comparisons conducted by John in the lecture? L: 1.19The main questions of interest were tested using three a priori planned comparisons. Because the homogeneity of variance assumption was met, p = .49, the analyses based on equal variance are reported. A significant difference was found between the Method of Loci condition and the other three groups combined, t(96) = 2.78, p = .006; however, no significant difference was found between the Drug condition and the two cognitive methods in combination, t(96) = 1.78, p = .08, or between the two cognitive-based methods (Imagery and Method of Loci), t(96) = 1.35, p = .18.
- 39How is 'a priori' written?In italics
- 40Are you more likely to get a significant result for the same test using a post hoc approach or planned comparison?Planned comparison as stats rewards you for the a priori approach with more statistical power.
- 41How would you get the f score equivalent of a t score?t-score x t-score
Informações
- Descrição
- FlashCards sobre Lecture 8, criado por Samar Castelli em 02-08-2018.
- Criado por
- Samar Castelli · ago 2018
- Atualizado
- 6 ago 2018
- Visualizações
- 1
- Avaliação
Mais sobre Planned Comparisons and Post Hoc Tests