Data week 7

winniesmith2's version from 2017-11-13 13:02

Section 1

Question Answer
Repeated measures designs (within-participants design)There are many situations in which we might want to study the same individual under a number of different conditions. This is because the research design involves taking repeated measures of the same individual(s) for different treatments or conditions.
Describe repeated measures ANOVA Repeated measures ANOVA is the equivalent of the one-way ANOVA, but for related, and not independent groups. It is an the extension of the paired t-test . A repeated measures ANOVA is also referred to as a within-subjects ANOVA. This week we will look at a One Way Repeated Measures ANOVA.
Advantages of repeated measures designsThe main advantage of repeated measures designs is increased statistical power. This is because the design removes the effects of individual differences. A second advantage is that fewer participants are needed for most experimental designs. This is a big advantage if data is difficult to obtain because we cannot find the participants for our experiment.
Disadvantages of repeated measures designs The main disadvantage of a repeated measures design is that the independent variable may be confounded with order of testing or carry over effects. Practice effects, fatigue effects, contrast effects, demand characteristics.
Practice effects subjects get better at the task over time because of practice, so that they perform best in the later conditions
Fatigue effects subjects get worse at a task over time because of fatigue, so they perform worse in the later conditions
Contrast effects a noisy condition experienced after a quiet condition might be perceived as even noisier than it normally would be.
Demand characteristics being in more than one condition makes it clear to subjects what the independent variable is. They may behave how they think you want them to in later conditions
Solutions to problems of Order Effects for repeated measures designs Randomising the order of testing and counterbalancing order of testing (ex. pg 10)
Another disadvantage of repeated measures designs; missing data -Repeated measures designs run into problems when there is missing data (e.g. one of the participants doesn’t turn up for the third condition of testing). -One solution is to exclude participants with missing data from the analysis (however losing participants means losing power). This is always a risk with a repeated measures design. -Other solutions involve estimating what values the missing data would take (multiple imputation) but this is far more complicated and beyond the scope of this module.

Section 2

Question Answer
1. How do the scores in one-condition vary from those in another? This is our between-groups variance.
2. How do participants vary in their average scores? How much does a participant’s mean score differ from the grand mean? This is a measure of participant variance.
3. How much error variance is there? for example p 13If participant 2 has faster recall scores than participant 1 for List A, we would expect that person to perform similarly for lists B and C. Any deviation from this can be considered experimental error.
Partitioning the variance- what is there not in repeated measures designthere is no between groups variation due to individual differences as the participants in all conditions are the same
How F is calculated in a repeated measures designF= π‘π‘’π‘‘π‘€π‘’π‘’π‘›βˆ’π‘”π‘Ÿπ‘œπ‘’π‘π‘  π‘£π‘Žπ‘Ÿπ‘–π‘Žπ‘›π‘π‘’ (π‘€π‘–π‘‘β„Ž π‘–π‘›π‘‘π‘–π‘£π‘–π‘‘π‘’π‘Žπ‘™ π‘‘π‘–π‘“π‘“π‘’π‘Ÿπ‘’π‘›π‘π‘’π‘  π‘Ÿπ‘’π‘šπ‘œπ‘£π‘’π‘‘) / π‘€π‘–π‘‘β„Žπ‘–π‘›βˆ’π‘”π‘Ÿπ‘œπ‘’π‘π‘  π‘£π‘Žπ‘Ÿπ‘–π‘Žπ‘›π‘π‘’ (π‘€π‘–π‘‘β„Ž π‘–π‘›π‘‘π‘–π‘£π‘–π‘‘π‘’π‘Žπ‘™ π‘‘π‘–π‘“π‘“π‘’π‘Ÿπ‘’π‘›π‘π‘’π‘  π‘Ÿπ‘’π‘šπ‘œπ‘£π‘’π‘‘). This formula for F takes into account that the participants are the same in each condition. Variation due to individual differences is removed from the top and bottom of the equation. This gives rise to a more sensitive and powerful statistical test (basically we have removed a big chunk of error variance by having the same participants in each group).
Assumption: Homogeneity of Variance - Sphericity This is a very complex assumption about the variances and covariances in the data set. SPSS will calculate sphericity for you. Sphericity means that the variances of the differences between all combinations of related groups (levels) are equal. Violation of sphericity is when the variances of the differences between all combinations of related groups are not equal. Sphericity replaces the homogeneity of variance assumption in independent measures ANOVA.
Homogeneity of variance -Variance is a measure how data are spread around the mean (variance ο‚Ί SD2). -Two or more variances between or within groups are compared: similar in size = homogeneous. -Use Levene's test of Homogeneity of Variance (between groups design) and Mauchly’s Test of Sphericity (within measures design).
Recap repeated measures ANOVA- what to do One Way Repeated measures ANOVA is the same as One Way Independent Measure ANOVA except for one thing. The same participants take part in all of the treatments/ conditions. This means that individual differences can be removed from the error variance. The error variance is reduced in size because individual differences are removed. This means that we have more chance of finding a significant treatment effect if one is present.