click below
click below
Normal Size Small Size show me how
STATS module 3
stats midterm
Question | Answer |
---|---|
NHST is a method for deciding a difference likely exists, but does not speak to the ________ of that difference. | size |
In NHST, what value is associated with the t-test as an index of magnitude? | p-value. probability value |
Experiment 1: independent samples t-test p = .001 vs. Experiment 2: independent samples t-test p = .01. What conclusion can be drawn from this information? | The chance explanation is less viable in Experiment 1 than Experiment 2 |
What two things influence the p-value as a result of their influence on the t-value? | 1. size of mean difference 2. sample size |
Is concluding the null hypothesis is very unlikely the same as concluding the difference (magnitude) is large? Why? | No, With large samples, high precision could lead us to conclude a modest difference is unlikely a result of chance factors |
What is the key principle of Bayesian Statistics? | for our data we calculate a Bayes factor (ratio of the likelihood of the alternative hypothesis relative to the likelihood of the null hypothesis) |
What does a Bayes factor of 1 mean? | equal likelihood of alternative relative to null |
What does a Bayes factor less than 1 mean? | the null hypothesis is more likely than the alternative |
What does a Bayes factor greater than 1 mean? | the alternative hypothesis is more likely than the null |
What is the Bayes factor threshold for moderate evidence that the alternative hypothesis is more likely than the null? | 3 |
What is the Bayes factor threshold for strong evidence that the alternative hypothesis is more likely than the null? | 10 |
What is raw effect size? | for example: the size of the difference between the two means (mean 1 - mean 2), or unstandardized regression coefficients |
When might it be helpful to use raw effect size? | when the outcome variable of interest (the dependent variable) is on a metric that is meaningful and readily interpretable in light of some clear criteria |
What are two situations in which using raw effect size would be problematic? | 1. The outcome variable is not easily interpretable with respect to specifiable criteria. 2. One needs to compare effects with outcome variables that are on different metrics |
Generally, why do we use standardized effect size indices? | These indices can be applied to measures of different metrics and express magnitude of effects in common metric regardless of original scaling. |
What are two commonly used indices often associated with independent samples and repeated measures t-tests? | 1. Cohen's d 2. Pearson's r |
Generally, what is Cohen's d and what does it express? | One of the most widely used effect size indices, expresses magnitude as a standard difference between means |
What is Cohen's d for independent samples? | the mean difference divided by the pooled standard deviation of the two samples |
What subscript represents the Cohenβs d for an independent samples t-test? | s |
How is the Cohen's d influenced by sample size? | it is not |
ππ increases as the mean difference __________ and the standard deviations __________ | ππ increases as the mean difference increases and the standard deviations decrease |
What are the minimum and maximum values of ds? | min = 0, no max |
How can a ds value of .5 be interpreted? | indicates the difference between the means is half the size of the dependent variableβs standard deviation |
What ds value indicates the difference between the means is as big as the standard deviation of the dependent variable? | ds = 1.00 |
How can a ds value of 2.00 be interpreted? | indicates a mean difference twice the size of the standard deviation of the dependent variable |
What are Cohen's guidelines for interpreting a ds value? | 0.2 (small), 0.5 (medium), and 0.8 (large) |
What subscript represents the Cohenβs d for a repeated measures t-test? | av |
What is Cohen's d for repeated measures t-test? | dav = (D bar)/ Avg.S |
What is Avg.S? | (S1+S2)/2 aka the mean of the standard deviations |
What is the key difference between dav and drm? | πππ£ ignores the magnitude of correlation between sets of observations while drm takes it into acount |
When __________ in both sets of observations are equal, πππ£ and πππ are equal. | standard deviations |
πππ£ will tend to be more similar to ds than πππ except when? | r is low and the difference between standard deviations are large |
πππ is _______ conservative than πππ£ but is considered overly conservative when r is _______ | πππ is more conservative than πππ£ but is considered overly conservative when r is large |
Is Cohen's d positively or negatively biased? | Cohenβs d is a positively biased estimate of the population effect size |
What effect size measure is often used as a more conservative estimate of effect size? | Hedges g |
For what situations can you use Pearsons R and what does it express? | 1. the strength and direction of association between two continuous variables 2. the strength of association between a binary categorical variable and a continuous variable |
When squared, what does Pearson's r express? | the proportion of variance in the dependent variable accounted for by group membership |
What is the range of r values? | from -1.00 to 1.00 with .00 indicating no association |
What are Cohen's guidelines for interpreting an r value? | .10 (small), .30 (medium), and .50 (large) |
What is the relationship between effect size and power? | it is necessary to make assumptions regarding the effect size when doing power calculations |
Large effect sizes do not directly imply practical significance. What other factors are important to consider? | 1. Metric can be hard to interpret without reference to more concrete reference criteria. 2. Durability of an effect might also be relevant in addition to its size. 3. Cost/benefit analysis also can determine practicality. |
What are some examples of sample values for which we can calculate the associated standard error? | 1. sample means 2. raw effect size 3. standardized effect size indices |
What is the common specified confidence interval? | 95% |
What does a confidence interval indicate? | that there is a 95% chance that the interval you have calculated contains the population |
What determines the width of a confidence interval? | standard errors |
What two factors influence standard error? | 1. sample size 2. variability around the means |
What does a confidence interval containing 0 indicate? | the effect is not statistically significant |
What do CIs and effect sizes have in common with NHSTs? | 1. They rely on the same information and are affected by the same factors 2. They simply express this information in a different way |