30 Cards in this Set
| Front | Back |
|---|---|
|
(What is?) Pooled Variance
|
The weighted averages of the variances from two groups which places more influence or importance (i.e. weight) to the group that has the largest sample size (is what?)
|
|
Why do we use Pooled Variance?
|
We use it because we get a more accurate estimate of natural variability by averaging the estimates from both of our groups
|
|
Imagine that you conducted an Independent Samples t-Test, and found the following data for the two groups in your experiment. Group 1: , n1 = 100. Group 2: , n1 = 10. When you calculate pooled variance (), which value will it be closer to? 65, 100, 30
|
100. Because group 1 has the largest sample size (n=100), it will receive more weight or importance in the calculation of the pooled variance. Thus, the resulting score will be closest to 100.
|
|
What are the two ways in which all t scores can be interpreted in two ways?
|
Whether from a single, related, or independent samples t-Test, all t- scores can interpreted as a location on the sampling distribution and as a ratio of effect to natural variability.
|
|
Using t=4.00 as an example, how can all t- scores be interpreted in two ways?
|
A t-score of 4.00 tells us that the evidence we got in our experiment fell 4 standard errors above the population mean (the mean we expected) on the sampling distribution. Also, a t-score of 4.00 tells us that our effect was 4 times greater than natural variability.
|
|
What is the main difference between Independent Samples and Related Samples?
|
Independent is characterized by to groups that are not related in any important way while Related is characterized by two samples that are related (or correlated).
|
|
How can people be "related" in a related sample?
|
Same people in each condition or different people with something in common
|
|
What does it mean by not related in an independent study?
|
There are different conditions that exist (i.e. experiment vs control) or different people in each condition.
|
|
What type of research project would we have for an one sample t-Test?
|
Comparing scores from a single sample to a hypothetical standard represented by the population mean (mue).
|
|
How would the example "number of new words recalled during a memory test" be used in an one sample?
|
We would show a group of randomly chosen students a list of new words and then record how many they remembered. We might then compare their recall to a score of 75% correct, because we think that would indicate signifiant learning.
|
|
What type of research project would we have for independent samples?
|
Comparing the sample means (or scores) from two entirely unrelated groups to see if the means are different. One of the groups is usually a control group which did not receive an experimental treatment.
|
|
How would the example "number of new words recalled during a memory test" be used in an independent sample?
|
We might have one group see a list of words, then ask the participants to recall the words (control group). Then we might have the other group recall the words after eating a peppermint candy, which is supposed to boost recall (experimental group). If any differences between means emerges…
|
|
What type of research would we have in a related sample?
|
Comparing two sample means taken from a) the same individuals, who were tested twice or b) two groups of individuals who are somehow related (twins, husband and wife, members of republican party).
|
|
How would the example "number of new words recalled during a memory test" be used in a related sample?
|
We could do the same study as described in the independent sample, except use twins; one twin in the control group and the other in the experimental group.
|
|
What is the main advantage of using a related samples design?
|
To increase the Power of the t-Test
|
|
Why would a related samples design usually give us more power to detect that effect than an Independent samples design?
|
Because there is a true effect to detect (Null is false)
|
|
What is the main disadvantage to a related samples design?
|
The concern about "carry over" and "order effects". In other words, sometimes the treatment given in one condition can have a long term effect, which then makes it impossible to observe behavior under normal circumstances.
|
|
Using the peppermint/ recall experiment example & suppose peppermint really does improve long term memory., how can the concern about carry over and order effects influence results?
|
If we start our experiment with the experimental condition (with candy) and follow that with the control condition (without candy), then the control condition will be contaminated (lasting effects from the peppermint can affect performance in our control condition). Hence, related samples…
|
|
True or False: We can usually detect the effect with a much smaller sample size, so we dont need as many participants
|
True
|
|
In an independent sample design, what serves as our representation of natural circumstances?
|
The control condition serves as our representation or approximation of natural circumstances. This is because it is a condition in which we do not intentionally influence behavior (no experimental manipulation or independent variable)
|
|
In a related sample design, what serves as our representation of natural circumstances?
|
The same individual (or matched individuals such as twins) serve as their own baselines (or controls) for comparison. The assumption is that before we apply our independent variable, behavior is caused only by natural variability (the control condition in an independent sample design) and…
|
|
Using the fundamental form of the T- Test, explain how a related samples design increases power.
|
Related sample design lets us remove some of the natural variability from our dataset by comparing the same individuals to themselves. We can cancel out any source of natural variability (i.e. personality) when we compute the difference score. This happens because individuals serve as the…
|
|
How causes the decrease in standard error in a related samples t- test?
|
We have removed a source of natural variability from our dataset.
|
|
You are interested in studying whether gender affects relationship satisfaction for incoming college freshman. You collect data from a sample of men and a sample of women at the end of the fall semester and compare their avg scores. What sample will you use?
|
Independent
|
|
You are interested in whether college student's excitement over college basketball changes over the course of the season. You collect data from UL students before and after the season to answer your research question. What sample will you use?
|
Related
|
|
What is a possible interpretation of a p value= .05
|
5% of the time I will reject the Null Hypo when it is in fact true
|
|
When comparing the sampling distribution of means (null hypo) and a sampling distribution for alt hypo, what does a large effect mean?
|
The sample mean is much larger than the population mean in terms of standard deviations
|
|
If we increase sample size, the central limit theorem tells us that the standard error will decrease. What does a decrease in standard error mean?
|
It means we theoretically have a more accurate measure of the population
|
|
What is the advantage of using pooled variance?
|
It is a more accurate estimate of the variance across both samples.
|
|
What is Power?
|
The probability of finding a significant difference if the effect that you are looking for is real.
|
PSYC 301: EXAM 2