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;
178 Cards in this Set
- Front
- Back
|
Which is always possible, an experiment or an observational study? |
observational study |
|
|
If done correctly, which controls lurking variables, an experiment or an observational study? |
experiment |
|
|
The investigators control which subjects get what treatments in which one, an experiment or an observational study?
|
experiment |
|
|
Give an example in which an experiment changed the conclusions of an observational study. |
heart attacks in menopausal women with hormone replacement therapy |
|
|
Were the observational studies wrong that said women at menopause that had hormone replacement therapy had fewer heart attacks? |
no |
|
|
What was the lurking variable that in observational studies made it appear that hormone replace at menopause made women have fewer heart attacks?
|
how much the women cared about their health
|
|
|
Give a lurking variable other than wealth or smoking that may explain why it appear that drinking wine appears to be better than beer or hard liquor in observational studies. |
diet |
|
|
Give a lurking variable other than wealth or diet that may explain why it appear that drinking wine appears to be better than beer or hard liquor in observational studies. |
smoking |
|
|
Give a lurking variable other than diet or smoking that may explain why it appear that drinking wine appears to be better than beer or hard liquor in observational studies. |
wealth |
|
|
How could it be proven that wine is better than beer or hard liquor when it comes to health? |
experiment |
|
|
Suppose a large florist is deciding whether or not to accept a shipment of roses. The florist asks a recently hired employee to go into the truck where the shipment is and get a sample of 10 roses. What do you think this employee will do? |
get 10 nice roses |
|
|
Suppose a large florist is deciding whether or not to accept a shipment of roses. The florist asks a recently hired employee to go into the truck where the shipment is and get a sample of 10 roses. There is a good chance the employee will pick 10 nice roses. Would you be surprised if after accepting the shipment the florist is not happy with the overall quality of the roses? |
no |
|
|
Give a bias with Mall Sampling other than against tuff looking individuals and in favor or teenagers. |
in favor of retired people |
|
|
Give a bias with Mall Sampling other than against tuff looking individuals and in favor or retired people. |
in favor of teenagers |
|
|
Give a bias with Mall Sampling other than in favor of retired people and in favor or teenagers. |
against tuff looking individuals |
|
|
True or false: Getting a good sample is usually pretty easy to do. |
F |
|
|
The reasons people get bad samples is because of ignorance and ________. |
agenda |
|
|
The reasons people get bad samples is because of agenda and _________. |
ignorance |
|
|
Are volunteer response samples good? |
no |
|
|
Give an example of a volunteer response sample. |
Ann Landers asking her readers to write in about whether they wish they didn’t have children. |
|
|
The AFA (American Family Association) has online polls. Usually these polls will have what kind of bias? |
bias in favor of their points of view |
|
|
The AFA (American Family Association) got upset when an online poll about same sex marriage showed 2-1 in support of it. What happened? |
liberal groups ambushed it |
|
|
The ultimate way to sample is to get a what kind of sample? |
random |
|
|
How often are SRS’s possible? |
rarely |
|
|
Is it hard to get a bad sample? |
no |
|
|
Is it hard to get a good sample? |
yes |
|
|
Give an example of undercoverage. |
survey by phone – people without phones are not covered |
|
|
Give an example of nonresponse. |
survey in which not all surveys are completed |
|
|
What is the problem with undercoverage and nonresponse? |
the people not responding or not covered may be different in important ways |
|
|
Suppose a large city is deciding whether or not to use tax money to build a new stadium for its NFL football team. A newspaper is curious what the residents think and so they send out a mail questionnaire (that instantly makes it clear the NFL team is affected, but it’s not clear until you read closer that tax money is involved) to 10,000 addresses picked at random. Do you think all 10,000 questionnaires will be returned? |
no |
|
|
Suppose a large city is deciding whether or not to use tax money to build a new stadium for its NFL football team. A newspaper is curious what the residents think and so they send out a mail questionnaire (that instantly makes it clear the NFL team is affected, but it’s not clear until you read closer that tax money is involved) to 10,000 addresses picked at random. Do you think that even half will be returned? |
no |
|
|
Suppose a large city is deciding whether or not to use tax money to build a new stadium for its NFL football team. A newspaper is curious what the residents think and so they send out a mail questionnaire (that instantly makes it clear the NFL team is affected, but it’s not clear until you read closer that tax money is involved) to 10,000 addresses picked at random. Do you think people that would like a new stadium and those that do not will have the same rate of mailing the questionnaires back? |
no |
|
|
Suppose a large city is deciding whether or not to use tax money to build a new stadium for its NFL football team. A newspaper is curious what the residents think and so they send out a mail questionnaire (that instantly makes it clear the NFL team is affected, but it’s not clear until you read closer that tax money is involved) to 10,000 addresses picked at random. What sort of bias do you think will result if the newspaper relies only on the returned questionnaires? |
in favorof the stadium |
|
|
Suppose a large city is deciding whether or not to use tax money to build a new stadium for its NFL football team. A newspaper is curious what the residents think and so they send out a mail questionnaire (that instantly makes it clear the NFL team is affected, but it’s not clear until you read closer that tax money is involved) to 10,000 addresses picked at random. Should they put a story in their paper telling the residents what they think about the potential new stadium? |
no |
|
|
Does the wording of a question have much affect on the answers? |
yes |
|
|
Give an example in which the wording of a question could make quite a difference. |
do you think we are spending too much on “welfare” vs “help for the poor” |
|
|
Give an example of how a sensitive question might not give accurate results. |
do you use illegal drugs? |
|
|
Give an example of a question in which people are forgetful and the results may not be accurate. |
have you been to the dentist in the last 6 months? |
|
|
Give an example of a question asked by the wrong person that would make the results worthless. |
a uniformed police officer askingquestions about peoples’ opinion of the police departmentso |
|
|
Give an example of a question that begs a certain answer and hence the results can’t be trusted. |
has participating in our health program improved your health? |
|
|
We wish to perform an experiment to see whether an online version of a Stat course is better than an in class version. We have data from two teachers. Teacher A teaches an online class and the average grade point for the students in this class is 2.94. Teacher B teaches a regular class and the average grade point in this class was 2.33. So we conclude the online version is better. Three distinct problems are luck, Teacher A might be easier, and _______________. |
better students might be more likely to take the online class |
|
|
We wish to perform an experiment to see whether an online version of a Stat course is better than an in class version. We have data from two teachers. Teacher A teaches an online class and the average grade point for the students in this class is 2.94. Teacher B teaches a regular class and the average grade point in this class was 2.33. So we conclude the online version is better. Three distinct problems are luck, better students might be more likely to take the online class, and _______________. |
teacher A might be easier |
|
|
We wish to perform an experiment to see whether an online version of a Stat course is better than an in class version. We have data from two teachers. Teacher A teaches an online class and the average grade point for the students in this class is 2.94. Teacher B teaches a regular class and the average grade point in this class was 2.33. So we conclude the online version is better. Three distinct problems are better students might be more likely to take the online class, Teacher A might be easier, and _______________. |
luck |
|
|
What is a control group? |
a group that gets the standard treatment |
|
|
The three principals of experimental design are to divide subjects at random, use a control group, and ____________ |
use a lot of subjects |
|
|
The three principals of experimental design are to divide subjects at random, ________________, and use a lot of subjects. |
use a control group |
|
|
The three principals of experimental design are ______________, use a control group, and use a lot of subjects |
divide subjects at random |
|
|
What does statistically significant mean? |
rarely would occur by luck |
|
|
What is a placebo? |
dummy treatment that an individual is told is the same as an actual treatment |
|
|
What is the purpose of a placebo? |
remove the placebo effect, some people respond well to any treatment |
|
|
What is a double-blind experiment? |
neither the individuals nor those dealing with them directly know what treatments are given |
|
|
What is the purpose of a double-blind experiment? |
doctors giving a placebo might not be as positive |
|
|
Statistically significant depends on difference and _________. |
sample size |
|
|
Statistically significant depends on sample size and _________. |
difference |
|
|
Give an example of how lack of realism can cause problems in an experiment. |
TV commercials with a captive audience vs how normal people watch TV |
|
|
In a matched-pairs experiment if each person gets both treatments, why is it still important to divide the people up at random? |
order of treatments might affect the results |
|
|
Two advantages of a block design are to make luck less likely to affect things and ______________________. |
to see what does better in each subgroup |
|
|
Two advantages of a block design are to see what does better in each subgroup and ______________________. |
to make luck less likely to affect things |
|
|
Are the conditions usually met exactly when doing CIs or HTs? |
no |
|
|
Is it rare to see any problems when doing CIs or HTs? |
no |
|
|
If you have an outlier that is found to be a real piece of data, should you remove it? |
no |
|
|
To be a good statistician what should you do about not meeting conditions once the data is collected? |
decide if the lack of meeting conditions is minor or major |
|
|
If you have an outlier that is found to be an incorrect piece of data and can’t be corrected, the best thing to do is what? |
remove it
|
|
|
Do CIs and HTs remedy basic flaws in the data? |
no |
|
|
Give an example where a SRS is called for and not met, and this causes the results to be useless. |
adults with normal vision, but using students in a psychology class |
|
|
Give an example where a SRS is called for and not met, and this causes the results to be useless. |
interested in all adults view on welfare and using a sociology class at Harvard |
|
|
Give an example where there was a high statistical significance of something occurring, but it was not what people thought at first. |
gastric freezing on ulcers |
|
|
In the gastric freezing example, we were pretty sure patients were getting better, at first doctors thought it was ______________ , but later experiments showed it was probably just because of placebo affect? The problem was at first the gastric freezing experiment was not controlled. |
gastric freezing |
|
|
In the gastric freezing example, we were pretty sure patients were getting better, at first doctors thought it was gastric freezing , but later experiments showed it was probably just because of _________ affect? The problem was at first the gastric freezing experiment was not controlled. |
placebo |
|
|
=In the gastric freezing example, we were pretty sure patients were getting better, at first doctors thought it was gastric freezing , but later experiments showed it was probably just because of placebo affect? The problem was at first the gastric freezing experiment was not _________. |
controlled |
|
|
Do outliers have much affect on the HTs and CIs we do in the class? |
yes
|
|
|
Does the margin of error in a CI fix nonresponse? |
no |
|
|
Does the margin of error in a CI fix undercoverage? |
no |
|
|
Does the margin of error in a CI fix biased data? |
no |
|
|
If your sample is not a random sample, can you be 95% sure that the CI has the correct answer for the parameter? |
no
|
|
|
There is only one thing the margin of error in a CI covers, what it that? |
luck |
|
|
What does the p-value mean in cases where the sample you use for the HT has problems with it? |
nothing |
|
|
____________, how hard it would be to believe the result, and the consequences are three things that affect how small we would desire the p-value or the significance level to be. |
how important it is to get it right |
|
|
How important it is to get it right, ____________, and the consequences are three things that affect how small we would desire the p-value or the significance level to be. |
how hard it would be to believe the result |
|
|
How important it is to get it right, how hard it would be to believe the result, and ____________ are three things that affect how small we would desire the p-value or the significance level to be. |
the consequences |
|
|
Should we always use the 5% significance level? |
no |
|
|
If you have a small sample size, what will happen to the p-value if the same behavior is seen with a larger sample? |
go down |
|
|
If you have a small sample size and the p-value is too high, should you just give up on rejecting Ho? |
no
|
|
|
Does practically significant mean the same as statistically significant? |
no |
|
|
A small difference that nobody would care about in the real world, but we are really sure about is _____________ significant, but not practically significant? |
statistically |
|
|
A small difference that nobody would care about in the real world, but we are really sure about is statistically significant, but not _____________ significant? |
practically |
|
|
When doing HTs is it best to first look at the data you collect before deciding on Ho and Ha? |
no
|
|
|
Is it a good idea to do many different HTs tosearch for things that are true? |
no |
|
|
Why is it not a good idea to do many different HTs to search for things that are true? |
good chance just by luck you will reject a true Ho |
|
|
Is it a good idea to do repeat the same HTs with different sets of data? |
yes
|
|
|
Why is it a good idea to repeat the same HTs with different sets of data? |
the more times it is repeated the more sure you are |
|
|
How often we will be able to exactly meet the condition for CIs and HTs to be mathematically precise? |
rarely
|
|
|
Generally speaking there are there more concern with doing HTs and CIs with small sample sizes or large sample sizes? |
small
|
|
|
Name one problem with doing HTs and CIs with large sample sizes. |
small differences will be statisticallysignificant |
|
|
If your degrees of freedom are not in the table what should you do? |
round down |
|
|
If you reject an Ho assuming fewer degrees of freedom than you actually have, will you be able to reject Ho with the correct degrees of freedom? |
yes
|
|
|
If you reject an Ho assuming more degrees of freedom than you actually have, will you be able to reject Ho with the correct degrees of freedom? |
no
|
|
|
If you give a 95% CI assuming fewer degrees of freedom than you actually have, you should be ____________ than 95% sure you have the correct answer in the CI? |
more |
|
|
worse |
|
|
better
|
|
|
better |
|
|
You are comparing two means and your sample sizes are 5 and 8. The samples are random. There are no outliers but the shapes of the sample data are quite different. Do you think it is OK to do HT or CI? |
no |
|
|
You are comparing two means and your sample sizes are 50 and 80. The samples are random. There are no outliers but the shapes of the sample data are quite different. Do you think it is OK to do HT or CI? |
yes |
|
|
You are comparing two means and your sample sizes are 5 and 8. The samples are random. There are no outliers and the shapes of the sample data are very close. Do you think it is OK to do HT or CI? |
yes |
|
|
You are comparing two means and your sample sizes are 50 and 80. The samples are random. There are no outliers and the shapes of the sample data are very close. Do you think it is OK to do HT or CI? |
yes |
|
|
You are comparing two means and your sample sizes are 5 and 8. The samples are random. There are two minor outliers and the shapes of the sample data are very close. Do you think it is OK to do HT or CI? |
no |
|
|
You are comparing two means and your sample sizes are 50 and 80. The samples are random. There are two minor outliers and the shapes of the sample data are very close. Do you think it is OK to do HT or CI? |
yes |
|
|
You are comparing two means and your sample sizes are 5 and 8. The samples are random. There are two minor outliers and the shapes of the sample data are quite different. Do you think it is OK to do HT or CI? |
no
|
|
|
You are comparing two means and your sample sizes are 50 and 80. The samples are random. There are two minor outliers and the shapes of the sample data are quite different. Do you think it is OK to do HT or CI? |
yes
|
|
|
You are studying a mean and have a sample of size 10. The sample data is symmetric with no outliers and the data was collected at random. Do you think it is OK to do HT or CI? |
yes |
|
|
You are studying a mean and have a sample of size 10. The sample data is not symmetric and there are no outliers and the data was collected at random. Do you think it is OK to do HT or CI? |
no |
|
|
You are studying a mean and have a sample of size 10. The sample data is symmetric with a minor outlier and the data was collected at random. Do you think it is OK to do HT or CI? |
no |
|
|
You are studying a mean and have a sample of size 100. The sample data is symmetric with no outliers and the data was collected at random. Do you think it is OK to do HT or CI? |
yes
|
|
|
You are studying a mean and have a sample of size 100. The sample data is not symmetric with no outliers and the data was collected at random. Do you think it is OK to do HT or CI? |
yes
|
|
|
You are studying a mean and have a sample of size 100. The sample data is symmetric with a minor outlier and the data was collected at random. Do you think it is OK to do HT or CI? |
yes |
|
|
You are studying the mean heights of all adult men and have a sample of size 1200. The sample data is all major league baseball players and it is symmetric with no outliers. Do you think it is OK to do HT or CI? |
no |
|
|
You are studying the mean number of gallons of milk sold per day by a store and your sample is 60 days all in a row. It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
no |
|
|
You are studying the mean number of gallons of milk sold per day by a store and your sample is 30 days starting with one day and picking every 7th day after that. It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
no |
|
|
You are studying the mean number of gallons of milk sold per day by a store and your sample is 30 days starting with one day and picking every 12th day after that. It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
yes |
|
|
You are studying the mean drying time of paint on 2x4’s sold by a home improvement store and your sample is 40 boards all from the same shipment and the wood is pretty much the same from shipment to shipment. It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
yes |
|
|
You are studying the mean drying time of paint on 2x4’s sold by a home improvement store and your sample is 40 boards all from the same shipment and the wood tends to vary quite a bit from shipment to shipment. It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
no |
|
|
You are studying the mean drying time of paint on 2x4’s sold by a home improvement store and your sample is 40 boards in which you choose 10 shipments spaced out over several months and then chose 4 boards from each at shipment (one off the top, two from the middle, and one off the bottom). It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
yes |
|
|
You are studying the mean drying time of paint on 2x4’s sold by a home improvement store and your sample is 240 boards all from the same shipment and the wood tends to vary quite a bit from shipment to shipment. It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
no
|
|
|
You are studying the percents of cats that prefer two different types of cat food and your sample is 42 cats that were basically all the cats of all the people you know well that would participate. It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
yes
|
|
|
You are studying the percents of people that prefer two different types of beer and your sample is 42 prisoners in county jail. It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
no |
|
|
You are studying the difference in average weights of boy 4th graders and girl 4th graders and your samples are all 52 4th grade boys from a school in Mississippi and all 32 4th grade girls from a school in Colorado. It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
no |
|
|
You are studying the difference in average weights of boy 4th graders and girl 4th graders and your samples are all 52 4th grade boys from a school in Mississippi and all 32 4th grade girls from the same school. It’s not a SRS, but do you think it would still be OK to do a HT or CI? |
yes |
|
|
For HTs and CIs for comparing means from two independent samples with small sample sizes, you want the samples to have similar __________ with no outliers. |
shapes
|
|
|
For HTs and CIs for comparing means from two independent samples with small sample sizes, you want the samples to have similar shapes with no __________. |
outliers |
|
|
How can you get a good idea about the shape of a distribution? |
histogram or boxplot |
|
|
How can you get a good idea if there are outliers? |
histogram or boxplot |
|
|
|
|
|
|
|
The best guess for the variance of the difference of the sample means from X and Y is ___________ |
|
|
|
|
|
What will be the shape? |
t distribution with 11 df |
|
What will be the mean? |
0 |
|
What will be the standard deviation? |
|
|
|
How do you go from how many successes (Binomial) to proportion of successes? |
divide by n |
|
|
If you divide the number of successes by n, the mean gets divided by ____. |
n |
|
|
If you divide the number of successes by n, the variance gets divided by ______ |
|
|
|
The mean of the binomial is np which divided by n is ______, which is the mean of p’. |
p |
|
|
|
|
|
|
|
|
If a population is normal, then dividing by n will give it what shape? |
normal |
|
|
The binomial is approximately normal when np and nq exceed _______ so p’ is also approximately normal under the same conditions. |
10 |
|
|
true |
|
|
p' |
|
|
q' |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
0 |
|
|
|
|
|
normal |
|
|
With a Test for Independence why are the Es = (row total)(column total)/(grand total)? For E that is for Row 2 and Column 3, E should be (grand total)(P(_________________)) = n(P(R2) P(C3)) because we assume that Ho is true which is that the rows and columns are independent. The best estimate for P(R2) = (R2 total)/n and for P(C3) = (C3 total/n) making the estimate (R2 total)(C3 total)/n. |
R2 & C3 |
|
|
With a Test for Independence why are the Es = (row total)(column total)/(grand total)? For E that is for Row 2 and Column 3, E should be (grand total)(P(R2 & C3)) = n(P(_____) P(C3)) because we assume that Ho is true which is that the rows and columns are independent. The best estimate for P(R2) = (R2 total)/n and for P(C3) = (C3 total/n) making the estimate (R2 total)(C3 total)/n. |
R2 |
|
|
With a Test for Independence why are the Es = (row total)(column total)/(grand total)? For E that is for Row 2 and Column 3, E should be (grand total)(P(P(R2 & C3)) = n(P(R2) P(_____)) because we assume that Ho is true which is that the rows and columns are independent. The best estimate for P(R2) = R2 total / n and for P(C3) = (C3 total/n) making the estimate (R2 total)(C3 total)/n. |
C3 |
|
|
With a Test for Independence why are the Es = (row total)(column total)/(grand total)? For E that is for Row 2 and Column 3, E should be (grand total)(P(P(R2 & C3)) = n(P(R2) P(C3)) because we assume that Ho is true which is that the rows and columns are ______________. The best estimate for P(R2) = (R2 total)/n and for P(C3) = (C3 total)/n making the estimate (R2 total)(C3 total)/n. |
independent |
|
|
With a Test for Independence why are the Es = (row total)(column total)/(grand total)? For E that is for Row 2 and Column 3, E should be (grand total)(P(P(R2 & C3)) = n(P(R2) P(C3)) because we assume that Ho is true which is that the rows and columns are independent. The best estimate for P(R2) = __________ and for P(C3) = (C3 total/n) making the estimate (R2 total)(C3 total)/n. |
R2 total / grand tot |
|
|
With a Test for Independence why are the Es = (row total)(column total)/(grand total)? For E that is for Row 2 and Column 3, E should be (grand total)(P(P(R2 & C3)) = n(P(R2) P(C3)) because we assume that Ho is true which is that the rows and columns are independent. The best estimate for P(R2) = (R2 total)/n and for P(C3) = ___________ making the estimate (R2 total)(C3 total)/n. |
C3 total / grand tot |
|
|
better |
|
|
worse |
|
|
better |
|
|
worse |
|
|
Can you still do all the calculations for CIs and HTs if the data is bad? |
yes |
|
|
Should you do all the calculations for CIs and HTs if the data is bad? |
no |
|
|
Two advantages of Non-parametric statistics are that they tend to be less assumptions and ____________. |
easy to apply |
|
|
Two advantages of Non-parametric statistics are that they tend to be easy to apply and ____________. |
less assumptions |
|
|
Give a disadvantage of Non-parametric statistics. |
waste info |
|
|
If you want to do a HT about the mean but the sample size is small and there is an outlier, you might instead do a HT about the __________ and use the nonparametric test called the Sign Test. |
median |
|
|
If you want to do a HT about the mean but the sample size is small and there is an outlier, you might instead do a HT about the median and use the nonparametric test called the _______________. |
Sign Test |
|
|
With a Test for Independence why are the Es = (row total)(column total)/(grand total)? For E that is for Row 2 and Column 3, E should be (grand total)(P(P(R2 & C3)) = n(P(R2) P(C3)) because we assume that Ho is true which is that the rows and columns are independent. The best estimate for P(R2) = (R2 total)/n and for P(C3) = __________ making the estimate (R2 total)(C3 total)/n. |
C3 total |
