Exam 3 cheat sheet stats PDF

Preview

Summary

Pamela Burch...

Description



Paired means o Use this when/graph: use purple sheet o Find the CI: go to summary stats in stats and find the mean of both columns, then subtract them to find the difference (xD), find T using the T-table and DF 1. Interpret CI for obs: I’m (%) % confident the population mean ___ are/is between ___ and ___ (units) less (-,-)/more (+,+) than ___. 2. Interpret CI for exper: I’m (%) % confident the mean ___ are/is between ___ and ___ (units) less (-,-)/more (+,+) than ___. 3. Interpret CI for (-,+): I’m (%) % confident there is no significant difference in the mean ___ for ___ and ___. 4. STATCRUNCH: (on masterpiece) sample 1: 1st column, sample 2: 2nd column, confidence interval for whatever percentage is in the problem, compute o Hypothesis test for paired means (experiment): 1. Write the HO: µ1 = µ2 and HA: µ1 (≠)µ2 OR HO: µD = 0 and HA: µD (≠)0 2. Alpha = .05 3. Conditions for an experiment: Random assignment, independent, 10% condition, Normal? If… 

Conditions for an observation: representative, independent, 10% condition, Normal? 

The population was already normal OR



n is “large enough” (n > 40) OR



n is between 5 and 40 and the population is not very skewed *the more skew the closer n needs to be to 40

4. T-stat: find using the masterpiece 

Interpret: = (mean on SC) is located (t-stat) SE to the (left , both sides ≠) of µ = 0.

5. P-value: P( > alpha? Fail to reject HO. P-value < alpha? Reject HO. 7. Conclusion for experiment: There is weak(FtR)/strong(R) evidence to show _____ caused the mean ___. Example (There is weak evidence to show smiling caused the mean age to be older than not smiling.) 8. Conclusion for observation: There is weak(FtR)/strong(R) evidence to show the population mean ___ is/are different(≠) /lower() than ___. Example there

is strong evidence to show the population mean prices at WinCo products are less than at Walmart 9. STATCRUNCH: use masterpiece to know where to start. Remember that is the mean  Sample 1: 1st column, sample 2: 2nd column, hypothesis test: HO: µD = 0 HA: µD (≠)0, compute 

Independent (Two-sample) means – o Use this when/graph: use purple sheet o Find the CI: the same as paired means o Hypothesis test for independent samples 1. Write the HO: µ1 = µ2 and HA: µ1 (≠)µ2 2. Alpha = .05 3. Conditions for an experiment: Random assignment, independent, 10% condition, Normal? If… 

Conditions for an observation: representative, independent, 10% condition, Normal? Same as paired means

4. T-stat: find using the masterpiece 

Interpret: μ1 - μ2= (samp diff on SC) is located (t-stat) SE to the (left , both sides ≠) of μ1 - μ2= 0.

5. P-value: P( 

> alpha? Fail to reject HO. P-value < alpha? Reject HO. 7. Conclusion for experiment: There is weak(FtR)/strong(R) evidence to show _____ caused/s the mean ___. Example (There is strong evidence to show using fluoride causes the mean number of cavities to be less.) 8. Conclusion for observation: There is weak(FtR)/strong(R) evidence to show the population mean ___ is/are different(≠) /lower() than ___. Example (This is weak evidence to show the population mean GPA for men in fraternities is different from men not in fraternities.) 9. STATCRUNCH: use masterpiece to know where to start. Remember that is the mean  Sample 1: 1st column, sample 2: 2nd column, hypothesis test: HO: µD = 0 HA: µD (≠)0, compute 

ANOVA for Independent samples –how to compute the F test statistic how to fill out the table by hand o Use this when/graph: use purple sheet

o Hypothesis test: 1. Write the HO: µ1 = µ2 = µ3 = µ4 HA: At least one µi is different. 2. Alpha = .05 3. Conditions for an experiment: Random assignment, independent, equal variances, Normal? same as paired means 

Conditions for an observation: representative, independent, equal variances, Normal?



Equal variances: largest s2(or square the largest std dev)/smallest s2 (or square the smallest std dev)< 4 (if n’s are different) or 9 (if n’s are the same)

4. F-stat: find using masterpiece 

Interpret: The variation in ____because of ___ is (f-stat) times bigger than because of random error.

5. P-value: P(F>(f-stat)) = p-value 

Interpret: Assuming HO is right, there is an (p-valu %) % chance of getting an F of (f-stat) or even greater.

6. Decision: p-value > alpha? Fail to reject HO. P-value < alpha? Reject HO. 7. Conclusion for experiment: There is weak(FtR)/strong(R) evidence to show the _____ caused/s at least one mean ___ to be different. Example (There is weak evidence to show the sighting method causes at last one mean shooting core to be different.) 8. Conclusion for observation: There is weak(FtR)/strong(R) evidence to show the population mean ___ is/are different for at least one ___. Example (There is strong evidence to show the population mean marketing manager salary is different for at least one region. 9. STATCRUNCH: use masterpiece to know where to start/ EXCEL  

select all the columns that are relevant, compute/ data analysis, ANOVA Single Factor,

ANOVA for Blocked samples –how to identify the response, treatment and blocking variables, how to compute the F test statistic using technology (Excel or StatCrunch), how to interpret the meaning of the F test statistic, what would have happened if you had not blocked (had used Independent samples instead), how to find the P-value and interpret its meaning, how to decide whether you will reject or fail to reject the null hypothesis, how to write a conclusion in terms of the problem (note the different ways to write a conclusion for observations vs experiments). How to read the Tukey output. o Use this when/graph: use purple sheet o Hypothesis test: 1. Write the HO: µ1 = µ2 = µ3 = µ4 HA: At least one µi is different. 2. Alpha = .05

3. Conditions for an experiment: Random assignment, independent, equal variances, Normal? same as paired means 

Conditions for an observation: representative, independent, equal variances, Normal?



Equal variances: largest s2(or square the largest std dev)/smallest s2 (or square the smallest std dev)< 4 (if n’s are different) or 9 (if n’s are the same)

4. Response = usually a number treatments = the thing that’s being repeated blocks = first column usually 5....