- StudyBlue
- Wisconsin
- University of Wisconsin - Madison
- Psychology
- Psychology 210
- Hendricks
- Though Questions Unit 2
Though Questions Unit 2
Psychology 210 with Hendricks at University of Wisconsin - Madison
About this deck
Textbook:
Statistics for the Behavioral SciencesCreated: 2010-03-22
Size: 45 flashcards
Views: 241
About StudyBlue
Dennis
Sign up (free) to study this.
- Population - uses greek letters - entirety of what your studying-variability=sigma, center is ?
- Sample - subset of population, usually representative of the population - english letters-center = M, varib =s
- sampling distribution - distribution of sample means, variab=?M. mean=?, infer from samples about population
- Empirically derived is trial and error. take repeated samples of a population and find all the means. then find the mean of those means. --CON-time consuming
- Population derived - based on CLT1 -CON- need to know population
- the bigger the sample size the more clustered around the, the narrower it is.
- sampling error decreases
- ?M
- mean age of jurors
- battery life
- any problem relating M's
- only 5/100 times the ? will not fall with the intervals set by the means
- can infer things about population through the samples
- without the CLT we can not predict/set probability
- Scien:what you want to find --alt hypoth
- Stat: what you want to disprove -null hypoth
- focus on null because easier to prove some thing wrong once than prove it right for every possible case
- directional-one tailed-shows an increase or decrease in the treatment--better power
- non directional--two tailed--more conservative and safer--when in doubt use this --shows a change(significance)but not which direction
- reject null when you shouldn't have
- claim significance when there is none
- purely happens by chance--experimenter has no control over it
- represented by alpha
- the lower alpha, the higher beta, the lower power. = bad
- there is a 5% chance of committing a type 1 error
- means found in that region are considered significant and not due to chance
- can create false hope for patients
- money
- integrity
- reputation for science and scientist
- time--delays
- may not consider treatments that actually work
- claim no significance when there actually was significance
- represented by beta
- fail to reject the null
- Factors:
- ---sample size
- ---diff to be detected
- ---variabilty
- ---test type (directional? dependent?)
- type 1 happens purely buy chance, the experimenter has no control and did nothing wrong
- type 2- experimenter choses sample size and the other factors listed previously and therefore has more blame in committing of a type 2 error
· Says that 95% of the sample means will fall within a given confidence int.?the ? will fall w/in this as well
· Both hypoth and conf int based on alpha level?foundation material is similar
· Hypoth testing based on null hypoth, conf int based on actual outsomes (range)
· Suggested as an alt to hypoth testing
· Follow up to 2 tailed· Interval estimate is confidence int
· In std dev units
· Shows how much a treatment actually effected a condition
· Measure of magnitude· Significance only tells if something had an effect or not, effect size tells you how much of an effect it actually had
· Easier to have significance but small effect size with large n· Ability to detect a change in the data if such a change actually occurred
· Power = 1 ? ?
· Can see if a treatment will be effective before testing is actually done
· Power increases as type 2 error decreases· Sample size increase power increase?width shrinks
· Direction -> non directional?alpha decreases, beta increases, power decreases
· Discrepancy to be detected- position of alt curve changes?difficult small differences
· Alpha changes- decreases, beta increases, power decreases?goal line changes
· Variability ? position shifts, variability increases, power decreases26. What are some criticisms of, or concerns regarding, traditional hypothesis testing?
· Arbitrary significance?not based on actual consequences of type 1 error
· Dichotomous logic--black and white!
· Overemphasis on significance
· Inadequate attention to other factors that influence significance- i.e. sample size, variance(poor control)· Confidence interval?alpha level still arbitrary
· Alpha level based on risk ?what constitutes risk?
· Report actual probability and reader can determine significance
- what percentage of the total variance is accounted for by the treatment/study
- measure of magnitude in percent
- how much influence th IV had on the DV
· Confidence int:---The actual data found, 95% of the sample means should fall within this interval?still based on alpha (CLT)?not dichotomous like hypoth testing
· Hypoth. Testing focuses on the null (what didn?t happen)
o -black and white-no gray.
· Effect size---how many std dev did the treatment chance the control
· Proportion of variance tells how much of the total variance can be accounted for by the study done
- the center is around mean1 -mean2, which is 0. the variability is sm1-m2. -normal distribution(approaching even more normal with increasing n
- made of differences between means
- very robust
- when equal n errors can just be averaged together, when unequal they are weighted the one wit the bigger n as more weight than the the other
· Correlation equation subtracts out consistent indiv differenced right away
· Sd/?n uses the individual differences variation right away
· It is multiplies by the error that gets subtracted out therefore the better the correlation the lower the error.
· The better groups are matched the better the correlation the lower the error
· Power is greater for dependent because there is less error
· Based on degrees of correlation- r increases power increases· PROS ? better power, more correlation, less error, subtract out individual differences
· CONS ? difficulties in matching, carryover effects, loss of extremes
· Don?t need to know variability or difference to be detectedàneed to know to know power, so good that we don?t have to know them
· Estimate ratio of cohens d?what effect size do you want?About this deck
Textbook:
Statistics for the Behavioral SciencesCreated: 2010-03-22
Size: 45 flashcards
Views: 241
About StudyBlue
Dennis