psych 202 midterm
Which of the measures of central tendency is most affected by the presence of an outlier? Select one: a.Median b. Mode c. Mean d. They are all equally affected
c mean
in a normal curve, how are the mean median and mode related
they are all equal to each other
One of the problems with NHST is that it encourages dichotomous thinking, or seeing the world in black and white. Select one: True False
true
A study's method section reports that they had 40 subjects in group 1 and 38 subjects in group 2. What degree of freedom should be reported if the researchers conduct a t-Test? Select one: a. 76 b. 40 c. 38 d. 78 e. 77
a 76
Which DF term is reported in the results of a one way repeated measures ANOVA? Select one: a DFerror b. DFwithin-groups c. DFparticipant d. DFthanosdidnothingwrong
a DFerror
On average, what value is expected for the t statistic (t obtained) if the null hypothesis is true: Select one: a. 0 b. 1 c. .05 or less d. Depends on the sample means
a 0
An experimenter runs a study where he has 32 subjects and 4 conditions. Which of the following represents the degrees of freedom for within? Select one: a. 28 b. 31 c. 3 d. 32
a 28
describe standard deviation
a computed measure of how much scores vary around the mean score -generally want it to be small because it decreases variability in the population
The average squared deviation from the mean of a sample is called the... Select one: a. variance b. variability c. standard deviation d. range
a variance
Meta-analysis is widely regarded as one of (if not the best) ways to summarize data across a number of studies and get a firm estimate of the importance of a a finding in the real world. In what ways does NHST pose problems for meta-analysis? Select one or more: a. It is easier to publish significant effects, so studies that fail to find significance are often not published (the file drawer problem) b. Researchers who take an NHST approach often fail to report confidence intervals or the necessary statistics to calculate confidence intervals. c. p values are easier for researchers and lay audiences to intuitively understand compared to confidence intervals. d. NHST approaches make it difficult to generalize research findings beyond the current sample in any study.
a, b
what is a two independent groups design
o A single independent variable with two levels; each participant receives one and only one level of the IV
what are advantages of within subjects designs
o Compare same person under the influence of multiple levels of the IV § Instead of comparing apples and oranges we are comparing apples and apples o Reduce the influence of individual differences § Error variance (denominator of statistics) decreased § Increases likelihood of detecting true effect
how do you determine significance of a independent samples t test
o Compare your t valye to a critical value of t o Statistically significant § Difference due to treatment variance § T obtained is larger than t critical o Not significantl § Difference due to error variance § T obtained less than t critical
why should you use a paired samples t test
o Comparing the ratio of between subject variation to within subject variation o Can also think of as how much variance you can assign to treatment in one group and how much error is in that treatment o Within and error is partially due to individual differences o Each participant serves as his/ her own matched partner across treatments o By having participants act as their own controls, you limit the influence of individual differences o Therefore you can reduce error variance by removing variability associated with individual people (giving you more power)
describe the critical value of t
o Critical t defines the region of rejection o If we exceed this value, we can say with confidence that the groups are systematically different o Statistically significant difference between the means o Largest ratio of (between group variance)/ (within group variance) that were willing to attribute to chance o Anything larger is unlikely to be due to error variance
what are descriptive and inferential statistics
o Descriptive statistics: presenting, organizing, summarizing data o Inferential statistics: drawing conclusions about a population based on a sample
describe the hypotheses for ANOVA
o HO: All the means are equal § Group 1= group 2= group 3 o Ha: not all the means are equal § At least one mean is different
what are assumptions of ANOVA
o Homogeneity of variance (s relatively similar in each group) o Normality of DV o Independent observations
what is the critical value of F
o Just like in a t distribution the critical value of F varies o It depends on two dfs
describe effect size for independent group t test
o Measure of the magnitude of an effect: how different is group A from group B o Small- 0.2, medium- 0.5, large- 0.8 § Rough guidelines § Can go above one
how do you determine the independent samples t test df
o N1-1 + n2-1= df o Or N-2
describe the two types of alternative hypotheses of an independent samples t test and paired samples t test
o Non directional hypothesis (2 tailed) § Mean of group 1 is not equal to mean of group 2 § Think there is a difference but don't know which one is more o Directional hypothesis (1 tailed) § Mean of group 1 is larger than mean of group 2 § Mean of group 1 is less than mean of group 2
describe the null and research/ alternative hypotheses
o Null hypothesis: assume that there is no difference between groups/ no relationship between variables o Research/ alternative hypothesis: statement of a difference between groups/ relationship between variables
describe ANOVA
o One categorical IV with 2+ levels o One continuous DV (interval or ratio) -an ANOVA just tells us whether one group is different from the others
describe why statistical significance is either/ or for independent groups t test
o P values aren't stable o So, you should always report the exact p values and frame research questions in terms of how much o Significance tells us that sample difference likely occurs in population but does NOT tell us anything about the size of the effect
at what point do we say that a mean is improbable
o Region of rejection § means that fall within the region of rejection are so improbable (unlikely to occur) that we reject that they are representative of the population +/- 1.96
how can you choose what type of central tendency to use
-consider measurement scale § Interval/ ration scale: mean § Ordinal scale: median § Nominal scale: mode -consider shape of distribution § normal: mean § skewed: median § multimodal: mode
how do you determine effect size and CI for paired samples t test
-effect size: cohens d but adjust denominator -CI: same logic as independent groups
how do you compute variance
-first X minus the mean -then square it -sum all those numbers together -divide by n-1
how do you report statistically significant results of a independent samples t test in APA format
-identify test being used -identify two levels of the IV -identify descriptive statistics and the dirction of the effect (SD) -results of the statistical test (t(df), p, d) -provide brief interpretation of what the results mean in terms of the variables
how to you report a statistically significant result for paired samples t test in APA
-identify test being used -identify two levels of the IV -identify descriptive statistics and the dirction of the effect (SD) -results of the statistical test (t(df), p, d) -provide brief interpretation of what the results mean in terms of the variables
how do you report a non statistically significant result for paired samples t test in APA
-identify test being used -say that the variable did not differ because of whatever you were doing -descriptive stats and SD -stats block
how do you report non significant results of an independent samples t test in APA
-identify test being used -say that the variable did not differ because of whatever you were doing -descriptive stats and SD -stats block
what are NHST critiques
-it is kind of confusing (its intuitive and hard to understand) -encourages dichotomous thinking (world is not black and white) -we always have a chance of committing a type 1 error -p values are not stable (they flucuate)
how do you calculate CIs
-mean score + or - MOE
what are the properties of the normal curve
-mode=median=mean -symmetrical and asymptotic -standard deviations and % of observations
how do you compute standard deviation
-square root of variance
describe the standard error of the mean (SEM)
-standard deviation of the distribition of means o average amount that the sample means deviate from the population mean o n= size of sample o the population has more variability than the SDM § σ > σM
how do you calculate ANOVA (F)
-systematic variance/ error variance -between group variance/ within group variance
how do you calculate the paired samples t test
-t=Mdiff/SMdiff -standard error of the difference scores -not done by hand
what are post hoc tests used for and what are examples
-used to figure out which means are different from each other when an ANOVA test is significant -correct the alpha level -help prevent the multiple comparisons problem -bonferroni tests, holm test, games-howell test, LSD test
Adam develops a new scale measuring humor perception, gives it to every Furman student, and finds µ = 60, σ = 3. He thinks that psych students might have better senses of humor, and randomly samples 9 psych majors and finds their average humor score is 62. What can Adam conclude about the humor of psych majors at Furman? A. Psych majors at Furman have a statistically reliable better sense of humor than the rest of Furman students. B. Psych majors at Furman have the same sense of humor as the rest of campus.
A. Psych majors at Furman have a statistically reliable better sense of humor than the rest of Furman students.
What does a p value of .023 tell us? A. That the analysis was statistically significant. B. That the analysis was NOT statistically significant. C. That there was a large effect size. D. There there was a small effect size. E. More than one of the above (say which).
A. That the analysis was statistically significant.
________ are to the sample as __________ are to the population A. statistics; parameters B. central tendency; variation C. parameters; central tendency D. parameters; statistics
A. statistics; parameters
A researcher is interested in whether Working Memory Capacity (the ability to remember and manipulate information) predicts reading comprehension. She has WMC scores for a sample of 600 college students at her institution and recruits 30 low WMC students and 30 high WMC students to complete a reading task. She predicts that the students with high WMC will score higher on a comprehension quiz than students with low WMC. The researcher finds a t value of 2.37 and finds she has a statistically significant effect. What do we know about tcritical? A. t crit is less than 2.37 B. t crit is more than 2.37 C. t crit is equal to 2.37
A. t crit is less than 2.37
A researcher is interested in whether Working Memory Capacity (the ability to remember and manipulate information) predicts reading comprehension. She has WMC scores for a sample of 600 college students at her institution and recruits 30 low WMC students and 30 high WMC students to complete a reading task. She predicts that the students with high WMC will score higher on a comprehension quiz than students with low WMC. Is this researcher examining relationships between variables or looking for a difference between groups? A. A relationship between variables. B. A difference between groups.
B. A difference between groups.
A 5th grader is in the 14th percentile for height. Without knowing the mean or standard deviation of the population of 5th grade heights, what do we know about the 5th grader's Z score? A. It will be close to 0. B. It will be positive. C. It will be negative. D. Trick question; we can't know anything about the z score given what we know.
C. It will be negative.
A researcher is interested in whether Working Memory Capacity (the ability to remember and manipulate information) predicts reading comprehension. She has WMC scores for a sample of 600 college students at her institution and recruits 30 low WMC students and 30 high WMC students to complete a reading task. She predicts that the students with high WMC will score higher on a comprehension quiz than students with low WMC. Which of the following is the best description of her research hypothesis? A. H0: µhigh = µlow B. H0: µhigh > µlo C. Ha: µhigh = µlow D. Ha: µhigh > µlow E. Ha: µhigh ≠ µlow
D. Ha: µhigh > µlow
how do you calculate sum of squares for ANOVA
SS=sum(X-M)^2 § Within · Measure of each scores deviation from its own group mean · Sum (X-M)^2 § Between · Measure of each group mean deviation from total mean · sum (M-Mtotal)^2 § total · measure of each score's deviation from total mean · sum (X-Mtotal)^2
A statistically significant result is defined as a result that has a ________ probability of occurring if the ________ hypothesis is true. Select one: a. Small; null b. Small; alternative c. Large; null d. Large; alternative
a. Small; null
An article in The Paladin reports that "65% of Furman students (± 2%) are in favor of shortening the final exam schedule to just three days." The poll that was used to generate this statistic had 50 respondents. What does the confidence interval tell us about the attitudes of Furman students towards compressing the final exam schedule? Select one: a. That the true proportion of Furman students who are in favor of compressing the finals schedule is between 63% and 67%. b. That 95% of Furman students are likely in favor of compressing the finals schedule. c. Trick question: without a p value it is impossible to know how many Furman students are actually in favor of compressing the exam schedule. d. Because the confidence interval doesn't cross 0%, we can't conclude that a significant proportion of Furman students endorse compressing the finals schedule.
a. That the true proportion of Furman students who are in favor of compressing the finals schedule is between 63% and 67%.
you're doing a study where you have a large number of groups completing a task. You run an ANOVA and find a significant F value. The next step in your process is to run post-hoc tests. In this situation, you don't have a strong hypothesis about what differences you'll find, but you do want to be extra cautious about drawing any conclusions from the test. Which post-hoc test should you do? Select one: a. The Bonferroni multiple correction procedure b. Games-Howell c. LSD d. Tukey's HSD
a. The Bonferroni multiple correction procedure
Eta squared, and other effect size measures for ANOVA, represent what? Select one: a. The amount of variance in the dv that we can attribute to the independent variable. b. the amount of variance in the dv that we can attribute to the dependent variable. c. the amount of variance in the IV that we can attribute to unsystematic variance. d. Trick question. ANOVA doesn't have effect sizes.
a. The amount of variance in the dv that we can attribute to the independent variable.
Why do researchers conduct post-hoc tests after running an omnibus ANOVA? a. To determine which specific groups are different from which other groups b To determine the between-groups degrees of freedom c. As part of calculating mean squares d. To determine the effect size of the ANOVA
a. To determine which specific groups are different from which other groups
What can you conclude from the following information? t(48) = 4.14, p < .01 Select one: a. We can generalize this result to the population b. This is a small to medium effect c. There is a less than 1% chance that the IV did not affect the DV d. A total of 48 people were in the study
a. We can generalize this result to the population
This is defined as "a measure of the lack of symmetry or lopsidedness of a distribution. Select one: a. skewness b. kurtosis c. variance d. mean e. standard deviation
a. skewness
Given a dataset where the dependent variable is nominal, which of the following is the best measure of central tendency to use? Select one: a. the mode b. the median c. the standard deviation d. the mean
a. the mode
In Fry 2019, the author describes an unusual study done by David Speigelhalter where they put a subject into an fMRI and found sixteen sites in the subject's brain that activated in response to their stimuli. The unusual part of this study was that the subject was a "4lb Atlantic Salmon which 'was not alive at the time of scanning'." What point was Speigelhalter making with conducting this study? Select one: a.To demonstrate that chance findings will occasionally appear to be statistically significant b. To highlight the weaknesses in the peer review process c. To protest against the use of human subjects in psychology research d. To demonstrate why meta-analysis is an inferior statistical technique to classic NHST
a.To demonstrate that chance findings will occasionally appear to be statistically significant
The t-test is the appropriate statistic to use when: a.you are looking for differences between two groups b. you are looking for a correlation between two continuous variables c. you want to determine how unusual a sample is compared to a population d. you are trying to figure out if your friend likes green tea or black tea more
a.you are looking for differences between two groups
when computing for variance, why do we subtract 1 from the sample
as a correction given that we dont have access to all data points (population) -it is an adjustment that increases variability
what is sample variance
average of each scores squared difference from the mean -a way to try and quantify the spread
t(48) = 4.14, p < .01, d = .3, 95% CI [4.2-9.1] Based on the information above, t critical must be... Select one: a. Greater than 4.14 b. Less than 4.14 c. Less than .01 d. Between 4.2 and 9.1
b. Less than 4.14
What term captures systematic variance in the paired samples t-test? Select one: a. The difference between the two means of the IV (M1 - M2) b. The average difference score (Mdiff) c. The standard deviation of the difference between means. d. The standard error of the difference between means.
b. The average difference score (Mdiff)
What can you conclude from the following information? t(48) = 4.14, p < .01, d = 0.30, 95% CI [4.2-9.1] Select one: a. We can't generalize this result to the population b. This is a small to medium effect c. There is a less than 1% chance that the IV did not affect the DV d. A total of 48 people were in the study
b. This is a small to medium effect
Imagine the same study as before, but imagine the researchers reported the results like this: "The meditation group showed a(n) _________ but ______ decrease in anxiety symptoms, d = 0.09, 95%CI [0.08, 0.22]". Fill in the blanks: a. moderate effect size; significant b. statistically significant; small c. insignificant; small d. moderate effect size; statistically insignificant
b. statistically significant; small
Researchers conducted a study examining whether meditation can reduce feelings of anxiety and stress. They conducted an independent groups design with a meditation group and a control group. They reported their key results to be t(58) = 3.45, p = .013, d = 0.49. What can we conclude about the strength of their finding? a. the difference between groups on the dv, while statistically significant, is small. b. they reported a medium effect size, which means a moderate difference between scores on the dv. c. the large effect size they reported means that two scores on the dv are wildly different d. it is impossible to determine the effect size without more information.
b. they reported a medium effect size, which means a moderate difference between scores on the dv.
A one-way ANOVA is used instead of an independent samples t-test when a study has... Select one: a. 1 IV with 2 levels (between-groups) b. 1 IV with 2 levels (within-groups) c. 1 IV with 3 or more levels (between-groups) d. 2 IVs regardless of whether they are manipulated between or within subjects
c. 1 IV with 3 or more levels (between-groups)
With regard to the normal curve, _________ of scores are within 1 standard deviation of the mean. Select one: a. 97.50% b. 13.59% c. 68.26% d. 2.50%
c. 68.26%
Which of the following is not a way to quantify the variability of a sample? Select one: a. variance b. sum of squares c. degrees of freedom d. mean squares e. standard deviation
c. degrees of freedom
"children who play video games will differ in their hand-eye coordination from children who do not play video games." This hypothesis is an example of: Select one: a. directional hypothesis b. nonfalsifiable hypothesis c. nondirectional hypothesis d. contradictory hypothesis
c. nondirectional hypothesis
The standard deviation represents... Select one: a. the number of scores necessary to calculate a measure of central tendency b. the relationship between a sample and a population c. the average amount of variability in a set of scores. d. the difference between the highest and lowest scores in a data set
c. the average amount of variability in a set of scores.
Researchers conducted a study examining whether meditation can reduce feelings of anxiety and stress. They conducted an independent groups design with a meditation group and a control group. They reported their key results to be t(58) = 3.45, p = .013, d = 0.49. How many subjects were in their experiment? a. 56 b. 58 c. 59 d. 60 e. 3.45 f. 0.49
d 60
Which of the following types of hypotheses states that there is no relationship between your variables? Select one: a. Population hypothesis b. Alternative hypothesis c. Research hypothesis d. Null hypothesis
d. Null hypothesis
You read a research paper that includes a histogram for the DV. Which of the following is on the y-axis? Select one: a. The scores on the dependent variable b. The independent variable c. The mean for each level of the independent variable d. The frequency with which each level of the DV appears
d. The frequency with which each level of the DV appears
If someone has a z score of -1.5, this means... Select one: a. Their score is equal to one and a half times the standard deviation b. Their score is at the 98.5th percentile c. The score is 1.5 points lower than the mean d. Their score is one and a half standard deviations below the mean
d. Their score is one and a half standard deviations below the mean
In the reading, Cumming identifies which of the following as being a problem with p values and the general approach of NHST? Select one: a. p values are likely to differ wildly, even with a direct replication. b. p values encourage researchers to think dichotomously (an effect either exists or not) c. p values fail to communicate the size of an effect d. all of the above are problems associated with p values.
d. all of the above are problems associated with p values.
Which of the following is NOT a source of variability in an ANOVA? Select one: a. within groups b. between groups c. treatment d. post-hoc e. error
d. post-hoc
Which of the following ratios best describes all of the statistical tests we've covered this semester? a. within-group variance / between-group variance b. treatment variance / systematic variance c. unsystematic variance / systematic variance d. error variance / treatment variance e. systematic variance / error variance
e. systematic variance / error variance
what are types of central tendencies and describe them
mean: average score median: middle score mode: common score
what does NHST stand for
null hypothesis significance testing
what are assumptions about the t test
o 1. Homogeneity of variance: variability will be same across groups o 2. Each observation is independent o 3. DV is normally distributed in population
describe the normal curve
o Remember to think of a distribution as a histogram o We can describe a normal curve with two properties § µ: mean § σ: SD o mean=median=mode o symmetrical o asymptotic: curve never touches the end (it never reaches zero) o with a normal curve you can separate it by standard deviations
what are the three principles of the central limit theorem
o SDM will approximate the normal curve, regardless of the shape of the population o mean of the SDM is equal to the population mean μ = μM o the population has more variability than the SDM . σ > σM
what is the t distribution
o The t distribution is bell shaped o The distribution changes depending on the df o After df 30, it is approximately normal
why would means differ
o Total variance= systematic variance + unsystematic variance o Having more systematic variance suggests our IV has an effect o Having more unsystematic suggests it does not o Statistical tests compare the ratio of systematic variance/ unsystematic variance
what is the critical value of t
o Unlike in a z distribution, the region of rejection is a t distrubtion varies on § Alpha § 1 tail or 2 tails § Degrees of freedom (df)
what is the independent groups t test
o Used to tell if two groups are similar or different o Technically: a comparison of whether mean differences in a sample are generalizable to the population from which that sample was drawn -between subjects t test (posttest or pretest/ posttest)
describe how NHST is process of inference
o We get our sample (ideally using a representative technique) o We measure our sample (with a valid and reliable measure) o We decide if our sample is unusual or just the result of chance) o We draw a conclusion about the population based on what we observed in the sample
what is the one sample z test
o allows us to calculate the probability of obtaining a specific sample mean from the population
describe z score
o deviation of a score from the mean in units of standard deviations § count of SD aware from the mean o note that a positive z score is above the mean o how many (count of) SD away from the mean
desribe effect size for ANOVA
o n^2 (eta squared) (weird looking n) § n^2=SSbtw/ SSttl o interpreting n^2 § small- .01, medium- .06, large- .14
what are the steps to how all statistical tests work
o step 1: identify your alternative hypothesis o step 2: identify the null hypothesis o step 3: step the alpha level or level of risk o steps 4-6: pick the correct test statistics + some other stuff o step 5: check to see if your test statistics (based on your sample) lands in the region of rejection o step 6a: if it does, we reject the null hypothesis o step 6b: if it does NOT, we fail to reject the null hypothesis
what is a distribution
representation of the frequency with which all possible values occur in a data set
what does sample describe and what does population describe
sample= sample statistics population= population parameters
what makes a normal distribution
symmetrical and unimodal
describe variability
the amount of spread in the distribution of scores -range: distance between 2 most extreme scores -variance -standard deviation -indicative of how typical the average is
when do you have a type 1 error
when you say there is a relationship but there is not one (false alarm)
when do you have a type 2 error
when you say there is no relationship but there is one (missed effect)
what do you use the z score for and what do you use the z statistic for
z score= individual z statistic= group
what is a CI
§ A confidence interval (CI) is a range of values that will likely contain the true value (the population value) § AKA with 95% CI- if we were to run this experiment 100 time 95/ 100 times the confidence interval would capture the true value § Cis not only tell us precision, they tell us significance (if we care) and give use a clue about replication
how do you interpret error bars
§ Bars that don't overlap suggest differences § Bars that cross 0 are N.S.
what is central tendency
§ Center of distribution: indicative of "typical" performance
describe the post hoc comparison process
§ Comparing each group to every other group in your design § Generate a p value for each comparison (can also generate Cohen's D)
how do you report ANOVA results in APA
§ F: F(dfbtw, dfwtn)= F obtained, p= x, n^2=x -report ANOVA results (plus effect size) -report post hoc tests -tell what is significant
describe the LSD test
§ Fishers least significant difference · Run a set of t tests · No correction to alpha rate (so inflation will occur) · Only use if you have strong a priori predications
how can you tell the difference between a bar graph and a histogram
§ Histogram is continuous data · Not categorical § Bar graph is useful for categorical data § Bar graph has space between bars where the histogram does not
what are the 2 key ideas for 95% CI interpretation independent samples t test
§ If finding is statistically significant the 95% CI will not include 0 § Small error bars indicate a more precise estimate of the mean
describe the Games-Howell test
§ Instead of the long dreadful computation we will click a box in jamovi § Reasons to like G-H · Adequately controls alpha inflation · Has relatively high power · Works with unequal sample sizes · Works with unequal variances
how are CIs intuitive
§ Point estimates: single number estimate of population parameters § Interval estimates/ MOE: range of plausible estimates; also tells us about precision
describe the Holm test
§ Slight variation of Bonferroni that is more powerful, but keeps the type 1 error rate inflation corrected § Always use holm in place of Bonferroni
describe number of peaks in a distribution
§ Unimodal distribution: there is one mode · Mode is most common score in a data set § Bimodal distribution: there are two peaks in a distribution § Multimodal: many peaks throughout the distribution § rectangular distribution: the distribution is completely flat
describe the bonferroni tests
§ Very conservative test that modifies alpha to reflect the number of comparisons being made § Corrected alpha= alpha/ #comparisons § Conduct a series of independent samples t tests using the corrected (more extreme) value of alpha § To be considered significant, p values for these post hoc t tests now have to be less than the corrected value of alpha · This makes it difficult to find a statistically significant effect § Aka low power
why do we use +/- 1.96 in the region of rejection
§ What is the probability associated with a z score of 1.96? (area beyond the curve) § Only a 5% chance that such a sample mean comes population § AKA p< .05
how do you calculate mean squares for ANOVA
§ average sum of squares § between · estimate of systemic variance · SSbetween/ dfbetween § Within · Estimate of unsystematic variance · SSWithin/ dfWithin
how do you find degree of freedom for ANOVA
§ dfTotal= N-1 · N= total number of participants in study § dfBetween= r-1 · r= number of groups (levels of IV) § dfWithin= N-r · N= total number of participants · r= number of groups § dfTotal= dfBetween + dfWIthin
describe the omnibus F test
§ if our one-way ANOVA yields significant results, we reject the null hypothesis § but we don't know how many differences there are or where they are § post hoc comparisons · Bonferroni/ holm procedures · Games-Howell procedure · LSD
what are key differences that the z statistic has compared to the z score
§ mean of a sample instead of single observation § instead of population mean, we use the mean of the sampling distribution § instead of using standard deviation of population were using standard error of mean
describe symmetry in distributions
§ symmetric: the right side is the exact same as the left side · can draw a line down the middle and fold it perfectly in half § asymmetrical: the data set is skewed (scores trail off in one direction) · positively skewed: tail pointing to positive scores · negatively skewed: tail pointing to negative scores
how can you separate a normal curve by standard deviation
· 34.13% between mean and 1 standard deviation o 68.26% of the observations fall within + or - 1 SD of the mean § 13.59% between + or - 1 and 2 SD · 95.44% of the observations fall within + or - 2 SD of the mean § 2.15% between + or - 2 and 3 SD · 99.74% of the observations fall within + or - 3 SD of the mean
describe the paired samples t test
· Paired samples t test will have the same subject test two times § Within subjects t test or repeated measures t test § Observations in groups can be related (same subject, husband wife pairs, matched design) -repeated or concurrent measures
why should you run an ANOVA instead of doing a bunch of t tests
· The probability of type 1 error increases with each test o Probability= 1-(1- alpha) to the c power o C= number of tests · Risk of type 1 error with 3 t tests? o 1- (1-.05) to the third power= .14 o In other words, your type 1 error rate increases rapidly with multiple comparisons