Stats practice
Do you have a true Independent variable?
Is yes, use Regression analysis
Type of statistic to predict group membership
Logistic regression
T-test
Looks at the differences between two groups The IV must have ONLY two groups
The mathematical average
Mean
The center value
Median
The most commonly occurring value
Mode
MANOVA
Same as ANOVA but you can study two or more related DVs while controlling for the correlation between the DV
MANCOVA
Same as MANOVA but adds controls of one or more covariates that may effect the influence of DV
Is this study interested in INTERVENTIONS (group differences)?
T-test, anova, manova, chi square
ANOVA
Tests the significance of group differences between two or more groups (but doesn't tell you what the difference is)
Several ways to refer to the significance level of a test:
The finding is significant at the 0.05 level. The confidence level is 95 percent. The Type I error rate is 0.05. The alpha level is 0.05. α = 0.05. There is a 1 in 20 chance of obtaining this result (or one more extreme). The area of the region of rejection is 0.05. The p‐value is 0.05. p = 0.05.
The width of the confidence interval is related to the confidence level, standard error, and n such that the following are true:
The higher the percentage of confidence desired, the wider the confidence interval. The larger the standard error, the wider the confidence interval. The larger the n, the smaller the standard error, and so the narrower the confidence interval.
Types of Statistics for Group Differences
chi-sqaure t-test Anova ancova Manova Mancova
A Type I error can occur when the null hypothesis is
correct.
If the significance level, α, is increased then the chance of a Type II error will
decrease
If the significance level, α, is increased then the power will
increase
If the significance level, α, is increased, then the chance of a Type I error will
increase
A Type II error can occur when the null hypothesis
is incorrect
A null hypothesis should be rejected when the p‐value is:
less than the significance level.
The conditional probability of obtaining the sample statistic, or one more extreme, given that the null hypothesis is true is called the:
p-value
ANCOVA
same as ANOVA but adds control of one or more covariates that may influence DV
Bottles of water have a label stating that the volume is 12 oz. A consumer group suspects the bottles are under‐filled and plans to conduct a test. A Type I error in this situation would mean
the consumer group concludes the bottles have less than 12 oz. when the mean actually is 12 oz.
The smaller the significance level p:
the more stringent the test and the greater the likelihood that the conclusion is correct
Selecting the significance level α will determine
the probability of a Type I error
What kind of statistic do you use to test for differences in the means of groups?
two groups: t -test more than two groups: ANOVA
The power of a test can be increased by
using a larger sample size
Which is not a typical significance level?
.25
Standard Deviation
1. How much scores deviate from the mean 2. Is the square root of the variance 3. is the most commonly used measure of spread
Selecting the significance level α is based on:
1. desired strength of evidence against the null hypothesis. 2. an acceptable probability of a Type I error.
Types of Statistics to determine Degree of Relationship
Bivariate correlation Multiple regression Path analysis
Chi Square
Compares observed frequencies to expected frequencies.
Nominal or ordinal data (discrete or categorical)
Use chi square
Is this study interested in description (association)?
Use correlations, factor analysis, and path analysis
IS this study interested in explanation or prediction?
Use: regression, logistic regression, discriminant analysis
A measure of how spread out a distribution is
Variance
A type II error occurs when
a null hypothesis is not rejected but should be rejected. Saying there are no differences or effects when there actually are.
A type I error occurs when
a null hypothesis is rejected but should not be rejected. Saying there is an effect/difference when there is actually not a statistically sig effect or difference
All other things being equal:
a smaller confidence interval is always more desirable than a larger one because a smaller interval means the population parameter can be estimated more accurately.