Module 9: Inferential Statistics Flashcards
Why do researchers use inferential statistics?
To determine cause and effect
Compare differences between groups
To determine relationships amount variables
Estimate a population parameter
Statistical inference consists of what 2 techniques
Estimation of parameters
Hypothesis testing
What is point estimation?
The computation of a value (by the computer) that will estimate your population parameter
Why do researchers do an interval estimation instead of point estimation?
D/t the margin of error associated with using one point like in point estimation. Interval estimation uses a range of values around which the population parameter will lie
What is the terms used to describe the range of values within which a population parameter is estimated to lie?
Confidence interval (95%)
What does the statement:95% of scores lie within 2 SD of the mean? (CI)
We’re 95% confident that the correct value for our population mean (mu) lies within that range of scores
Hypothesis testing uses _______ criteria for deciding whether a hypothesis is supported by data.
objective
What is a major factor influencing standard error of the mean?
Sample size
What does hypothesis testing help to distinguish?
Whether differences in the population occurred by chance or were true population differences
What type of tailed test is typically used if a researcher has stated a directional hypothesis?
One tailed (nondirectional= two tailed, many will run two tailed just ot be sure with a directional)
What does the null hypothesis state?
That there is NO relationship between variables
Describe alpha related to level of significance. (value, who set by, probability of ____)
0.05, 5% of the time results occurred by chance
Set by researcher
Probability of making a Type I error
Describe statistical significance, p value. (value, set by, allow us to ____)
<0.05= statistical sig (less than 5% of the time the results occured by chance)
>0.05= not statistical sig (over 5% of the time the results did occur by chance)
Set by/calculated by computer
Allows us to know if we can accept or reject the null
Which statistical tests are higher level and more powerful?
Parametric
In parametric statistical testing what level of measure of data is the DV and the IV?
IV: nominal (ex-gender, ethnic groups)
DV: interval
In parametric statistical testing we assume that data is _______ distributed.
normally
normality assumption
Parametric statistical tests allow researchers to compare _____ ______ .
group means
Paired t-test, independent t-test, ANOVA
Parametric statistical tests also allow researchers to determine ________ among two variables, at _____ level.
relationship
interval
(Pearson’s r- correlation)
In nonparametric statistical testing what level of measurement are the DV and IV?
IV: nominal
DV: non-interval level
What do we assume about nonparametric statistical testing in relation to the normality assumption?
We assume they do NOT meet the normality assumption (data is skewed)
Nonparametric statistical testing can compare group means and relationships between two variables at _________ level
non-interval level
Comparing groups: Chi Square, Mann-Whitney U, Kruskal-Wallis
Relationship: Phi coefficient, Spearman’s rho
Which parametric procedure is used for testing differences in group means?
t-tests
When is a Bonferroni correction applied by researchers?
When running multiple tests with the same data to establish a more conservative alpha and avoid a Type I error (this correction runs the risk of creating a Type II error by decreasing alpha to 0.017 for example)
Parametric:
When means for two sets of scores are dependent, researchers use a _____ _ _____. Often a pretest/posttest design.
Paired t-test
Parametric:
When means for two sets of scores are independent, researchers use an _____ _ _____. Measured one time only
Student’s (Independent) t-test
Non-parametric:
In an independent two-group situation where the dependent variable is non-interval level or the distribution is markedly non-normal, what test is used?
Mann-Whitney U test
Non-parametric:
When non-interval level data are paired (dependent) and measured two times which test is used?
Wilcoxon signed ranks test
involves taking the difference between paired scores and ranking the absolute difference
Which parametric test is used to test differences between means when there are three or more groups?
ANOVA
What statistic is computed using an ANOVA?
F-ratio
What nonparametric test is similar to the Mann Whitney U in that it is based on assigning ranks to the scores of various groups but is comparing greater than two groups and a one-way test for independent samples is desired?
Kruskal-Wallis
What nonparametric test is used for more than two dependent groups where data is measured by repeated measures?
Friedman Test
Which test is used to test hypotheses about group differences in proportions?
Chi-square test
computed using observed frequencies and expected frequencies
What are the indices that describes correlational relationship between two nominal level, ordinal level, and interval level dichotomous variables?
Nominal: Phi coefficient
Ordinal: Spearman’s Rho
Interval: Pearson’s-r