Statistics and Research Flashcards
Alpha (level of power)
Alpha determines the probability of rejecting null hypothesis when it is true. The value of alpha is set by the researcher before collecting or analyzing data. Usually .01 or .05
ANCOVA (Analysis of covariance)
The analysis of covariance is a version of the ANOVA that is used to increase the efficiency of the analysis by statistically removing variability in the DV that is due to an extraneous variable. When using the ANCOVA, each person’s score on the DV is adjusted on the basis of his score on the extraneous variable.
Between-Groups Designs
are experimental research designs that allow a researcher to assess the effects of the different levels of one or more IVs by administering each level or combination of levels to a different group of subjects.
Between-Groups Designs
are experimental research designs that allow a researcher to assess the effects of the different levels of one or more IVs by administering each level or combination of levels to a different group of subjects.
Central Limit Theorem
derived from the probability theory and predicts that the sampling distribution of the mean (a) will approach a normal shape as the sample size increases, regardless of the shape of the population distribution of scores, (b) has a mean equal to the population mean, (c) has the standard deviation equal to the deviation referred to as the standard error of the mean.
Central Limit Theorem
derived from the probability theory and predicts that the sampling distribution of the mean (a) will approach a normal shape as the sample size increases, regardless of the shape of the population distribution of scores, (b) has a mean equal to the population mean, (c) has the standard deviation equal to the deviation referred to as the standard error of the mean.
CHI-SQUARE TEST (Single/Multi Sample)
is a nonparametric statistical test that is used with nominal data. It is a goodness of fit type of test. (O bserved - E xpected)squared divided by E xpected. The answer minus 1 gives degrees of freedom. Standard acceptance of Null Hyp starts after .05
CHI-SQUARE TEST (Single/Multi Sample)
is a nonparametric statistical test that is used with nominal data. It is a goodness of fit type of test. (O bserved - E xpected)squared divided by E xpected. The answer minus 1 gives degrees of freedom. Standard acceptance of Null Hyp starts after .05
Cluster Analysis
is a multivariate technique used to group people or objects into a smaller number of mutually exclusive and exhaustive subgroups (clusters) based on their similarities.
Cross-validation / Shrinkage
refers to validating a correlation coefficient on a new sample. The correlation coefficient tends to shrink on cross-validation. Shrinkage is greatest when the original sample size is small and the number of predictors is large.
Discriminant Function Analysis
is the appropriate multivariate technique when 2 or more continuos predictors will be used to predict or estimate a person’s status on a single discrete (nominal) criterion.
Effect Size
is the measure of magnitude of the relationship between the dependent and independent variables; useful for interpreting the relationship’s clinical or practical significance. Use Cohen’s d (difference between groups in standard deviations) or eta squared (percent of variance in DV accounted for by variance in the IV)
Experimental Research (Quasi & True)
conducting an empirical study to test hypothesis about relationships. TRUE experiments hallmark is random assignment into groups.
Experimental Research (Quasi & True)
conducting an empirical study to test hypothesis about relationships. TRUE experiments hallmark is random assignment into groups.
Experimentwise Error Rate
The prob of making a type I error. As the number of statistical comparisons increases so does the EER.
External Validity
The degree to which a study’s results can be generalized to other people, settings and conditions. Threats include: pretest sensitization, multiple-treatment interference,
Factorial ANOVA (2-way, 3-WAY ANOVA)
appropriate when study includes 2 or more IVs and a single DV.
Internal Validity
Is it a causal relationship? Are we measuring what we say we are measuring?
Internal Validity
Is it a causal relationship? Are we measuring what we say we are measuring?
Interval recording / Event Sampling
method of behavioral sampling that involves dividing a period of time into discrete intervals and recording whether the behavior occurs in each interval. It’s useful for behaviors that have no clear beginning or end. Event sampling is useful for rare occuring variables / behaviors.
LISREL
a structural equation (causal) modeling technique that is used to verify a predefined causal model or theory. Allows 2-way non-recursive paths and takes into account observed variables.
MANOVA (Multivariate analysis of variance)
One or more IVs and 2 or more DVs each measured on a different interval scale. Helps reduce experimentwise error rate, and increases power.
Measures of Central Tendency
MMMR
Mean = Sum the divide by N
Mode = Which is repeated most, can be more than 1 or none.
Median = The number in the middle, if two add them together then divide by 2.
Range = line up in numerical order. subtract the smallest from the largest.
Mixed ANOVA (Split-plot)
used when study includes both between and within groups variables.
Mixed Designs
Type of factorial design in which at least 1 IV is a between-groups and one is a within-subjects variable.
Mixed Designs
Type of factorial design in which at least 1 IV is a between-groups and one is a within-subjects variable.
Moderator and Mediator Variables
Moderators affect the strength or direction of the relationship between IVs and DVs. If treatment is greater for men than women, gender is a moderator variable. Mediators explain for the relationships between IVs and DVs; Authoritative parents have positive impact on academics due to its impact on self-efficacy beliefs.
Multiple Regression / Multicollinearity
m-regression is a technique that is used for predicting a score on a continuous criterion based
Multiple Regression / Multicollinearity
m-regression is a technique that is used for predicting a score on a continuous criterion based performance of 2 or more continuous predictors. Multiple IVs impact on the DV. Ex: Distance traveled and number of deliveries impact on time it takes to complete route. Multicollinearity are the relationships between the IVs not the IV to DV. Ie: sea salt and regular salt both make a salty flavor, hard to tell which is responsible.
Nominal Curve / Areas under the normal
bell-shaped distribution that is defined by math formula. Usually; 68% of results fall in the range that are plus and minus 1 standard deviation, 95% fall between plus or minus 2 and 99% fall between plus or minus 3.
Nominal Curve / Areas under the normal
bell-shaped distribution that is defined by math formula. Usually; 68% of results fall in the range that are plus and minus 1 standard deviation, 95% fall between plus or minus 2 and 99% fall between plus or minus 3.
One-Way ANOVA / F Ratio
Used to compare means of two or more groups when a study includes 1 DV and 1 IV measured on an interval scale. The F-ratio indicates if any group means are sig different. When treatment has an effect, F-Ratio is greater than 1.0
Parametric Tests
More powerful than non-parametric. Include Student’s T tests; ANOVA.
Nonparametric Tests
Less powerful than parametric. Used to analyze nominal or ordinal data. Include chi-square, Mann-Whitney U test and Wilcoxon matched-pairs test.
Path Analysis
structural equation (causal) modeling technique used to verify a pre-defined causal model or theory.
Probability Sampling
simple random; stratified random; cluster sampling.
Random Error
unpredictable; sampling error & measurement error.
Random Error
unpredictable; sampling error & measurement error.
Randomized Block ANOVA
used to control for extraneous variable
Scales of Measurement
Nominal-
Ordinal-
Interval-
Ratio-
Standard Deviation
square root of the varience
Statistical Power
refers to the probability of rejecting a false null hypothesis. Can not be directly controlled, but can be increased by increasing sample size, maximizing the effects of the IV, increasing size of Alpha, reducing error.
Student’s t-Test
Parametric test used to compare 2 means.
Systemic Error / Extraneous variables
systemic = predictable. extraneous = confounding / systemic errors that affect Rx between IV and DV
Trend Analysis
used to assess linear and nonlinear trends when the IV is quantitative.
Type I Errors
when a true null hypothesis is rejected. The prob of making a type I error is alpha.
Type II Errors
False Null is retained. Prob of making II error is equal to Beta.
Within-Subjects Design
Each participant receives, at different times, each level of the IV. Comparisons on the DV are made within subjects rather than between groups.
