Statistical Tests/Treatments from Sir G and Atty G Flashcards
These tests significance of INTERVAL/RATIO DATA, using RANDOM SAMPLING with an UKNOWN POPULATION
Parametric Tests
These tests use a NON-PROBABILITY SAMPLING with a CATEGORICAL DATA
Non-parametric tests
The conclusions or assumptions from a parametric test can _____ to the population
be generalized
The conclusions or assumptions from a non-parametric test can _____ to the population
not be generalized
We expect ___ data and a ____ distribution from a non-parametric test
Homogenous; skewed
Two tests that fall under parametric sampling
T-test and ANOVA
The only tests with a single letter name that you should expect from a parametric test, the rest sa non-parametric na
Z, T, and F (ANOVA is also known as the F-test)
Parametric and non-parametric tests are used by researchers when they’re trying to ____ data
Compare
Regression analyses are used by researchers when they’re trying to ____ data
Predict
Treatments like the pearson r, spearman rho, phi coefficient, tetrachoric correlation are used by researchers when they’re trying to ____ data
Correlate/associate/link
Under correlation, this is the term used to show how much two scores vary together
Covariance
This is the is the result of a correlation or the mathematical index of the correlation
Correlational coefficient
The correlational coefficient describe the ___ and ___ of the relationship
direction and magnitude
The correlational coefficient ranges from
-1 to +1
This value of the correlational coefficient means no correlation at all
0
This correlational treatment is used when the data are continuous or interval
Pearson r
This correlational treatment is used when the data are ordinal in nature
Spearman rho
TEST ITEMS are ____ in nature
Ordinal (either right or wrong lang)
You can also use interval data for spearman rho because
Interval data can be CONVERTED to ordinal data
This correlational treatment is used when one or more of the data are TRUE DICHOTOMOUS
Meaning pwede two true dichotomous data, pwede one true + one artificial dichotomous data
Phi coefficient
This correlational treatment is used when you have TWO ARTIFICIAL DICHOTOMOUS (AD) data
Tetrachoric correlation
This correlational treatment is used when you have ONE CONTINUOUS/INTERVAL and ONE AD data
Biserial
This correlational treatment is used when you have ONE CONTINUOUS/INTERVAL and ONE TRUE DICHOTOMOUS (DT) data
Point-biserial
You use this correlational statistical treatment when doing ITEM DISCRIMINATION
Point-biserial
This correlational treatment is used when you want to know the interscorer agreement of 3 people or more with a test using ORDINAL DATA
Kendall’s Coefficient Test of Concordance
or Kendall’s W
This correlational treatment is used when you want to know the interscorer agreement using INTERVAL DATA
Kappa Statistic
Under Kappa Statistic, ___ is used when you have 3 raters and above
Fleiss’ Kappa
Under Kappa Statistic, ___ is used when you want to assess the agreement between 2 raters
Cohenn’s Kappa
This correlational treatment is used when you want to measure the NON-ASSOCIATION of two variables
Coefficient of Alienation
This correlational treatment is used when you want to know the suggested PERCENTAGE SHARED by two variables
JCoefficient of Determination (r2)
This test of prediction is used if you only have one predictor, and one factor being predicted
Linear regression
This test of prediction is used if you have MANY PREDICTORS predicting ONE FACTOR
Multiple regression
This test of prediction is used in FACTOR ANALYSIS
Multiple regression
This test of prediction is used if you are FILTERING a SERIES of predictors to look for the best predictor
Stepwise regression
This test of prediction is also known as BETA WEIGHTS and it shows how much a variable CONTRIBUTES to the another variable or the whole test
Standardized regression coefficients
This is also used in FACTOR LOADING
Standardized regression coefficients
A PARAMETRIC test that COMPARES groups of participants that aren’t related (independent) in any way
Independent samples t-tests
Independent samples t-tests is sometimes called
Between subjects design
A PARAMETRIC test that COMPARES groups of participants that are RELATED in some way
Paired samples t-test/dependent samples t-test/repeated measures design
When participants in the first group are the same participants in the second group, this is PARAMETRIC COMPARATIVE TEST will be used
Repeated measures design
Also known as the f-test
One-way ANOVA
A PARAMETRIC test if you have TWO LEVELS (treatments/conditions) of a SINGLE IV
One-way ANOVA
A PARAMETRIC test if you have TWO IVs having TWO OR MORE CONDITIONS
Two-way ANOVA
A PARAMETRIC test that needs ONE CATEGORICAL IV and a CONTINUOUS DV
Repeated measures ANOVA
A PARAMETRIC test like one-way ANOVA but for RELATED/DEPENDENT GROUPS
Repeated measures ANOVA
The extension of a DEPENDENT T-TEST/repeated measures design
Repeated measures ANOVA
Repeated measures design is also called
Within-subjects ANOVA or ANOVA for correlated samples
This is used when the ANOVA is significant and you want to know which conditions have the SIGNIFCANT DIFFERENCE
Post-hoc tests
How do post-hoc tests compare conditions?
Two at a time
The post-hoc test used for one-way ANOVA/between groups
Scheffe post-hoc
The post-hoc used for two-way ANOVA/among groups
Tukey post-hoc
This NON-PARAMETRIC TEST is used to determine and significant difference between the EXPECTED FREQUENCIES and the OBSERVED FREQUENCIES
Chi-square tests
This NON-PARAMETRIC TEST is used to determine the NUMBER OF RESPONSES that FALL IN DIFFERENT CATEGORIES for a single qualitative (lahat ng non-para quali) variable
Chi-square goodness of fit
This NON-PARAMETRIC TEST is used to know the difference between two populations that are HOMOGENOUS in some characteristics
Chi-square of homogeneity
This NON-PARAMETRIC TEST is applied to categorical data to evaluate how likely it is that ANY OBSERVED DIFFERENCE between the sets AROSE BY CHANCE
Pearson chi-square test
The pearson chi-square test is also known as
chi-square of correlation
Considered the nonparametric alternative to the independent t-test
Mann-Whitney U test
Non-parametric equivalent of the paired samples t-test
Wilcoxon test
Non-parametric equivalent of one-way independent measures ANOVA
Kruskal-Wallis H test
Non-parametric alternative for one-way repeated measures ANOVA
Friedman test
Formula for Z score
X1 + x bar / SD
Basic statistics is divided into two:
Descriptive and Inferential
The 3 requirements of a parametric test
1) Normally distributed
2) Homogenous variance
3) Interval or ratio data
The 3 requirements of a non-parametric test
1) Doesn’t need normal distribution
2) Doesn’t need a homogenous variance
3) Nominal or ordinal data