Research Methods exam 1
Nonparametric Statistical Tests
- 1-Way Chi Square -2-Way Chi Square - Spearman's Rho Rank Correlation - Mann-Whitney U Test - Wilcoxon Signed Rank Test - Kruskal Wallis 1-Way ANOVA - Friedman's Test
Tips for Spotting Correlations
- Relationship (Variables) - Continuous ( Quantitative) - "Related"
if there are 2 or more independent variables, automatically we know it is a _______________
ANOVA (some type)
extension of ANOVA in which main effects and interactions are assessed on dependent variable scores after the DV has been adjusted for by the DV's relationship with one or more covariates
ANCOVA
parametric statistical analysis used for comparing means among MULTIPLE treatment groups; used when there is either more than 2 groups to compare AND/OR more than 1 dependent measure
ANOVA
Statistical Tests that Require Post Hoc Tests
ANOVAs
2 Way ANOVA Between Subjects
Different people test each condition, so that each person is only exposed to a single user interface.
Nested ANOVA
Each group is divided into 2 or more random subgroups. Tests to see if there is variation between groups or within nested subgroups of attribute variable.
ANCOVA (Analysis of Covariance)
Extension of ANOVA where main effects & interactions are assessed on dependent variable scores after the dependent variable has been adjusted for by the dependent variable's relationship with 1 or more covariates.
probability of rejecting at least one true hypothesis, of making at least one type 1 error
Familywise Error Rate (FWER)
nonparametric statistical analysis used to test the equality of more than 2 population probability distribution; equivalent to One Way Repeated Measures ANOVA
Freedman's Rank Test
Repeated Measures ANOVA Nonparametric Test
Friedman Test (Means of 3+ related samples)
A researcher believes college students have better hearing ability than college students. Maximum hearing frequency (measured in megahertz) is measured 200 high school and 200 college students.
Independent T test
nonparametric statistical analysis used when more than 2 groups are being compared; equivalent to one way ANOVA
Kruskal Wallis One Way Anova
1-Way ANOVA (Between Subjects) Nonparametric Test
Kruskal-Wallis Test (Means of 3+ independent samples)
extension of ANOVA in which main effects and interactions are assessed on a combination of dependent variables; tests whether mean differences among groups on a combination of DV's is likely to occur by chance
MANOVA
nonparametric statistical analysis used to compare two samples from two populations; equivalent to an independent t test
Mann - Whitney U Test
Independent T-Test Nonparametric Test
Mann-Whitney U test (Means of 2 independent samples)
Pearson Correlation
Measures the nature & strength between 2 variables of the quantitative type.
A company is designing a new app that uses daily variables collected from its users (resting heart rate, steps/day, exercise minutes and exercise heart rate) to predict aerobic fitness level (VO2max). The company measures these variables in 500 subjects, as well as directly measuring VO2max in those subjects.
Multiple Regression
creates an equation that uses more than one variable to predict the outcome variable (7 SITE SKIN FOLD EXAMPLE)
Multiple Regression
Friedman's Test
Nonparametric alternative to analyze a 1-Way Repeated measures ANOVA.
Spearman's (Rho) Rank Correlation
Nonparametric measure of correlation. 2 sets of ranks that may be assigned to the sample values of x & y. Used when at least 1 value is nonparametric
Wilcoxon Signed Rank Test
Nonparametric measure used to assess differences between paired (dependent) sample.
compares observed frequencies with expected frequencies when there's 1 independent variable (EX: rolling a dice); also known as a "goodness of fit" test
One Way Chi Square (Goodness of Fit)
A researcher wants to measure whether any of 4 different fertilizers (A, B, C, and D) improve grass seed germination time. 20 land plots are given fertilizer A, 20 land plots are given fertilizer B, 20 land plots are given fertilizer C, and 20 land plots are given fertilizer D. Germination time is measured on each plot of land.
One way ANOVA between
A researcher wants to see if a program to improve forearm strength in baseball pitchers will improve ball spin rates (revolutions per minutes). 12 pitchers undergo an 12 week training program. Ball spin rates are measured before the program, after 4 weeks, after 8 weeks and after the program is completed in each pitcher.
One-Way ANOVA within
A researcher wants to examine the effect of caffeine dose on heart rate. 25 groups of subjects are chosen for the study. On Day 1, subjects are given 200mg of caffeine and heart rate is measured 2 hours later. On Day 2, the same subjects are given 400mg of caffeine and heart rate is measured 2 hours later. On day 3, the same subjects are given 600mg of caffeine and heart rate is measured 2 hours later.
One-way ANOVA (within subjects) Caffeine (independent) on heart rate (dependent) 25 groups of subjects (population) Day 1: subjects are given 200mg of caffeine and heart rate is measured 2 hours later Day2: the same subjects are given 400mg of caffeine and heart rate is measured 2 hours later Day 3: the same subjects are given 600mg of caffeine and heart rate is measured 2 hours later. Measured on same subjects: within subjects
A researcher wants to know if the thickness of cardiac muscle around the left ventricle of the heart (measured in mm) is related resting systolic blood pressure (measure in mmHg). Both variables are measured in 200 subjects.
Pearson correlation
Factorial ANOVA
Possible to have 2 or more independent variables
used when you are performing many independent or dependent statistical tests at the same time; allow us to specific exactly where significant differences occur
Post Hoc Adjustments
if a correlation is poor, then there is no point in doing a ______________
Regression
when you have more than 2 dependent t tests, that is equivalent to a _______________________
Repeated Measures ANOVA
1 Way ANOVA (Within Subjects)
Repeated measures design in which at least 1 of the factors consists of repeated measurements on the same subject or experiment. Units, under different conditions, &/or at different time points.
Pearson's Correlation Nonparametric Test
Spearman Correlation (Relationship between 2 continuous variables)
A researcher wants to know whether this year's college freshman at PBAU have lower resting systolic blood pressure than normal. Blood pressure has been measured in college freshman for the last 20 years at PBAU. The researcher wants to examine whether this year's group of freshman differ from all of the freshman ever tested at PBAU.
Student T test
parametric statistical analysis that compares 2 groups on one dependent measure; simple, easy, quick
T test
2 Way ANOVA Within Subjects
The same person tests all the conditions (all user interfaces).
parametric statistical analysis in which at least one of the factors consists of repeated measurements on the same subjects or experimental units, under different conditions and/or at differential time points; basically a paired t test for multiple groups
Two Way ANOVA
compares observed frequencies with expected frequencies when a sample has 2 dimensions of variables; also known as a "test of independence" test
Two Way Chi Square (Test of Independence)
A research wants to measure the effect of glucosamine supplementation on aerobic enzyme activity in mice. One group of mice is given 300mg of glucosamine per day for 14 days. Another group of mice was given a placebo for 14 days. Aerobic Enzyme activity (measured in μg/ml) was measured before and after the 14 days in both groups.
Two Way Mixed ANOVA
Mann Whitney U Test
Used to compare 2 samples from 2 populations, nonparametric.
Independent T-Test
Used when measuring the difference between 2 group means with separate sampling sources.
Dependent (Paired) T-Test
Used when measuring the differences between 2 group means with the same sampling source.
T-Test Student's T Test (Z-Test)
Used when there's only 1 group; want to test against hypothetical or population mean.
Covariate
Variable that is related to the dependent variable which you can't manipulate but want to account for in relationship with the dependent variable
Dependent (Paired) T-Test Nonparametric Test
Wilcoxon Signed Rank Test (Means of 2 related samples)
nonparametric statistical analysis used to assess differences between paired (dependent) sample; equivalent to dependent t test
Wilcoxon Signed Ranked Test
when you have more than 2 groups in an independent t test, that is equivalent to a _______________________
Within Subjects One Way ANOVA
different people test each condition, so that each person is only exposed to a single user interface
Between Subjects Design
MANOVA (Multiple Analysis of Variance)
By measuring multiple dependent variables, you increase chances for finding a group difference.
Chi Square...
No parametric test equivalent
Kruskal Wallis 1-Way ANOVA
Used when more than 2 groups are being compared. Significant treatment effect means at least 1 population median differs from another.
the same person tests all the conditions
Within Subjects Design
Correlation, Regression
_________________ describes the strength of a linear relationship between 2 variables; a ______________ tells us how to draw the straight line described by correlation
time, numbers, etc
parametric
Tips for Spotting Chi Squares
- Comparison (Frequencies) - Categorical (Qualitative)/ Nonparametric - "Frequencies"
Tips for Spotting T-Tests
- Comparison (Means) - Categorical and continuous - "Difference"
Parametric Statistical Tests
- Pearson Correlation - Multiple Regression - T-Test- Student's T-Test (Z-Test) - Independent T-Test - Dependent T-Test - 1-Way ANOVA, Within and Between Subjects - 2-Way ANOVA, Within and Between Subjects - Mixed Model ANOVA - Nested ANOVA - ANCOVA - MANOVA
A researcher examines whether a 3-month peer-mentoring program alters the number of days of detention served by 250 9th graders. Detention days are documented in each subject for 1 month before and 1 month after the peer-mentoring program. The data is not normally distributed.
--
A researcher want to examine whether AP Calculus Exam scores are significantly different in 4 different school districts. AP scores on the Calculus Exam are collected from students in each school district. AP scores are AP tests are scored on a 1 to 5 scale (5 - Extremely well qualified, 4 - Well qualified, 3 - Qualified, 2 - Possibly qualified, 1 - No recommendation)
--
A researcher wants to know if dance training for 10 weeks improves vertical jump height. 15 subjects engage in 10 weeks of dance training 3 days a week. Maximum vertical jump height (in centimeters) is measured before and after the 10 weeks in all subjects.
--
A researcher wants to know whether this year's college freshman at PBAU have lower resting systolic blood pressure than normal. Blood pressure has been measured in college freshman for the last 20 years at PBAU. The researcher wants to examine whether this year's group of freshman differ from all of the freshman ever tested at PBAU.
--
A researcher wants to measure the effect three different herbicides (A, B and C) on seed production of soft white wheat grain (a type of grain). 30,000 grain seeds are randomly treated with either herbicide A, B or C. Time of seed germination in all seeds to see which herbicide results in the highest seed germination times.
--
A researcher wants to see if a new anti-diuretic drug (ADD) affects urine output. One group of subjects receives the drug for 14 days and another separate group does not receive the drug. Urine output (in milliliters) is measured in both groups before and after the 14 days.
--
A researcher wants to see if the honors students have different views of their college experience than non-honors students over 4 years of college. A group of 20 honors (HON) students and a group of 20 non-honors (NON) students are asked to rate their satisfaction with college (on a 1-5 scale). Both groups are asked to give ratings 4 different times (freshman year, sophomore year, junior year, senior year).
--
A researcher wants to use a variety of golf skills to predict golf scoring average. The researcher collects data from 120 golfers. This data includes driving distance (in yards), greens hit in regulation and average number of putts per round.
--
A team of researchers hypothesize that there is an association between gender and risky driving behavior. They observed a number of drivers around West Palm Beach and noted whether or not they displayed any risky driving behaviors (i.e. speeding, dangerous overtaking, going through red lights, etc...). Drivers who displayed such behavior were classified a 'risky' and those who did not display such behavior were classified as 'safe'. The gender of the driver was also noted.
--
Mixed Model ANOVA
1 Variable is between subjects, the other is within subjects.
1 Way ANOVA (Between Subjects)
2 or more groups, 1 dependent variable
The purpose of the study was to analyze muscle contractile speed (using tensiomyography (TMG) of the rectus femoris muscle (RF) using TMG muscle properties after different lower limb strength training protocols (multple sets; DS= drop set; eccentric overload; FW= fly wheel; PL= plyometics). after baseline measurements, 14 male strength trained athletes completed 1 squat training protocol per week over a 5-week period in a randomized controlled order. Muscle contractile speed (Tc) was measured before (pre), 0.5 HR (post-train), 24 hr (post-24), and 48 hr (post-48) after the training interventions
2 way ANOVA within parametric- quantitative
A researcher wants to measure whether electrolyte supplementation alters sweat rate in a group of 20 subjects. Sweat rate (ml/hour) is measured in all subjects both before and 5 hours after supplementation.
dependent T test
Not Significant
if "r" actual < "r" critical, then correlation is __________________
Significant
if "r" actual > "r" critical, then correlation is __________________
parametric statistical analysis used when measuring the differences between 2 group means with separate sampling sources
independent t test
Pearson Correlation
measures the nature and strength between two variables of the quantitive type; PARAMETRIC
non-parametric statistical tests often don't use means, but use __________ instead
median
study design in which one variable is between subjects, and another variable is within subjects
mixed model design
an insurance company attempts to PREDICT fitness level to calculate health risk. in 800 subjects, they collect data on maximum heart rate (beats/min), resting heart rate (beats/min), age (years), weight (lbs), % body fat and measured VO2max (O2/kg/min). data from this group is applied used to estimate VO2max in the entire population
multiple regression whenever the word predict its gon be multiple regression
Spearman's Rho
non-parametric measure of correlation; measure of the strength and direction of the association that exists between 2 ranked variables; ranks are correlated, not actual values
category data (survey)
nonparametric
parametric statistical analysis used for comparing 2 or more groups with 1 dependent variable; most basic form of this category of test
one way ANOVA
a study examines whether the % of 715 fatal bicycle accidents differs based on season (fall, winter, spring, summer). 140 occurs in the fall, 98 occurs in the Winter, 197 occurs in the spring, and 280 occurs in the summer
one way chi square when there is % it is going to be chi square
A single gene in a gene family codes for different isoforms of a protein. In the normal population, these isoforms occur at different rates (isoform A 64.4%, isoform B 18.2%, isoform C 10.5%, isoform D 7.9%). Researchers measure protein isoforms in 90 lung cancer patients to see if the proportion of these isoforms is significantly different. what test would best analyze this data
one- way chi squared observed
Correlation
quantitive measure of the relationship among 2 or more variables; means there is a relationship, not a cause and effect; describes the strength of a linear relationship between 2 variables
Geologists want to assess whether there is a significant relationship. between the number of earthquakes (>2.0 richter scale) measured each day in Trona, CA and the local high temp (*F) each day. the data to be analyzed are as followed
spearman correlation
A researcher want to see is if final exam grade of students in an introductory HHP course is related to the number of classes missed during the semester. These data are not normally distributed.
spearman rho
regression
statistical method based on correlation that allows the prediction of one variable from another; tells us how to draw the straight line described by correlation
parametric statistical analysis used when we have only one group and we want to test it against a hypothetical mean or population mean; not used a ton because we don't have great data from the population
students T test (Z test)
researchers assess the effect of an diaphragm muscle training (DMT) protocol on inspiratory muscle pressure. 23 subjects engage in a 14 week DMT protocol, while 18 subjects in a control group perform no training over the same time. inspiratory muscle pressure is measured before and after the 14 weeks
two way mixed ANOVA
general population
we use correlations to make inferences about the ____________ ____________
A researcher wants to examine the role of highest level of education (high school, college, graduate school) and gender on the number of shoes owned. 400 subjects were divided into separate groups based on these two variables. The number of shoes owns was measured in all subjects.
B/N Subjects Two Way ANOVA
compares observed frequencies within groups to their expected frequencies; allows us to assess differences in frequencies
Chi-square
most common measure of effect size
Cohen's d
2 Way Chi Square
Compares frequencies with expected frequencies when a sample has 2 dimensions of variables.
1 Way Chi Square
Compares observed frequencies with expected frequencies when there's 1 independent variable.
variable that is related to the dependent variable (which can't be manipulated), but you want to account for its relationship with the dependent variable; must be interval data (common examples include age, time, height, weight)
Covariate
ANCOVA's are a recommended experimental research design because the inclusion of ______________, can have the effect of reducing the _______ ______ ______ and therefore, increase the _______ ____ _________
Covariates, Mean Square Error, Power of Analysis
Multiple Regression
Creates equation that uses more than 1 variable to predict the outcome variable. Takes more than 2 variables and determines which are most highly correlated.
parametric statistical analysis used when measuring the differences between 2 group means with the same sampling source; typically used pre vs post
Dependent T Test (Paired Sample/Matched Pair)