Kin 371 Final
Abstract contains
1. Start with general statement about the relationship of key study variables 2. Problem statement 3. How this study attempts to address problem 4. Purpose statement 5. Conclusion statement (payoff statement to why this research matters
Non-parametric tests
Mann‐Whitney U test -(independent ttest) •Wilcoxon match‐pairs signed‐ranks test (dependent t‐test) •Kruskal‐Wallis ANOVA by ranks (ANOVA) •Friedman two‐way ANOVA by ranks (RM ANOVA) •Spearman rank‐difference correlation-(Pearson's r
Chi squared uses what type of data
Nominal
Three parts of reliability
Observed score True score Error score
Ways of presenting literature
Oral presentation Poster presentation
Effect size can be calculated by
Pearson's R - Correlational studies Cohen's D - Experimental studies Predicted variance r^2 - correlation / regression
alpha level (level of significance)
Probability of chance occurrence •Set beforehand •"level of significance" •Usually set at .05 or .01 -Magnitude of Type I error
Type 1 error
Rejecting null hypothesis when it is true
Repeated Measures ANOVA
Same people complete a measurement of DV 3+ times
Poster presentation cons
Sometimes viewed as less prestigious •Can get "buried" in the poster room •May stand there hours without any good questions •Hard to fit everything •Many question
Factor analysis
Statistical technique used to group similar sets of data into basic components
Before your study you need to know
Statistical test (test distribution) •Expected Effect size (or previously published ES) •αlevel •Power (1 - B)
split-half correlation method
Taking the entire test as a whole and then the test is divided into halves even #s on one half and odd #s on 2nd half). The correlation between the half scores yields a reliability coefficient. Can be used for essay tests and attitude measures
Chi-square test
Testing the observed versus the expected Procedure for rank order data
Parametric data assumes
homogeneity of data. normal distribution curve
Variability
in a set of numbers, how widely dispersed the values are from each other and from the mean
1 IV 2 groups with 1 DV
independent t test Dependant t test
multivariate analysis of variance (MANOVA) aims to answer what question?
is there a difference among groups 1, 2, or 3 on more than 1 DV?
T-tests aim to answer what question?
is there a difference between group 1 and 2 on y?
Pearson's Correlation Test
linear relationship interval or ratio data
negative skewness
mean < median most of the data is at the end of the graph
positive skewness
mean > median most data is at the beginning
central tendency
mean, median, mode
If chi square statistic is smaller than the critical value....
fail to reject the null hypothesis
type 2 error
failing to reject a false null hypothesis
Research proposal should NOT be talked about in...
first person
to do a prediction of a regression line they must be....
found to be correlated
What tense is used in the method
future
univariate analysis
1 DV 1+ IV
Sample size for qualitative study
1 to 20
1 IV 3 groups 1 DV
1 way ANOVA 1 way ANCOVA
Multiple regression aims to answer what question?
Does x1, x2, x3, predict y?
purpose statement is located at the...
end of the introduction
Measurement always has...
error
Chi square is used when there is...
expectations
beta value (beta coefficient)
- Magnitude of type 2 error - as alpha increases beta value decreases
Post Hoc Testing
- conducted after you do an ANOVA - does multiple pair-wise comparisons to determine differences among groups - A vs B A vs C B vs C
Coefficients range from
-1 - 1
Oral presentation cons
-Any question is fair game •Technology fails when you wish it would not •More anxiety •Not in control of time frame•Less interaction with people
Confidence intervals
-Based on the fact that measurements have errors -usually set at 95% -need a critical value CI = mean +or- (SE x critical value)
Poster presentation pros
-Can view a large number at once •Go at your own pace •Great way to highlight the main points •Personal interactions •Can be less stressful
Oral presentation pros
-Practice of presentation skills •Show your research to a good/large audience •In control of what is presented •Seen as prestigious
Probability Sampling Methods
-Random selection -Stratified random sampling -systematic sampling
Introduction contains
-What is not known in the research -Purpose statement -Hypothesis -NO direct quotes
Power measures what? How is power measured?
-probability of rejecting the null hypothesis when the hypothesis is false -Ranges from 0-1 - (1 - B value)
goal is to hit an effect size of
.5
effect size values (meaningfulness)
0.2 = somewhat meaningful 0.5 = educationally meaningful 0.8 = clinically and practically meaningful
4 types of validity
1.Logical 2.Content 3.Criterion 4.Construct
Sample size for small survey
100-250+
T tests always have
2 levels of IV's
Multivariate
2+ DVs 1+ IV's
canonical correlation
2+ X variables and 2+ Y variables
Sample size for experimental study
30 per group
normal distribution curve percentages
34.1%, 13.6%, 2.1%, .1%.
Sample size for correlational study
75-200+
Sample size for large survey
800+
Example: on average the VO2 increased in the participants t(58) = 3.27 p < .05 What does this mean?
95% chance that the data will be replicated 5% chance that the data was due to chance
Inter-observer agreement
A measure of reliability of observers; the degree to which two or more observers concur that specific events or behaviors have occurred.
nonparametric statistics
A statistical method wherein the data is not required to fit a normal distribution. Often used ordinal data. Good for ranks. Inferential.
2+ IVs and 1 DV
ANOVA ANCOVA
Analysis of Covariance (ANCOVA) aims to answer what question?
After accounting for variable c, is there a difference among 1, 2, or 3 on Y?
ANOVA (analysis of variance) aims to answer what question?
Are there differences between groups 1, 2, and 3 on y Determine difference in 3+ groups
Contingency table example
Athletes and nonathletes respond to: "A baseball player who traps a fly ball between the ground and his glove should tell the umpire that he did not catch it" Agree, disagree, or no opinion •450 participants in total
intraclass correlation
Can asses change over time Error variance and systematic variance inter-tester reliability
The accuracy with which a 12‐min run estimates maximal oxygen consumption (VO2 max) in a group of male high school seniors represents a. logical validity b. content validity c. construct validity d. concurrent validity
Concurrent validity
criterion validity
Concurrent: correlate with a gold standard •Predictive: do scores predict some criterion?•Example: ACT, SAT, and GR
Comparing test items with the course objectives (course topics) checks which type of validity? a. content b. predictive c. concurrent d. construct
Content
Multivariate has multiple...
DV's
Objectivity is the...
Degree to which different testers obtain similar scores on the same participan
construct validity
Degree to which scores assess a hypothetical construct •Known group difference method •Correlation•Confirmatory factor analysis
Method contains
Design Participant description Measures Procedure Planned analysis
Dependant T tests
Do two related groups differ?
Independent t tests
Do two sample means differ from each other?
Logical Validity (face validity)
Does the test appear to measure what it claims to measure? •Test is valid by definition
Content validity
Does the test appear to sample the course appropriately? •Used with class examinations and attitude instruments
Regression aims to answer what question?
Does x predict y?
MANOVA example
Example: Among college students (Freshman, Sophomore, Junior, Senior), do they differ on sleep, minutes of PA, and minutes of sedentary behavior?
Analysis of covariance example?
If we control for motor proficiency is there a difference in PA levels among 3rd, 4th, or 5th graders?
Inference Statistics
Inference is the generalization of data to a bigger group
Objectivity shows which type of reliability?
Inter tester reliability Meaning it should be replicated no matter which person is giving the test
Correlation data aims to solve what two questions?
Is there a relationship between X and Y? Is there an association between X and Y?
standard error of measurement
The standard deviation of test scores you would have obtained from a single student who took the same test multiple times Example: •Mary's body fat percentage score was 20.1% •SE of measurement = 2.4% •We are 95% confident that Mary's true fat percentage will fall between 15.3 and 24
strength of a relationship
True variance/ error variance (r^2)
contingency table
Two‐way classification of occurrences and groups used for computing the significance of differences between observed and expected score
Multiple t tests increase what?
Type 1 error
Contingency table is used for...
a chi square with two or more categories with two or more groups
standard error of estimate
a measure of variability around the regression line - its standard deviation
multiple regression
a statistical technique that includes two or more predictor variables in a prediction equation 2 or more IV's and 1 DV
Cohen's d
a value that is the difference of means divided by the pooled standard deviation
ratio variable
a variable that meets the criteria for an interval variable but also has a meaningful zero point Temperature
interval variable
a variable used for observations that have numbers as their values; the distance (or interval) between pairs of consecutive numbers is assumed to be equal. no meaningful zero
bivariate correlation
an association that involves exactly two variables
Prediction equation
an equation suggested by the points of a scatter plot that is used to predict other points
Cronbach's coefficient alpha
average of all possible split-half correlations
nominal
categorical data
Probability is statistics that is expressed by
confidence interval usually p < .05
non probability sampling methods
convenience, quota, snowball/referral
omega squared measures
determines the strength or magnitude of the difference in the mean scores Meaningfulness
Factorial ANOVA aims to answer what question?
do two or more IV's with multiple groups differ from a DV?
Line of best fit is found by
drawing a line that has the least sum of squares
more participants =
more power
If chi square statistic is larger than the critical value...
null hypothesis is rejected
Spearmans p uses what type of data
ordinal data
effect size
quantitative measure of the strength of your findings
Frequently populations are not______ so we use______ to justify
randomly selected, post hoc
Oridnal
ranked data (racers place when they finish)
Statistical significance in a linear data set is influenced by
sample size
unit of analysis
specifies whether data should be collected about individuals, households, organizations, departments, geographical areas, or some combination. school mean vs. student mean
structural equation modeling
statistical method that attempts to establish relationship to another hypothesis/DV through data that proves one DV
homogeneity of variance
the assumption that the variances are equal for the two (or more) groups you plan to compare statistically
interclass correlation
the most commonly used method of computing correlation between two variables; also called Pearson r or Pearson product moment coefficient of correlation
Assumptions to generalize are fueled by
theory and previous research