Chapter 14: Quantitative Data-Statistical Analysis Flashcards
Things to keep in mind for this LOOOOONG chapter:
1- you will not be asked to calculate any of the parameters described
2- Research studies are not published and peer-reviewed if they used the wrong type of statistical test or measure to analyze their study.
Quantitative Studies:
Levels of Measurement (5):
1) Nominal
2) Ordinal
3) Interval
4) Ratio
5) Continuous variable
Quantitative Studies: Levels of Measurement
_______ measurements are two values that can be accurately described by their ratio; the highest level, and the numbers included are able to have a meaningful zero.
Give an example:
Ratio
Example: Weight
A 200-pound person is clearly twice the weight of a 100-pound person.
Quantitative Studies: Levels of Measurement
______ measurements rank people on attributes with a number that clearly specifies a distance between the two.
Give an example:
Interval
Example:
An IQ test
Quantitative Studies: Levels of Measurement
_______ measurements rank people based on an attribute, such as ADLs.
Give an example:
Ordinal
Example: ADL measurements
4 = being independent
1 = being completely dependent
The number relates to the rating of an attribute.
Quantitative Studies: Levels of Measurement
_________ are variables with interval and ratio measurements.
Continuous variables
Quantitative Studies: Levels of Measurement
_______ measurement is the lowest level and involves using numbers to designate attributes. In other words, things are indicated by a numeric value, but that numeric value can’t be used mathematically.
Give an example:
nominal
Example: Eye color
Brown can be 1
Blue can be 2
Green can be 3, etc.
Quantitative Studies:
_________ Statistics are used to synthesize and describe data.
________ are calculated values such as averages and percentages.
_______ is a descriptive index from a sample.
Descriptive statistics
Parameters
Statistic
Quantitative Studies: Descriptive Statistics
Name the 4 Types of Descriptive Statistics
1) Frequency distribution
2) Symmetric distribution
3) Skewed distribution
4) Normal distribution
Quantitative Studies: Descriptive Statistics
A ______ distribution is a bell-shaped curve that has a lower peak than other types of distributions.
SD within 1?
SD within 2?
normal
For a normal distribution or bell-shaped curve,
68% of values will be within 1 SD of the mean.
95% of values will be within 2 SD of the mean.
Quantitative Studies: Descriptive Statistics
________ distributions are ones where if the graph is folded in half the two halves would be superimposed.
Symmetric
Quantitative Studies: Descriptive Statistics
In a _______ distribution, the majority of the data peaks to one side.
skewed
Quantitative Studies: Descriptive Statistics
A _______ distribution takes a list of values and organizes it lowest to highest and includes a count.
Give an example:
frequency
Example:
If there were 10 values and 5 of them were 80, it would list 80 once and put a count of 5 and a percentage of 50%.
Quantitative Studies:
______ ________ includes methods to determine a central value for a set of numbers.
This includes _____, _____, and ______.
Central tendency
Mode, Median, and Mean
Quantitative Studies: Central Tendency
The ____ is the number that occurs most frequently in a distribution.
mode
Quantitative Studies: Central Tendency
The ______ is the point in a distribution that divides scores in half.
median
Quantitative Studies: Central Tendency
The _____ equals the sum of all values divided by the number of participants.
mean
Quantitative Studies: Central Tendency
Measures of central tendency can be the same for two different sets of distributions.
This is known as _______.
variability
Quantitative Studies: Variability
The _____ is the highest minus the lowest score in a distribution.
range
Quantitative Studies: Variability
The ___________ is a variability index calculated based on every value in a distribution; it is the average amount of deviation of values from the _____.
standard deviation
average amount of deviation of values from the MEAN
Quantitative Studies: Variability + Standard Deviation
The lower the standard deviation, the more _________ the distribution.
homogenous
What Human attributes occur as bell-shaped curves?
height and intelligence
Quantitative Studies: Bivariate Descriptive Statistics
If a frequency distribution is based on two variables instead of one, a _____________ can be used to depict the data.
crosstabs crosstable (Crosstabulations) -Two-variable descriptive statistics
Correlation
Correlation coefficient
Quantitative Studies: Bivariate Descriptive Statistics
Crosstabulations are used for what 2 kinds of data?
Used with nominal or ordinal data
The most widely used correlation statistic is Pearson’s r (the product-moment correlation coefficient).
Spearman’s rho is a correlation coefficient calculated for values on an ordinal scale.
A correlation matrix displays correlation coefficients in tables displaying rows and columns.
Quantitative Studies: Bivariate Descriptive Statistics
_________ are methods to describe relationships between two variables, and utilizes the ________ _________, which is the intensity and direction of the relationship between two variables.
Correlations
correlation coefficient
Quantitative Studies: Bivariate Descriptive Statistics
Correlation coefficients range from _________.
range from -1.0 to 0 to 1.0.
Quantitative Studies: Bivariate Descriptive Statistics + Correlation coefficients:
A +1.0 would mean a _______ relationship.
Values b/t 0 and +1.0 represent a _____relationship.
Values b/t -1 and 0 represent a ______ relationship
Perfect relationship
Positive relationship
Negative relationship
Quantitative Studies: Bivariate Descriptive Statistics
Name the 2 Types of Correlation Coefficients:
Which one is the most widely used correlation statistic?
1) Pearson’s r –> MOST WIDELY USED
2) Spearman’s rho
Quantitative Studies: Bivariate Descriptive Statistics + Correlation Coefficients:
The product-moment correlation coefficient =
Pearson’s r –> MOST WIDELY USED
Spearman’s rho is a correlation coefficient calculated for values on an ordinal scale.
Quantitative Studies: Bivariate Descriptive Statistics + Correlation Coefficients:
________ is a correlation coefficient calculated for values on an ordinal scale.
Spearman’s rho
Quantitative Studies: Bivariate Descriptive Statistics + Correlation Coefficients:
What is used to display correlation coefficients in a table with rows and columns?
A correlation matrix displays correlation coefficients in tables displaying rows and columns.
Quantitative Studies: Describing Risk
________ risk is the proportion of people who experienced an undesirable outcome in each group.
How do we calculate this and what is it called?
Give an example:
Absolute risk
Absolute risk reduction is a comparison of the two risks.
Example:
one group of people had education on falls and are the experimental group, and the control group did not have education on falls.
10% of individuals in the experimental group had falls, while 30% of the individuals had falls in the control group.
Experimental group AR = 0.10
Control group AR = 0.30
Calculation:
(Experimental group AR) - (Control group AR) x100 =
.30 - .10 = .20 or 20%.
This suggests that 20% more of the control group would not have fallen with the intervention.
Quantitative Studies: Describing Risk
_________ is the ratio of two odds, and gives us the odds of an undesirable effect.
How do we calculate this?
Give an example:
Odds ratio
Calculation:
(# who didn’t fall) / (# who did fall) = odds ratio
Example:
10% of individuals in the experimental group had falls, while 30% of the individuals had falls in the control group.
For the experimental group,
70 / 30 = 2.33
For the control group,
90 / 10 = 9.00
Odds ratio = 2.33 / 9 = 0.259
This means the odds of falling with education is 25%, and the odds of falling without education is 400% more likely.
Quantitative Studies: Describing Risk
The number needed to treat estimates is ……?
How we calculate this?
Give an Example:
how many people need to receive an intervention to prevent one undesirable outcome
Calculation:
1 / Absolute risk reduction
Example:
1 / .20 = 5
Five people would need to receive education on falls.
Quantitative Studies:
________ statistics use the law of probability to test research hypotheses with data.
Use ____ to test research hypotheses.
Inferential statistics
Use data to test research hypotheses.
Quantitative Studies: Inferential Statistics
A confidence interval relates to the probability of being right. A confidence interval of __%-__% is the goal for most researchers.
confidence interval of 95%-99%
GOAL FOR US:
To understand that in a research study, seeing the researchers describe obtaining a confidence interval of 95% or 99% indicates a high probability of the data being right.
Quantitative Studies: Inferential Statistics
______ ______ of the Mean is the standard deviation of the mean of a sample.
The higher the number –> ?
What sample size (small or large) will help reduce the chance of outlying data?
Standard Error of the Mean (SEM)
The higher the number –> the more error associated with the sample selected.
- In other words, a lower SEM reflects a more appropriate sample size.
LAAAAARGE
Quantitative Studies: Hypothesis Testing
Hypothesis testing uses ______ criteria to determine if the hypothesis is supported or rejected.
uses objective criteria
Quantitative Studies: Hypothesis Testing
______ ________ means the results are not likely to be due to chance.
What Standard Deviation value indicates this?
What Confidence Interval (%) indicates this?
Statistical significance
Falls within 2 standard deviations of the mean
Confidence intervals of 95% or greater represent statistically significant values.
Quantitative Studies: Hypothesis Testing
A _______ result means that the results could be due to chance.
nonsignificant
Quantitative Studies: Hypothesis Testing
A ____ hypothesis is when there is no relationship between the independent variable and dependent variable.
null hypothesis
Quantitative Studies: Hypothesis Testing
A type __ error is rejecting the null hypothesis that is in fact true.
Type __ error assumes the null hypothesis is true when in fact the independent variable did have an effect.
Type I
Type II
Quantitative Studies: Hypothesis Testing
A power analysis is calculated to determine the chance of a type __ error.
Type II = power analysis
Quantitative Studies: Hypothesis Testing
Alpha, or level of significance, is calculated to determine the chance of a type __ error.
Type I error = alpha
Quantitative Studies: Hypothesis Testing
Is it possible to eliminate Type I and Type II errors?
No, IMPOSSIBLE to eliminate type I and type II errors
So, statistical tests will aim to be statistically significant.
Quantitative Studies: Specific Statistical Tests
Name the 3 Specific Statistical Tests:
What value determines Statistical significance for all 3 tests?
1) T-tests
2) Analysis of Variance (ANOVA) test
3) Chi-Squared Test (X^2)
A p value of <0.05 indicates Clinical Significance
Quantitative Studies: Specific Statistical Tests
T-tests, ANOVA tests, and Chi-squared tests use _____ to determine Clinical Significance.
p value
<0.05 = Clinical Significance
Quantitative Studies: Specific Statistical Tests
_________ is used to test mean group differences of three or more groups.
Calculates a __ value, which varies for each study.
Analysis of variance (ANOVA) test
ANOVA calculates a F value
Quantitative Studies: Specific Statistical Tests
A parametric test for testing differences in two group means is called a _______.
Calculates a __ value, but the individual value is not important.
T-test
Calculates a t value, but the individual value is not important.
Quantitative Studies: T-test
_______ t-test is appropriate for two different groups of people.
_______ t-test is used when one group is tested at two different points in time.
Independent t-test
Paired t-test
Quantitative Studies: Specific Statistical Tests
A _______ test focuses on the difference in proportion.
Calculates the value of ______, which varies from test to test.
Give an example:
chi-squared test
Calculate the value of chi-squared, which varies from test to test.
If blood sugars were controlled in 60% of patients with a diabetes educational intervention, versus 40% of patients without, is the 20% real?
Quantitative Studies: Multivariate Statistical Analysis
Name the 3 Types of Multivariate Statistical Analysis Tests:
1) Multiple Regression
2) Analysis of Covariance (ANCOVA)
3) Logistic Regression
Quantitative Studies: Multivariate Statistical Analysis
_________ is a combination of ANOVA and multiple regression.
This test is good when there is not control through _______.
ANCOVA attempts to control _________.
Analysis of Covariance (ANCOVA)
ANCOVA is good when there is not control through randomization.
ANCOVA attempts to control covariates of confounding variables.
Quantitative Studies: Multivariate Statistical Analysis
___________ measures several independent variables, called ______ variables in this type of test.
A statistic called a _____________ is calculated and known as R.
Multiple Regression
Predictor variables
multiple correlation coefficient (R)
Quantitative Studies: Multivariate Statistical Analysis
___________ measures several independent variables, called ______ variables in this type of test.
A statistic called a _____________ is calculated and known as R. R varies from _________.
Multiple Regression
Predictor variables
multiple correlation coefficient (R)
R varies from 0.0 to 1.00.
Quantitative Studies: Multiple Regression
________ gives a % of how influential the predictor variables were.
A higher R-squared value means –> ?
R-squared
A higher R-squared value means –> it is more likely the predictor variables accounted for the variation.
Quantitative Studies: Reliability Measurements
What are the 3 Types of Reliability Tests?
Indicate what each test calculates:
For all of these, scores closer to ___ means a higher level of reliability.
1) Test-Retest Reliability
- Intraclass correlation coefficient (ICC)
2) Interrater reliability
- Cohen’s kappa
3) Internal consistency reliability
- Coefficient alpha or Cronbach’s alpha
For all of these, scores closer to 1.0 means a higher level of reliability.
Quantitative Studies: Reliability Measurements Intraclass correlation coefficient is used for \_\_\_\_\_\_\_ reliability and ranges from 0.00 to 1.00.
test-retest reliability
Quantitative Studies: Reliability Measurements
Cohen’s kappa is used for dichotomous classifications and is used for _______ reliability; helpful to determine if two individuals would rate something similarly.
Interrater reliability
Quantitative Studies: Reliability Measurements
Coefficient alpha aka Cronbach’s alpha is used for __________ reliability, which measures how often components of a multi-component tool measure the same attribute
Internal consistency reliability
Quantitative Studies: Validity Assessment
Name the 3 Types of Validity
1) Content Validity
2) Criterion Validity
3) Construct Validity
Quantitative Studies: Validity Assessment
_____ validity determines if the content of the items adequately reflects the construct of interest.
An index of ___ or higher indicates good validity.
Content Validity
- Experts meet to rate each item to determine a content validity index.
An index of 0.90 or higher indicates good content validity.
Quantitative Studies: Validity Assessment
______ validity concerns the extent to which a measure is truly measuring the target construct. It uses various procedures such as Pearson’s r value, or an independent group t-test.
Construct validity
Quantitative Studies: Validity Assessment
______ validity concerns the extent to which scores on a measure are consistent with a gold standard criterion.
Two terms fall under this type of validity, What are they?
Criterion validity
1) Sensitivity
2) Specificity
Quantitative Studies: Criterion Validity Assessment
________ is the ability of a measure to correctly screen or diagnose a condition.
Sensitivity
Think: SENSITIVITY = HOW MANY TIMES DID I GET CLOSE TO THE TARGET
the ability of a test to correctly identify patients WITH a disease.
Quantitative Studies: Criterion Validity Assessment
________ is the measure’s ability to correctly identify non-cases or screen out those without the condition.
Specificity
Think: SPECIFICITY = HOW CLOSE TO A TARGET
the ability of a test to correctly identify people WITHOUT the disease
Quantitative Studies: Criterion Validity
Sensitivity vs Specificity using Walter’s Example of a flu swab diagnostic test:
A specific flu swab would….
A sensitive flu swab would…
A specific flu swab would NOT be positive for strep or other conditions.
A sensitive flu swab would be positive for the flu MOST of the time.
Quantitative Studies: Criterion Validity
Sensitivity vs Specificity using Walter’s Example of a flu swab diagnostic test:
What if a flu swab always caught the flu, but also would show positive for strep?
highly sensitive, but has LOW specificity
Quantitative Studies: Criterion Validity
Sensitivity vs Specificity using Walter’s Example of a flu swab diagnostic test:
What if a flu swab ONLY detected the flu, but only detected half the cases of flu successfully?
highly specific, but has LOW sensitivity
