Research and EBP Flashcards
a ____ is about what will happen if the hypothesis does not come true
null hypothesis (pg.92)
example:
hypothesis: X will be greater than Y
null hypothesis: X will not be greater than Y
what are the 4 principles that the ASHA code of ethics reflects?
- the welfare of persons
- professinal competence
- responsibilities to the public
- responsibilities to the profession
(pg. 92)
what is the PICO format question?
p=population of interest I=intervention C= comparison O=outcomes (pg.93)
example:
in stroke survivors with verbal anomia (P), does training in the use of gestures (I) as compared to no gesture (C) facilitate word retrieval (O)
____ is how well the study is testing or describing what it purports to be testing or describing
internal validity
external validity
internal validity (pg. 93)
____ is the generalizability of results
internal validity
external validity
external validity (pg.93)
the strongest treatment evidence comes from what kind of study?
meta-analysis (pg.93)
what type of study is based on multiple randomized controlled clinical studies and also from systematic reviews
meta-analysis (pg.93)
the generalizability of the study results is enhanced when the subject sample is large or small?
large (pg.94)
and randomly selected from the population of interest
the _______ can be described as a condition that can change behavior
independent or dependent variable?
independent (pg.94)
example:
what is the difference in participants ability to recognize speech under conditions of high versus low noise level?
independent= noise level
the ____ can be described as the behavior that may change
independent or dependent variable?
dependent (pg.94)
example:
what is the difference in participants ability to recognize speech under conditions of high versus low noise level?
dependent= ability to recognize speech
what are the 4 different levels of measurement?
- nominal (identity)
- ordinal (identity and magnitude)
- interval (identity, magnitude, and equality)
- ratio (identity, magnitude, equality, true zero)
(pg. 94)
______ referes to how much we can depend on a measure
reliability (pg.94)
how can reliability be measured ?
- test-retest method
- comparing perfomance or alternate or parallel forms of a test
- internal consistency of measurement (split-half method in which performance on one half of the test is compared to performance of the other half from the same test)
(pg.94)
what are potential sources of errors when taking measurements?
- systematic errors
- intermittent equipment malfunction
- day-to-day challenges
- the behavior of the researcher
(pg.94)
True/False: a small standard error suggest higher reliability?
True (pg.94)
what are the 2 ways of estimating measurement root associates with the researcher?
- interobserver agreement=how consistent 2 or more researchers are in making a particular measurement
- intraobserver agreement= how consistent one researchers is when making the same measurement more than once
(pg. 94)
_____ refers to how well the test items measure the characteristics or behaviors
content, criterion, or construct validity
content (pg.95)
____ refers to how well the measure correlates with an outside criterion that is known to be a good indicator of the characteristic or behavior
content, criterion, or construct validity
criterion (pg.95)
____ refers to how well the measure neglects a theoretical construct of the characteristic or behavior of interest
content, criterion, or construct validity
construct (pg.95)
what are possible threats to internal validity?
- subject selection
- history
- reactive pretest
- statistical regression (if selected cuz of extreme scores, their scores on subsequent administration of the same measure may be closer to average even with no treatment)
- researcher bias
- test environment
- Hawethorne effect (subject’s awareness that they are in a research study)
(pg.95)
in ________ research, data are obtained using numerical measures
quantitive (pg.95)
________ research, the researcher manipulates one or more independent variables and observes the effect on the dependent variable
experimental, descriptive, or mixed
experimental (pg.95)
_______ research, it is not possible for the researcher to manipulate the independent variable because it is a subject characteristic
experimental, descriptive, or mixed
descriptive (pg.95)
______ research is often used in speech-language pathology
experimental, descriptive, or mixed
mixed (pg.95)
example: in a study measuring the dependent variable or oral reading accuracy, the descriptive component may be the comparison of 2 groups based on subject attributes and the experimental component may be the manipulation of an independent variable
_______ design involves comparison of 2 or more groups of subjects
between-subject or within subject
between subject (pg. 95)
in ___________ design, the behavior of the same subjects is studied under different conditions
between-subject or within subject
within subject (pg.96)
______ research involves exploration of factors that may underlie behavior
quantitative or qualitative
qualitative (pg.96)
Interval or ratio measurement data may be described in terms of….
- central tendency
- variability (how much scores vary from the average)
- skewness (lack of symmetry of the distribution scores)
- kurtosis (the general shape of the distribution of scores)
(pg.97)
central tendency can be measures in what 3 ways?
- mean= the arithmetic average
- median=the middle score of he distribution
- mode= the most commonly occurring score
(pg.97)
what are the measurements of variability?
- range= the lowest score to the highest score
- variance= how far each score in the distribution varies from the mean score
- standard deviation= the average amount that all the cores in the distribution deviate from the mean
(pg. 97)
a small standard deviation indicates more or less homogeneity?
more (pg.97)
when does a normal distribution of data occur?
when the middle scores occur more often and the lower and higher scores do not occur often. it results in a bell shape
(pg.97)
when can parametric statistics be used?
- normal distribution of data
- interval or ratio level of measurement
- if 2 or more data distributions will be analyzed and compared
(pg.97)
when are nonparametric statistics used?
when paramedic statistics cannot be used (pg.97)
what are the 2 types of errors that can be made in research?
type I= a true null hypothesis is rejected
type II= a false null hypothesis is accepted
the level of significance is….
A. the probability of making a type I error
B. the probability of making a type II error
C. the probability of making a type I or type II error
A. the probability of making a type I error
pg.97
when the level of significance is ________ then the researcher usually decides to reject the null hypothesis and therefore to decide that the hypothesis is probably true
small (pg.97)
what number is considered small enough to reject the null hypothesis?
a level of significance of .05 or less (p < .05)
pg.97
what is a one tailed test and two tailed test?
- one tailed is directional, example of=X is greater than Y
- two tailed is nondirectional, example of =X and Y are different
(pg. 97)
which are stricter, one tailed or two tailed test?
two tailed (pg.97)
the _____ in the data should be reported so that the results can be interpreted.
- degree of freedom
- generally df=n-1, when n= the total number of scores in the data distribution
(pg.97)
what do correlational statistics evaluate?
the relationship among data (pg.97)
what are examples of correlational statistics?
- pearson product-moment (r), parametric, used with interval or ratio level data
- spearman rank-order correlation (p), nonparametric, used with ordinal level data
(pg.98)
a perfect positive relationship between 2 variables is indicated by _____, a perfect negative relationship between 2 variables is ____ and the absence of a relationship is _______
- 0, -1.0, 0
pg. 98
a ___ number indicates a weak relationship between 2 variables and a ___ number indicates a strong relationship between numbers
small, large (pg.98)
the square of the correlation coefficient is used for what?
-to assess practical meaning (pg.98)
what is regression analysis?
measures the degree to which the value of one variable can be predicted from the value of other variables (pg.98)
for nominal level data, associations between variables can be presented in a _____ table
contingency (pg.98)
the level of significance of any relationship among the nominal variables can be examined using ____ or _____
chi-square, contingency coefficient (pg.98)
-chi-square does not indicate the strength of the relationship whereas the contingency coefficient does measure the strength of the relationship
parametric procedures (single dependent variable, difference between 2 sets of data)
- z ratio=when sample are 30 or more
- t-test= when samples are less than 30
- independent t-test= used to compare 2 different groups
- dependent t-test= used for within group comparisons
(pg.98)
nonparametric procedures (single dependent variable, difference between 2 sets of data)
- mann-whitney U Test= for differences between groups; for ordinal level data
- Wilcoxon matched-pairs single-ranks tests= for examining changes within the group over time
(pg. 98)
parametric procedures (single dependent variable but more than 2 sets of data)
- ANOVA= allows simultaneous comparisons of several means; yield an F ratio
- a one-way ANOVA= used when there is only 1 independent or classification variable
- a 2 way ANOVA= used when there are 2 independent variables
(pg. 98)
nonparametric procedures (single dependent variable but more than 2 sets of data)
- Kruskal-Wallis=one-way ANOVA by ranks
- Cochram Q test=for no nominal level data from related samples
- friedman 2way ANOVA by tanks
- Chi-aqure test for independent samples
(pg. 98)
what to use when there is more than one dependent variable and one or more independent variables, the analysis procedures used are:
- MANOVA= multivariate analysis of variance
- ANCOVA= analysis of covariance
- multiple t-tests can be used, however the Bonferroni correction may be needed
(pg. 98)
_______ can help the SLP to understand the practical significance of data in a research study
effect size estimates (pg.98)
-one common effect size estimate is Cohen’s d, used to compare the means of 2 or more groups
_______ is often completed to measure the probability of rejecting a null hypothesis when the null hypothesis is false
power estimate ([g.99)
-generally, a minimum value of .80 is a acceptable measure of power
______ research is aimed at demonstrating the benefits of treatment through well-controlled studies with internal validity, statistical significance and practical significance
treatment efficacy, treatment effectiveness, treatment fidelity
treatment efficacy (pg.99)
____ is demonstrated when there is clinical improvement from the treatment when applied in the real-world context
treatment efficacy, treatment effectiveness, treatment fidelity
treatment effectiveness (pg.99)
______ is the degree to which actual implementation of the treatment in the real world is consistent with the prototype treatment administered in he controlled conditions of the treatment efficacy study
treatment efficacy, treatment effectiveness, treatment fidelity
treatment fidelity (pg.99)
the strongest evidence of treatment efficacy comes from…
- meta-analysis (pg.99)
2. systematic review
In a ______, accumulated evidence from multiple studies is analyzed statistically to evaluate the consistency of results and effect sizes across studies
meta-anayisis or systematic review?
meta-analysis (pg.99)
__________ are objective and comprehensive overviews of research focused on a particular clinical issue.
meta-anayisis or systematic review?
systematic review (pg.99)
_______ refers to how well the test detects that a condition is present when the condition actually is present
sensitivity or specificity
sensitivity (pg.100)
_____ refers to how well the test detects that a condition is not present when the condition actually is not present
specificity (pg.100)
_______ the number of true positives divided by the combined true and false positives
positive predictive value or negative predictive value
positive predictive value (pg.100)
______ the number of true negatived divided by the combined true and false negatives
negative predictive value (pg.100)
True/False: EBP considers how research out-comes data may apply differently to ethnic minority groups that were not included in the research studies
true (pg.100)
How can SLP’s obtain information about a client’s preference and values?
a cultural assessment (e.g interview, questions) (pg.100)
