Home > Preview
The flashcards below were created by user
arikell
on FreezingBlue Flashcards.

Primary objective of SSD
investigate the cause and effect relationship between the intervention and the effect on outcomes and identify patterns of change

IV of and SSD
intervention

DV of SSD
outcomes (also termed target behavior which is observable, quantifiable, and a valid indicators of treatment effectiveness)

Sample size of SSD
110 subjects

Phase A
baseline (prior to intervention) or withdrawal (after intervention)


Phases CZ
other interventions

Length of phases
consider type of patient, treatment, and expected rate of change in target behavior, use relatively equal lengths and extend the baseline until stability is achieved or true nature of the condition is presented

Requirement for repeated measures
must have consistency and reliability in measuring tools/techniques

Pattern of Reversal (withdrawal) design
ABAintervention is removed after the intervention phase

Internal validity of reversal design
 has strong internal validity compared to regular AB designs
 It demonstrates the causal relationship between effect and intervention

Limitations of reversal design
 ethical concerns with withdrawing treatmentif treatment is effective and you take it away
 subjects may also be unhappy about reversal and perform worse

When can reversal designs not be used?
 they cannot be applied to studies where the IV has strong longterm effects (ex. learning somethingcan't unlearn it)
 baseline will not be able to be attained again

When should multiple baseline designs be used
reversal designs are not feasible due to ethical concerns or the changes by the intervention cannot be reversed

Internal validity of multiple baseline designs
to show that changes occur only when the intervention is introduced no matter when intervention began

Requirements of multiple baseline designs
 repeated measurement of 3 or more data series (can be across subjects, across settings, or across behaviors)
 all baselines must exhibit sufficient stability
 the intervention must be applied to only the first series while other baselines are continued
 first series achieve stability treatment
 intervention introduced to the second series

Limitations fo multiple baseline designs
causal relationships may not be clearly demonstrated, implementation of the study can be time consuming

Multiple Baseline Designs should NOT be used when
there is only one subject in one setting and one target behavior (DV)

Essential feature of Alternating treatment design
rapid alteration of two or more interventions or treatment conditions and the capability to study multiple treatments

Internal validity of alternating treatment design
randomize the order of different treatments can enhance internal validity (advantage)

When to use an alternating treatment design
 treatment effects are immediate
 outcome is a clear consequence of one specific treatment
 outcome is capable of changing quickly
 treatments must be able to trigger those quick changes

When to NOT use an alternating treatment design
when the outcome takes time to change

Results of a single subject study do not involve
any traditional statistical methods

How are results from SSDs analyzed?
via interpretation of graphs (comparisons within each phase and between different phases)

Level (method of visual inspection)
mean DV in each phase (increases or decrease and amount of change based on lines)

Trend (method of visual inspection)
shows the rate of change in behaviors (direction pos or neg, speed, and patterns of changes)

Stability or Variability (method of visual inspection)
stability reflects the consistency of response over time

Timing (method of visual inspection)
observing how quickly the line changes

Slope (method of visual inspection)
shows rate at which the line changes and can also show a positive or negative relationship

Advantages of SSD
 observe change at an individual level
 streamline with clinical practice
 quick to know results and easy for clinicians to interpret them
 relatively inexpensive
 provide control for most threats to internal validity (same subject repeatedly measures and patterns can be compared)

Disadvantages of SSD
 lacks external validityinability to generalize results to other individuals and to other situations/settings
 lack welldefined principles to apply systemic statistical analysis, just visual inspection (can be subjective)

When to use SSD
 when target population is small
 when you want to focus on patterns of changes in DV of individual subjects induced by the intervention
 conduct the study along with clinical practice/in a clinical setting

Descriptive Statistics
used to transform data into an interpretable form and summarize a set of data from your sample

frequency distribution
allows us to begin to analyze the data by looking at the number of time each value occurred or its frequency (often expressed as percentages of the total distribution)

Null hypothesis
generally a statement that assumes that there is NO difference between two sets of values, which is the group means are not different

Alternative hypothesis
there IS a difference between the two sets of values

The purpose of an experiement
to REJECT (disprove) the null hypothesis in the target population, based on the data collected from the sample

P value <0.05
 there is a significant difference between the expected and observed
 REJECT the null hypothesis

P value >0.05
 there is NO significant difference between the expected and observed
 FAIL to reject the null hypothesis

Does the p value indicate magnitude
 NO!!!!! it is just an arbitrary and does NOT tell magnitude or guarantee clinical importance
 Ex. 0.01 not more significant than 0.05

Type 1 Error
 rejecting the null hypothesis when it is actually true
 incorrect rejection of a true null hypothesis
 level of significance (alpha) usually=0.05
 ex. think that an effect or relationship occurs when it doesn't

Type 2 error
 the null hypothesis is not rejected when it is false
 failure to reject a false null hypothesis
 denoted as beta= 1power and the power is normally 0.8

Effect size
a standardized and unites value that can be used to compare intervention effect on different outcome measures

Does effect size tell the magnitude?
YES! it tells the magnitude of treatment effect and helps to interpret clinical significance of results

A p value of <0.05 tells you
if the study is GENERALIZABLE to the population your sample represents

Interpret Cohen's D for continuous outcome
 0.2=small effect
 0.5=medium effect
 0.8=large effect

Dichotomous Outcome
normal variables are normal variables which have only two categories or levels

Number needed to treat
offers a measurement of the impact a medicine or therapy by estimating the number of patients that need to be treated in order to have a beneficial impact on one person

Confidence interval
a range of scores with specific boundaries, that should contain the population mean

95% CI
range of values that you can be 95% certain the true mean of population

If you increase the sample size
the width of the CI will be smaller and you decrease your margin of error

if you increase your confidence level to 99%
the width of the CI increases and you increase your margin of error

center line of forest plot
 indicates not difference between groups
 if CI line passes it then NOT significant

horizontal line of forest plot
95% confidence interval

Diamond of forest plot
 compiled or pooled estimate from all the studies
 location shows overall significance

square size
 weight of the study (related to sample size and quality)
 weight does not influence significance

I^2
 the percentage of observed total variation across studies that is due to real heterogeneity rather than chance (ranges from 0100%)
 Should NOT be larger than 75%
 do not recommend to combine studies with heterogeneity

Power analysis4 variables
power, level of significance, effect size, and sample size

t test
used to compare two means

independent (unpaired) t test
 used when two independent groups of subjects are compared
 independent groups must have independent sets of subjects with no inherent relationship derived

paired t test
used when each subject is exposed to both interventional conditions and subjects used as their own controls

One tailed test
 if we are only interested in one side of the probability distribution
 we are testing whether one mean is higher than the other

two tailed test
if we are interested to see whether two means are different from each other

Inappropriate use of multiple t tests
each time you conduct a t test you increase the amount of type 1 error present (usually 5%)

If you want to compare the effects of three different treatments you should use
ANOVA (NOT t test!!)

ANOVA
analysis of variance is a tool used for analyzing multilevel and multi factor experimental designs (used to compare 3 means)

ANOVA H0
means all of the groups/conditions are equal (allows for 3 or more groups/conditions)

ANOVA Ha
 not all the meals are equal
 DOESNT say how or which ones differ
 (need posthoc for that

T or F: excel can run a posthoc test to tell which ones are different
FALSE! excel can only run the ANOVA test that can tell you whether all the means are equal

Correlations
 a measure of association between two numerical variables
 used to describe the strength and direction of a relationship between two variables
 correlation≠causation
 always a number from 1 to 1


Calculate r if the data is not linear??
NO!!!!!!!

Regression
 used when researcher wants to use the known relationship to predict an outcome
 useful for a prognosis
 regression≠causation

Independent variable
 predictive variable (X)
 variable used to predict another variable

Dependent variable
 criterion variable (Y)
 variable that the analysis predicts

Regression line
best fit line using least square method excluding all outliers

residuals
 the distance between the y values and their corresponding estimated values on the line
 include other factors that effect Y but are not included in the regression model

Y=B0+B1X +residual
 B0 is the regression constant and the Y int (DV) representing the value of Y when X=0
 B1 is the regression coefficient which is the slope of the line
 Y is the predicted value of Y
 X is the IV on the x axis

R^2
 coefficient of determination (square of correlation coefficient)
 means how much percentage that the X can account for the variation of Y

P value <0.05 in regression model
means regression model overall predicts outcome Y significantly well

