exam 1 Flashcards
descriptive research
examines group differences, relationships among variables through observations
-DESCRIBING. NOT INTERFERE OR CHANGE
experimental research
looks at cause-effect through manipulation of certain variables on measures under controlled conditions.
experimental design
-manipulates 1+ variables while using CONTROL mechanisms to evaluate change in behavior/outcome measure
single subject experimental designs
small number of participants (usually less tan 5)
-ex control mechanisms= baselines, rapid alt, withdrawal
group experimental designs
large number of participants (usually greater than 10 per group)
-control mechanism= control group (random best)
correlational study
large number of participants
relationship among measures
measures at one time
early measures to later outcome measures
case study
descriptive research design
describe unique case
describe small treatment study- no control mechanism
dependent variable
outcome measure(s)
- the effect you measure
- measure that answers your research question
independent variable
the cause or factors that influence the measure or DV
active IV
manipulated by investigators
assigned IV
inherent characteristics of subjects, not manipulated by investigators, presumed to effect DV
part of a research paper
- abstract
- intro
- methods
- results
- discussion
nominal data
mutually exclusive categories. naming
ex. type of hearing loss (SN or conductive). type of communication disorder, etc
ordinal data
mutually exclusive categories rank orders
- mild, moderate, severe
- low,middle, high
interval data
mutually exclusive categories rank orders- CONTINUOUS MEASURES
- equivalent units throughout the scale
- ex. standard tests on scores
- NO ZERO (percentile ranks)
- likert type of rating scales
ratio data
rank orders
equivalent units
ABSOLUTE ZERO
-ex. duration, dB of HL, number of disarticulations, raw scores on tests, etc
negatively skewed data
floor effect
-our measure was too hard and everyone failed
positively skewed data
ceiling effect
measure was too easy, everyone clustered toward the high end
correlations
info about the direction and strength of relationships
- how associated 2 measures/variables are
- the closer to 1 the stronger the relationship
multiple regression correlation
look at relative predictive value/strength of relationship between numerous variables and an outcome variable
-looks at more than 2 measures
type 1 error
(p value) probability of saying there’s a difference between the groups when in actuality there’s not really a difference
type II error
finding that there’s not a difference between groups when there really is a difference
post hoc comparisons
when there’s more than 2 groups the ANOVA tells you there’s a difference somewhere - but post hoc tells you where the difference in within the groups
MANOVA
multivariate analysis of variance
more than one dependent variable
two way/three way ANOVA
more than one independent variable in a single analysis
-looks at significance for each IV, and also for interactions between the IVs