Term
|
Definition
| observation that is assumed to be evidence of the attributes or properties of some phenomenon |
|
|
Term
|
Definition
| single indicator of variable |
|
|
Term
|
Definition
| composite of multiple items |
|
|
Term
|
Definition
| bathroom scale (x) = actual weight (t) + clothes (e) |
|
|
Term
|
Definition
| Nominal, Ordinal, Interval, Ratio |
|
|
Term
| Nominal scale of measurement |
|
Definition
| unordered, sex (0=male, 1=female) |
|
|
Term
| Ordinal scale of measurement |
|
Definition
| ordering of categories, NOT equal spaced (1st, 2nd, 3rd) |
|
|
Term
|
Definition
| in order, equally spaced (IQ) |
|
|
Term
|
Definition
| equally spaced, absolute 0 |
|
|
Term
|
Definition
| Does what you claim you're testing actually test what you're measuring? |
|
|
Term
|
Definition
| does the content of the test sample all parts of the domain |
|
|
Term
| criterion-related validity |
|
Definition
| does this measure correlate with other known tests that measure this idea? |
|
|
Term
| Construct validity: discriminate validity |
|
Definition
| correlation between measure and another unrelated construct |
|
|
Term
| construct validity: convergent validity |
|
Definition
| correlation between measure and another measure that measures the same thing |
|
|
Term
| Threat to Construct validity |
|
Definition
| unreliable scores, poor construct definition, construct confounding, monomethod bias, mono-operation bias, evaluation apprehension |
|
|
Term
|
Definition
| consistent each time it's used |
|
|
Term
| reliability: internal consistency |
|
Definition
| split scale in half, give each half to group, compare halves and get correlation. if r=.80 or better, it's reliable |
|
|
Term
|
Definition
| A comes before B, A covaries with B, and no other explanations are possible |
|
|
Term
|
Definition
| inference about observed covariation between treatment and outcome reflect a CAUSAL relationship |
|
|
Term
| Threats to Internal Validity |
|
Definition
| ambiguous temporal precedence, history, maturation, regression, attrition, testing, instrumentation, selection |
|
|
Term
Ambiguous Temporal Precedence
(Threat to Internal Validity) |
|
Definition
|
|
Term
History
(Threat to internal validity) |
|
Definition
| events happening at same time as treatment that may change outcome |
|
|
Term
Maturation
(threat to internal validity) |
|
Definition
natural changes over time
(height/weight, IQ) |
|
|
Term
Regression
(threat to internal validity) |
|
Definition
|
|
Term
Attrition
(threat to internal validity) |
|
Definition
| experimental mortality (dropout) |
|
|
Term
Testing
(threat to internal validity) |
|
Definition
|
|
Term
Instrumentation
(threat to internal validity) |
|
Definition
| measure may change over time - because of researcher perhaps |
|
|
Term
selection
(threat to internal validity) |
|
Definition
| change may be due to conditions (Chicago) |
|
|
Term
|
Definition
| each element of population has same chance of inclusion |
|
|
Term
| random sampling (type of probability sampling) |
|
Definition
| same chance of getting sampled |
|
|
Term
| stratified random sampling |
|
Definition
Grouping population in smaller subgroups (strata) and sampling randomly from strata)
function: reducing sampling error |
|
|
Term
|
Definition
AKA cluster sampling, multi-stage sampling
big area, then smaller, then smaller, etc |
|
|
Term
|
Definition
| no intent to generalize, focused on process, etc. |
|
|
Term
availability sampling
(nonprobability sampling) |
|
Definition
|
|
Term
purposive sampling
(nonprobability sampling) |
|
Definition
|
|
Term
dimensional sampling
(nonprobability sampling) |
|
Definition
| what social research actually uses mostly |
|
|
Term
|
Definition
| can you generalize to other populations? |
|
|
Term
| Threats to external validity |
|
Definition
| interaction of causal relationship with UNITS, TREATMENT VARIATIONS, OUTCOMES, SETTINGS |
|
|
Term
interaction w/units
threat to external validity |
|
Definition
| only found effective on some people |
|
|
