click below
click below
Normal Size Small Size show me how
RMDA ch 5 and 6
| Term | Definition |
|---|---|
| reliability | asks if there's consistency (agreement) about a phenomenon |
| reliability coefficient | pearson's product coefficient (r); cohen's kappa |
| test-retest reliability | using the same or similar test twice to get similar results; e.g. intelligence shouldn't vary weekly |
| internal consistency reliablity | use questions that test same thing using dif methods (e.g. MC and essay on same test); or spilt half reliability (cronbach's alpha) |
| inter-rater reliability | using two raters (Cohen's kappa) to confirm observations are the same |
| what to do with unreliable data? | get another coder; retrain; adjust operational definition; use different measure, condition, test |
| reliability is | consistency |
| validity is | accuracy |
| possible to have reliability without validity if | using wrong methods, measurement, etc; consistently measuring incorrect results |
| construct validity | does the operational definition reflect the meaning of the variable? accuracy |
| are there alternate operational definitions? | what are implications of this operational definition? |
| face validity | scale or test appears to measure what its designed to (shallow analysis); may get socially desirable answers |
| content validity | does the measure assess what the researchers think are the key components of a topic/ construct? |
| predictive validity | validly predicts hypothesized outcome; can be predictively valid without being face or content valid (e.g. GRE and grad success) |
| concurrent validity | examines whether two or more groups of people differ on the measure in expected ways |
| concurrent means | existing or happening at the same time |
| convergent validity | scores on one measure are related to scores on another measure of same construct; convergent = coming closer/together |
| discriminant validity | when measures aren't related to other measures that are theoretically different |
| discriminant means | allows things to be separated from each other |
| e.g. discriminant | measure of depression should be unrelated to measure of favorite ice cream |
| reactive measurement | participants react to the knowledge they are being measured/observed, changes their behavior; confound |
| reactive bias increases with | clarity (face validity) of measure; face-to-face interview methods |
| reactive bias decreases with | computer methods; use unobtrusive methods; give people time to get adjusted |
| scales of measurement (MS) | nominal, ordinal, interval, ration |
| nominal scale | categorical; no numerical property; levels represent different categories of groups but one category not better than other |
| ordinal scale | order; levels from lowest to highest etc; intervals between levels don't have to be equal; rankings mean smth (e.g. first place) |
| interval scale | intervals between levels are equal in size, but may not be in a person's mind; no absolute zero; e.g. attitude, temperature |
| ratio scale | equal intervals; same as interval but with absolute zeroes; can be summarized w/ mean; e.g. time, height |
| order of values is known (MS) | ordinal, interval, ratio |
| "counts", frequency distribution (MS) | nominal, ordinal, interval, ratio |
| mode used for (MS) | nominal, ordinal, interval, ratio |
| median used for (MS) | ordinal, interval, ratio |
| mean used for (MS) | interval, ratio |
| can quantify difference between each value (MS) | interval, ratio |
| add or subtract values (MS) | interval, ratio |
| multiply or divide variables (MS) | ratio |
| has true zero (MS) | ratio |
| quantitative approaches | behaviors that can be quantified; assign numerical values to responses and measures; large samples; stat data analyses |
| e.g. quantitative approaches | surveys, correlational studies, experiments |
| qualitative approaches | observational measures; behavior in natural or limited setting (e.g. focus group); non-numerical; researcher draws conclusions |
| naturalistic observation | fieldwork; field observation (no manipulation); observations made in natural setting over period of time |
| techniques to collect data - naturalistic observation | observing and interviewing participants; surveying documents |
| goal of naturalistic observation | describe setting, events, persons; analyze categories that emerge; researcher interprets what occurred; focus on themes |
| no participation in naturalistic observation | non-participant observer; behavior may change; remain objective; never part of group |
| yes participation in naturalistic observation | participant observer; may know more but lose objectivity |
| naturalistic observation - concealment pros | useful in complex and natural settings; naturalistic behavior; most ecologically valid |
| naturalistic observation - concealment cons | no informed consent; cannot be used to study all issues; no control/manipulation; less useful for testing well-defined hyp |
| systematic observation | observation of specific behaviors in a specific setting; coding system; methodological issues (equipment), reactivity, reliability |
| case studies | description of an individual (e.g. patient H.M.); valuable for rare or unusual situations |
| psychobiography | type of case study where researcher applies psychological theory to explain the life of an individual (e.g. what Freud did) |
| archival research | uses previously compiled info to answer research questions; don't collect data, only analyze; e.g. Census records |
| cons of archival research | confirmation bias; could be dated info; irrelevant or not useful to your question; bound by when data collection started |
| limitations of archival research | difficult to obtain; original info collected by someone else (accuracy? ethics?) |