Methods Final Exam

What is needed to establish validity
Click the card to flip 👆
1 / 32
Terms in this set (32)
A version, B version participants take both and compare scoresalternate forms reliabilityrandomly split quiz questions in heals. score each half and compare themsplit halvescontinuous generation of split halvesCronbach's alphalook at correlations between each individual question and the total scoreAverage item totallook at correlation between each individual questionAverage itemconsensus that a measure represents a particular concept. usually established by expert panel. validation by consensusface validityrepresentation of the breadth of material that should be presentcontent validityhow similar your measure is to another measure that is measuring a similar conceptconcurrent validityhow similar your measure is to multiple measures that it should be similar toconvergent validityoperationalization. are you measuring what you think you're measuringconstruct validitythe ability of your measure to predict what it should be able to predictpredictive validitycan NOT be ranked or ordered; can only use mode EX: hair colornominal measurementcan only be ordered; can use mode, median EX: age, fruit intake, class rankordinal measurementcan be ranked and ordered; can have negative values; can use mode, median, mean; EX: temperatureinterval measurementcan be ranked and ordered; has an absolute 0 (no negatives possible); can use mode, median, mean; EX: blood sugar, servings of veggiesratio measurementhow you draw the sample for your study from the populationrandom selectionhow you assign the sample that your raw to different groups or treatments in your studyrandom assignment__________ uses random assignment, whereas _______ does not, typically due to the fact that there are naturally forming/occuring groupsexperiment, quasi experimenthas a threat to internal validitymultiple group threatanything that makes the group different from the beginning of the studyselection threatwhat type of study design minimizes social interaction threat? How?cross-over design; because both groups receive treatmentthe extent to how you can generalize the results of your study back to the general populationexternal validity