interrupted time series design
pretest-posttest nonequivalent groups design
stratified random sampling
dividing the larger population into subgroups, like gender, and randomly selecting the males seperately and the females seperately.
Ex: This type can be used if you know what 45% of the general population is white and you want to ensure that 45% of your sample population is white.
using whoever is convenient as your sample.
ex: standing outside a store and conducting a survey.
hand picking, or purposely choosing the individuals based on an assumption of how similar they are to the population you're trying to generalize to.
multiple time series design
nonequivalent control group design
correlation vs. causation
Involves getting the final sample of individuals by first sampling in larger subgroups.
Ex: at CSUSM you can't get lists of all the students there, but you could get a list of all the classes & randomly select the classes and go in and sample from each of those classes.
involves setting up a quota of the type of individuals you want. you can still set up subgroups, but now instead of random sampling, you use convenience sampling.
you get one or two to take the survey, gain their trust and then they go and convince others to take the survey.
Type of systematic observation. Is the behavior occurring at a particular point in time?
Type of systematic observation. Ex: Whether or not the behavior/event occurs at all during the 10 minute interval, or the next one, and so on.
continuous real-time measurement
give the test to a respondent at one point in time and then give it to them again at a later point in time
correlating the results from 2 halves of a measure, If you are measuring one construct you should have a high correlation between the 2 halves.
average item-total correlation
involves correlating each item with the total of all the other items (ex: item 1 with items 2-20). Average of everything else on the measure and then averaging all of the correlation scores.
The extent to which a test seems on its surface to be measuring what it purports to measure.
Does the instrument appear to be getting at the intended construct? Sometimes assessed by a panel of experts.
ability of the measure to distinguish between individuals that we know to begin with are different on this construct (ex: test to measure political conservatism, give it to a group of liberals and a group of conservatives you should be able to distinguish between these 2 groups).
ability of the measure to predict some future event/behavior of the respondents (ex: the SATs predicting how well you will do in college)
extent to which your measure agrees with other ways/measures assessing the same thing
determining that your measure does not produce the same results as measures of related but different constructs.
tendency to distort one's answers to items in a self-measure report, usually to make oneself look better or provide a more socially desirable answer
Type of unobtrusive measure involving physical traces. The wearing away of physical traces. Ex: counting the number of worn tiles in front of a museum exhibit to see how popular it is.
Type of unobtrusive measure involving physical traces. The building up of physical traces. Ex: if we counted the number of fingerprints on the glass at a museum exhibit to see how popular it is.