Business Research Terms - Convenience Sampling & Validity

a representative sample

allows you to make inferences on what a greater population believes

random sampling

only chance determines which elements will make it into the study (equal chance for all)

probability sampling

experimental research design, has randomization, explanatory design (answer "why" questions); allows you to make generalizations

nonprobability sampling (what we use)

cannot generalize to broader population, unknown chance of being selected, used when probability sampling isn't possible, used for explanatory research

confidence level

95%; if the same study were conducted 100 times with random samples drawn from same population, then 95 times the result would be the same

levels of measurement

nominal, ordinal, interval, ratio

nominal measurement

basic; no numerical data drawn from it; check boxes/categories

(i.e. gender, religious affiliation, etc.) - usually demographic info

ordinal measurement

grouping; meaning, but not clear implications; ranking order

(i.e. satisfaction --> high, mid, low)

interval measurement

category has distance meaning, but there is NO ZERO; scale

(i.e. IQ, etc.)

ratio measurement

equal intervals, and true zero

(i.e. age, # of kids, GPA, income)

social desiribility

tendency to say the socially acceptable things or to tell people what you think they want to hear; changes answers because of this

reactivitiy

people tend to change their behavior when they know they are being observed

random error

won't throw off results; not considered a measurement error

i.e. respondent misreads a question

patterned/systematic/nonrandom/consistent error

will throw off results/does not balance out

validity

is the instrument measuring what it says it is measuring?

(does the spiritual assessment measure spirituality)

construct validity

correlation with expected related concepts

face validity

does it make sense?

criterion validity

same result as known other measurement is the result

concurrent validity

test against other valid measure

content validity

does it fully represent concept

predictive validity

ability to predict events

testing reliability

interrater - 2 or more people all test & compare for consistency

intrarater - test same rater twice

sampling frame

a list of all of the elements in a population from which the sample is selected

sampling bias

the systematic tendency to over- or under-represent some segment of the population

nonsampling error

sampling error that is not related to the actual sampling procedure and includes inadequate sampling frames, high levels of attrition or nonresponse, and measurement and data entry errors

simple random sampling

choice of a subset of a population, i.e., the sample, utilizing a method that allows each element to have an equal chance of being chosen; the sampling frame is not subdivided and should not be ordered

systematic random sampling

nonrandom error; in measurement, those sources of error that have a pattern and will influence the overall distribution of scores

order bias

when selecting a sample, the ordering in the sampling frame corresponds with the random start and the sampling interval to create an under- or over-representation of subgroups within the population

disproportionate sampling

the proportion of each stratum in the sample does not correspond to the respective proportion in the population; small strata may be over-sampled

cluster sampling

a random sampling method in which aggregates of elements called clusters are randomly sampled first, and then the elements of the population are either all included (cluster sampling) or are randomly sampled (multistage random sampling) from the selected clusters

convenience sampling

a nonprobability sample in which the selection of sampling elements is based on the most available elements to constitute the sample

sampling error

the degree of difference between the sample and the population from which it was drawn

attrition

tendency of participants to withdraw prematurely from a research study

cutting scores

on a scale, a measure that sets the difference between two levels of a condition or performance

scale

in instrument construction, a score obtained by adding scores assigned to specific responses and also taking advantage of any intensity structure that might exist among the individual items, for instance by weighting the responses to the individual items differently

primary data

the collection of new data

secondary data

data that were previously collected for a purpose other than the research at hand

response rate

in a survey, the response rate is the percentage of all those selected in a sample who actually respond