-
conceptualization
process of coming to an agreement on term in corner of the sheet. theoretical definition.
-
indicators
identified to mark the presence or absence of a concept.
eg. love=holding hands, not touching shoulders
-
dimensions
subcategories of a concept
eg. family love, romantic love, platonic love
-
operational definition
created to define the procedures or steps used in measuring a concept. must be specific and unambiguous. HOW.
-
At the levels of measurement, the attributes must be...
exhaustive and mutually exclusive
-
Nominal measures
Categorical--Offers names or labels for attribute characteristics
eg. what is your primary source of news: internet, newspaper, radio, other
-
Ordinal level of measurement
Categorical-ranked ordered attributes. distance between is irrelevant.
eg. what is class standing: freshman, sophomore, junior, senior
-
Categorical levels of measurement
nominal and ordinal
-
continuous levels of measurement
interval and ratio
-
interval level of measurement
Continuous--distance relevant & standard, rank ordered, no true zero.
eg. choose the appropriate response: Newspapers are an important source of news for me: SA, A, N, D, SD
-
ratio level of measurement
continuous--measurements are based on a true zero point.
eg. what is your age? __
-
reliability
consistency. do you get the same results every single time?
eg. standing on a scale 100 times and getting the same number (even if that number is wrong).
-
validity
correctness. are we measuring what we say we're measuring?
eg. standing on a scale that says 392834 lbs, would be invalid.
-
Can results be reliable but not valid?
Yes.
-
Can results be valid but not reliable?
No.
-
Which is the highest level of measurement?
Ratio: You can transform a high level measurement into a lower level, but you can't do the reverse.
-
Types of non probability samping
convenience, purposive, snowball, quota
-
convenience sampling
asking the people most convenient to fill out a survey
-
purposive
choosing people based on your knowledge of the population
eg. needing business students, waiting outside upper-level buisness class
-
snowball
ask a few, get them to locate the others
-
quota
create a matrix of characteristics, assign proportions (%), collect people who match this. then use convenience, purposive, snowball or quota
-
About probability sampling
Sample should represent population, not have potential bias, ensure representativeness of all members of the population. Gives everyone an equal chance.
-
study/population
group you want to study
-
sample
group you want to participate in the study
-
sampling frame
the list of sampling units
-
observation unit
WHO is the data collected from
-
simple random sampling
establish a frame (list of people), use random # generator to choose people from the list.
-
systematic sampling
population size divided by sample size to get K. every kth person.
-
stratified sampling
create subsets (like m/f), create a matrix that allows you to determine how many of each you need, then use simple or systematic sampling
-
Name some challenges
- 1. Factual question, questionable answer.
- 2. Actions and reports of actions are different.
- -social desirability/acquiescence response bias
- 3. Respondent's attitudes, interests can appear unstable
- -result of taking survey, learning from survey, tiredness
- 4. Small changes in wording, big changes in answer.
- 5. Misinterpretation of word "often"
- 6. Respondents will answer even if they don't know
-
Closed question
researcher provides response options, mutually exclusive, exhaustive
-
open questions
respondent answers in her/her own words
-
how to make good questions
- -avoid double-barelled
- -avoid "not"
- -avoid bias writing (language)
- -write questions respondent is willing to answer
- -write questions relevant to respondent
-
strengths of survey research
- ·It is useful in describing a large population
- ·It enables the use of large samples
- ·It is flexible
- ·It offers many strengths with regard to measurement
-
weaknesses of survey research
- -in constructing response options, all participants' social experiences may not be accounted for
- -often removes social context from the study of communication
-
Content validity
how well a measure covers the range of meanings or the dimensions, included within the concept
-
construct validity
based on logical relationships among variables
-
representational validity
are measure's categories meaningful to the people who are being assessed?
|
|