Module 4 Section 2 Flashcards
Measurement
-assigning numbers to represent the amount an attribute present in a person or object
Advantages of Measurement
- Removes guessing out of gathering data
- tends to be more objective
- reasonably precise
Errors in measurement
-measurement errors are systemic and random error associated with a person’s score on a measure, reflecting factors other than the construct
Common factors contribute to measurement errors
1.Situational contaminants
-awareness of observers may affect their behaviours
2.Response-set biases
-characteristics of the respondents can interfere with
the accurate measures of target attribute
3. Transitory personal factors
- fatigue, hunger, anxiety and mood can affect
motivation to cooperate
4.Administration variations
-altering collection methods can alter results
5. Item Sampling
-sampling of items used to measure a characteristics
can contribute to error
Reliability
- the accuracy and consistency of a measuring instrument
Stability
-the ability to give consistent results when the same people are tested at different times
Measure with test re-test reliability
-researcher gives the same test to a sample on 2 occasions then compares the results
Reliability coefficient
-measure reliability with values between 0 to and 1. higher is equal to more reliability
Limitations of test-retest
-subjects remember the items being tested
-traits of interest change over time
-repeating could change response given
-repetition could bore the participants causing them to
respond differently
Internal consistency
-extent to wich all items on an instrument measure the same critical variables or attributes
Measuring techniques
Split half technique
-split it into 2 halves to make comparisons
Spearman brown formula
-used to calculate the correlation coefficient
Item total correlations
-correlation between each item and total item. above
0.25 are acceptable
Kuder-Richardson (KD-20)
-yes or no response
Cronbach’s alpha
- coefficient measure in likert scale with each other simultaneously
Equivalence
-the degree to wich 2 or more observers using a single instrument obtain the same results
Validity
-an instrument measures what it was suppose to measure
-it addresses the question of how well we measure social reality using our constructs about it
4 major types of validity measurements
1.Face validity
2. Content Validity
3. Criterion-related Validity
4. Construct validity
Face Validity
-looks like it is measuring the target construct
Test for with: expert opinion and panel of experts
Content Validity
-extent the instrument covers the factors or situation under study
Test for with: verify with other evidence
Criterion related validity
-is the extent to which a subjects performance coincides with behaviour
Includes
1. concurrent validity:
-one measure correlates with another measure of the
same phenomenon
test for with:
test instrument with another instrument
that measures the same and is valid
2. Predictive validity
- can accurately predict a phenomenon
- test for with
-using the instrument in a study and compare with
a future outcome
Construct validity
- is the extent to which the test measures a theoretical concept
- test for with: hypothesis testing and convergent and divergent approaches, contrasted groups, factor analysis, casual modeling
Trustworthiness of qualitative data
- Credibility
- Dependability
- Confirmability
- Transferability
Credibility
-the believability of findings
Ways to ensure credibility
- prolonged engagement and persistent observations
-increased time invested = increased credibility
2.Triangulation- uses multiple methods to determine true info from
info with errors
- uses multiple methods to determine true info from
- Peer debriefing
-receive feedback about data quality and
interpretation from peers - Member checks
-obtain feedback from participants - Searching for disconfirming evidence
-occurs through purposive sampling
-facilitated through such processes as prolonged
engagement and peer debriefs - Researcher credibility
-is the researcher qualified to conduct research
Dependibility
- stability and reliability of data
- will the results be consistent after time
Confirmability
- objectivity or neutrality of the data and the data interpretation
- 2 or more independent people would agree with datas relevance and importance
Transferability
-extent to which the qualitative findings have acceptability in other settings and groups
Selecting measurement tools
- psychometrics
- describe expertise in instrument construction. Should possess comprehensive subject expertise and knowledge and skill in test and scale constructions
Characteristics of a good instrument
- uniform set of items and response possibilities
- Clear and concise statements
- Only 1 idea per statement
- Negative items and double negatives (confusing)
- Restricted to a few variations
- Not provide clues to other items
- Should cover a broad area of defined behaviour
- Adequately cover defined behaviour
- Measure what its intended to measure (validity)
Steps to develop measurement instruments
- Define the concept to be measured
- Formulate the items
- assess items for content validity
- Develop instructions for respondents and user
- Pretest and pilot test the items
- Estimate reliability and validity