COMM 88 Lecture Notes - Lecture 7: Face Validity, Predictive Validity, Convergent Validity
4/24/18
●Measurement -- operationalizing variables (both IVs and DVs)
●Assessing reliability
○Inter-item reliability
■You want all items to be indicators of the SAME variable
●If so, you get a high internal consistency (high Cronbach's alpha)
●A good “unidimensional” variable/concept
●Example variable: candidate credibility
○What if credibility involves more than just trustworthiness? Would need
“multidimensional” scale
■Trustworthiness ---- untrustworthy
■Honest ---- dishonest
■Sincere ---- insincere
○Or
■Knowledgeable ---- not knowledgeable
■Experienced ---- inexperienced
■Competent ---- incompetent
○Cant put all 6 together because it will throw off credibility scale because they are really
two separate parts of credibility (someone can be knowledgeable but not honest)
■Different aspects of credibility — the sub-scales are each unidimensional but
together are multidimensional
●Make sure to evaluate reliability separately for each sub-scale
●Assessing reliability
○For measures using coders (e.g. behavioral observations):
■Inter-coder reliability
●Compare multiple coders
■Intra-coder reliability
●Compare multiple observations of the same coder
●Validity of measurement
○Does your measure really capture the concept you intend to be measuring?
■Good fit of measure with concept
○Subjective types of validity:
Document Summary
Measurement -- operationalizing variables (both ivs and dvs) You want all items to be indicators of the same variable. If so, you get a high internal consistency (high cronbach"s alpha) Cant put all 6 together because it will throw off credibility scale because they are really two separate parts of credibility (someone can be knowledgeable but not honest) Different aspects of credibility the sub-scales are each unidimensional but together are multidimensional. Make sure to evaluate reliability separately for each sub-scale. For measures using coders (e. g. behavioral observations): Compare multiple observations of the same coder. The measure looks/sounds good on the face of it . The measure captures the full range of meaning/dimensions of the concept. The measure is shown to predict (not cause) scores on an appropriate future measure. Ex: sat scores (your potential to achieve) college gpa (your achievement) The measure is shown to get same result as another measure of the same thing.