Class Notes (834,244)
Psychology (2,075)
PSYCH 339 (27)
Lecture

# MEASUREMENT PSYCH 339

4 Pages
170 Views

School
Department
Psychology
Course
PSYCH 339
Professor
Richard Ennis
Semester
Summer

Description
MEASUREMENT "The whole of science is nothing more than the refinement of everyday thinking.“ Albert Einstein *Consensus – “they say UW students are the best in Canada” – everybody agrees with that *Authority – “ I’ve been doing this job for 40 years and I know that UW students are the best in Canada” *Introspection (Rationalization/Intuition) – My own hypothesis, my own thinking – “I want someone from Laurier, forget someone from UW” Measurement (Test) Theory *Measurement consists of quantifying the thing being measured and quantifying something other than the thing being measured *Observation = True score + error (X = T + e) *There are two kinds of error in measurement Good Error *Random error (re) *X = T + re *Variance in scores caused by random fluctuations, deviations, or “noise” (caused by little inconsistencies in the measurement) *Errors that are higher & lower (more & less) than True Score *Why is it a good error? It is random if we take enough measures, those random chance might reach zero and if it reaches zero, it will result to a true score (if there are enough HR specialists you will reach zero) – as longa s there’s continuous data *Good error because: *It is measurable, quantifiable (error variance) *It is something to be seen easily *As sample size increases, random error approaches zero *Higher and lower deviations cancel each other *The combination of random error and enough HR staff, you will get the perfect or the right measure Bad Error *Systematic bias (sb) *X = T + sb + re *Variance in scores caused by a systematic, constant error in all scores *All scores are either inflated or depressed in the same direction *Errors that are consistently higher or lower (more or less) than True Score Bad Error *Bad error because: *It is NOT measurable, quantifiable *As sample size increases, systematic error does NOT decrease *Deviations are all in the same direction *No matter how many HRs are asked to do the measurement, there will be no difference on the deviations *Disguised as True Score (e.g. you may say that it’s included in andrew’s height but not) X = (T + sb) + re x X = (Tx+ T x + re Discrimination Good Discrimination * Selection based on Bona-fide Occupational Requirements (BFOR’s) *Basing the discrimination that is related to the job *E.g. I’ll choose the other one because she met the requirements more than the other one Bad Discrimination * Selection bas
More Less

Related notes for PSYCH 339
Me

OR

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Join to view

OR

By registering, I agree to the Terms and Privacy Policies
Just a few more details

So we can recommend you notes for your school.