VALIDITY vs RELIABILITY by Ivan Prasetya Because of
VALIDITY vs. RELIABILITY by: Ivan Prasetya
Because of measuring the social phenomena is not easy like measuring the physical symptom and because there a lot of MEASUREMENT ERROR in measuring social phenomena, so we should do VALIDITY & RELIABILITY TESTING.
VALIDITY vs. RELIABILITY • Measurement VALIDITY: How well an empirical indicator and the conceptual definition of the construct that the indicator is suppose to measure ‘fit’ together. • Measurement RELIABILITY: The independability or consistency of the measure of a variable.
Different tools are used for measuring different things A tape measure is used to measure length or distance. A thermometer is used to measure temperature. A stopwatch is used to measure time.
If we want to measure weight/mass, we can use these tools: Digital scale have the highest VALIDITY. (1 litre = 1 kg)
However, even though a digital scale has the highest validity, but if the subject is not cooperative, the test result will not be accurate: y t i bil If you keep jumping around while measuring your weight, you won't get an accurate result of your body weight. h g i H t i d i l Va w o L , y a i l Re And if you carry extra loads while stepping on a digital scale, you won't get an accurate result of your body weight either.
VALIDITY vs. RELIABILITY Low Reliability, Low Validity High Reliability, High Validity
GOODNESS OF MEASURES GOODNESS OF DATA RELIABILITY (Accuracy in measurement) VALIDITY (Are we measuring the right thing? ) LOGICAL VALIDITY (CONTENT) FACE VALIDITY STABILITY TEST-RETEST RELIABILITY PARALLEL-FORM RELIABILITY CONSISTENCY INTERITEM CONSISTENCY RELIABILITY SPLIT –HALF RELIABILITY CRITERION-RELATED VALIDITY PREDICTIVE CONGRUENT VALIDITY (CONSTRUCT)) CONVERGENT DISCRIMINANT CONCURENT
RELIABILITY • Indicates the extent to which the measure is without bias (error free) and hence offers consistent measurement across time and across the various items in the instrument. • In other words, the reliability of measure indicates the stability and consistency with which the instrument measures the concept and helps to assess the “goodness” of measure.
STABILITY OF MEASURES TEST-RETEST RELIABILITY PARALLEL-FORM RELIABILITY The Reliability coefficient obtained with a repetition of the same measure on a second occasion. When responses on two comparable sets of measures tapping the same construct are highly correlated.
TEST-RETEST RELIABILITY If infrastructure in the country is not good, your company will cancel your decision to do Foreign Direct Investment (FDI) in that following country. y t i l i b a i l e R w o L Answer (now): strongly disagree Answer (20 days later): agree
PARALLEL-FORM RELIABILITY Do you think that Susi Similikiti is beautiful? Answer: YES y t i l i b a i l Do you think that Tukul’s e wife is beautiful? R w o Answer: NOL Fact: Susi Similikiti is Tukul’s wife.
CONSISTENCY INTERITEM CONSISTENCY SPLIT-HALF RELIABILITY This the test of the Reflecting the correlations consistency of respondents’ between two halves of an answer to all the items in the instrument. measure.
PROBLEMS and PITFALL in VALIDITY
VALIDITY in QUESTIONAIRE DESIGN • How diligent are you? 5 = very diligent 4 = diligent 3 = indifferent 2 = lazy 1 = very lazy y t i lid Lo a V w Because respondents have to evaluate themselves, they tend to choose number (4) and (5), or at least number (3).
To measure the degree of diligent (as a student), we also can use this questionaire: • How much time do you spend to learn everyday? • How many books do you read every week? • How often do you go to library every week? Hi y t i lid a V gh The answer which revealed from questionaire above will tend to be more accurate to measure the degree of diligent.
Avoid question that ask two things at once, you won’t know which ‘bit’ people are answering: Lo a V w y t i lid
Avoid ambiguity: Apakah kalian sudah membeli buku sejarah demokrasi yang baru? a V w Apanya yang baru? ? ? Lo Bukunya. . . ? ? ? Sejarahnya. . . ? ? ? Atau. . . Demokrasinya? ? ? y t i lid
Avoid jargon/abbreviations/slang: Higher number of current GDP in a country will encourage your company to do more FDI in that country. y t i lid a V w Higher number of current Per capita Income in a country will make your company do more FDI in that country. Lo If the level of corruption in a country increased, your company will reduce their amount of FDI in that country.
Avoid not mutually exclusive options: I’m 20 years old now. a V w What age are you? 16 -20 20 -25 25 -30 35 -40 � � y t i lid Lo
Problems and Pitfalls • Avoid making questionnaire too long y t i lid • Typographical / spelling errors Lo a V w
Using SPSS Analyze >>> Scale >>> Reliability Scale If item delete >>> Continue >>> OK If Corrected Item-Total Correlation > r-table: VALID If Cronbach’s Alpha if item deleted > r-table: RELIABEL
Using SPSS Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted VAR 1. 1. 1 152, 7333 249, 0299 0, 2785 0, 9368 VAR 1. 1. 2 153, 2667 244, 7540 0, 3568 0, 9366 VAR 1. 1. 3 153, 0333 249, 5506 0, 3113 0, 9364 VAR 1. 1. 4 152, 9667 246, 9989 0, 4354 0, 9357 VAR 1. 1. 5 153, 2000 248, 1655 0, 3300 0, 9364 VAR 1. 1. 6 153, 9000 245, 6793 0, 4432 0, 9356 VAR 1. 1. 7 154, 0667 246, 0644 0, 4555 0, 9356 VAR 1. 1. 8 154, 1667 241, 3851 0, 5454 0, 9349 VAR 1. 1. 9 153, 7000 245, 7345 0, 3929 0, 9361 Item-Total Statistics
THANK YOU!!!
- Slides: 24