NIVRAHA151_00848 
Vorige  848 van 1810  Volgende 

klein (250x250 max)
gemiddeld (500x500 max)
groot
Extra Large
groot ( > 500x500)
Hoge Resolutie
Allen (PDF)

Deze pagina
Allen

794 HANDBOOK OF BUSINESS ADMINISTRATION distribution case, now, because of different train connections, belongs i» compartment 7. From time to time locations of postal stations change, do mail routes. He must, therefore, be able to inhibit the tendency to place certain letters in compartments 3 or 7, and must recall that, to make different connections, these letters now go into compartment 8. Our mail clerk's problem, then, is to memorize these 5,000 names, to keep frequent readjustments and changes in mind, and yet to distribute very accurately and very rapidly. Since the entire study is covered thoroughly in one of our annual reports,' I shall discuss here only the nature of tlie tests used to measure the capacities which analysis of the duties showed to be necessary, and to the criteria u to determine their validity. Of the fifteen tests which we tried, the most effective one was a simple substitution test which requires the competitor to study for ten minutes list of names, making mental associations between those names and key numbers, and thereafter to substitute the correct number for each name in the test list. As the competitor continues to make these substitutions, his task is simplified in direct relation to his ability to form the associations. Comparison of differences in employees' test scores and differences in their efficiency ratings showed t h a t this test, which we call t h e "sorting test, is excellent for predicting the api)licant's output as a distributor in the post office. Applicants differ widely in the number of items which they answe' in the time allotted, this number ranging from 20 to 350. The test ranking next in selectivity was a "following instructions" test, which measures differences in the degree of accuracy with w^hich competitors are able to follow directions in making changes in distribution schemes i in routing. That each of these tests measures qualifications essential for success in t h e job is demonstrated by a correlation^ of 0.503 between efficiency ratings and scores on one of the tests, and by a correlation of 0.647 between efficiency ratings and scores on the other test. The predictive value of this battery of special tests is indicated by a correlation with the criterion of 0.65 in the first trial, 0.67 in the second trial, and 0.69 in the third trial. These special tests measure qualifications required for performing the duties of clerks in the post office much better than do the general tests. By general tests I refer to what is popularly known as the intelligence test. Scores on the test of general intelligence correlated with the criterion in the first study 0.44, in the second study, 0.37, and in t h e third study, 0.40. Hoifever, the addition of the general test to the b a t t e r y of special teats increase the correlation from 0.69 to 0.715. What does a correlation mean in terms of selection value in any speciM case? Looking at this correlation chart (Fig. 1), which is only a sample and has no connection with the postal tests to which I have referred, we fin" ^ Fortysecond Annual Report of the U. S. Civil Service CommiaBion. 2 For practical purposes, think of the coefficient of correlation as the mathematic expression of the relationship found between any two variables, such as test scores an efficiency ratings. that efficiency i efficiency at the basis of various i SO on the test, who made this lower than aver itself would not scores and succei We also find above t h e averag efficient employe o I fto. 1.—Test SCO] Of the 18 perse 4e applicants in than average effic: 1 have introduc *ould not be co: '•»lue is also deter, '•"stribution of cas , ^'hat the statis '" actual results is This does not m 'wugh onethird of t A ratio close to f ' The statement in
Beschrijving voorwerp
Titel  Handbook of business administration 
Jaartal  1931 
Collectienaam  NIVRA Historisch Archief, UBVU gedigitaliseerd 
PPN  344556336 
Toegangsgegevens (URL)  http://imagebase.ubvu.vu.nl/getobj.php?ppn=344556336 
Signatuur origineel  NIVRAHA151 
Evaluatie 
Beschrijving
Titel  NIVRAHA151_00848 
Transcript  794 HANDBOOK OF BUSINESS ADMINISTRATION distribution case, now, because of different train connections, belongs i» compartment 7. From time to time locations of postal stations change, do mail routes. He must, therefore, be able to inhibit the tendency to place certain letters in compartments 3 or 7, and must recall that, to make different connections, these letters now go into compartment 8. Our mail clerk's problem, then, is to memorize these 5,000 names, to keep frequent readjustments and changes in mind, and yet to distribute very accurately and very rapidly. Since the entire study is covered thoroughly in one of our annual reports,' I shall discuss here only the nature of tlie tests used to measure the capacities which analysis of the duties showed to be necessary, and to the criteria u to determine their validity. Of the fifteen tests which we tried, the most effective one was a simple substitution test which requires the competitor to study for ten minutes list of names, making mental associations between those names and key numbers, and thereafter to substitute the correct number for each name in the test list. As the competitor continues to make these substitutions, his task is simplified in direct relation to his ability to form the associations. Comparison of differences in employees' test scores and differences in their efficiency ratings showed t h a t this test, which we call t h e "sorting test, is excellent for predicting the api)licant's output as a distributor in the post office. Applicants differ widely in the number of items which they answe' in the time allotted, this number ranging from 20 to 350. The test ranking next in selectivity was a "following instructions" test, which measures differences in the degree of accuracy with w^hich competitors are able to follow directions in making changes in distribution schemes i in routing. That each of these tests measures qualifications essential for success in t h e job is demonstrated by a correlation^ of 0.503 between efficiency ratings and scores on one of the tests, and by a correlation of 0.647 between efficiency ratings and scores on the other test. The predictive value of this battery of special tests is indicated by a correlation with the criterion of 0.65 in the first trial, 0.67 in the second trial, and 0.69 in the third trial. These special tests measure qualifications required for performing the duties of clerks in the post office much better than do the general tests. By general tests I refer to what is popularly known as the intelligence test. Scores on the test of general intelligence correlated with the criterion in the first study 0.44, in the second study, 0.37, and in t h e third study, 0.40. Hoifever, the addition of the general test to the b a t t e r y of special teats increase the correlation from 0.69 to 0.715. What does a correlation mean in terms of selection value in any speciM case? Looking at this correlation chart (Fig. 1), which is only a sample and has no connection with the postal tests to which I have referred, we fin" ^ Fortysecond Annual Report of the U. S. Civil Service CommiaBion. 2 For practical purposes, think of the coefficient of correlation as the mathematic expression of the relationship found between any two variables, such as test scores an efficiency ratings. that efficiency i efficiency at the basis of various i SO on the test, who made this lower than aver itself would not scores and succei We also find above t h e averag efficient employe o I fto. 1.—Test SCO] Of the 18 perse 4e applicants in than average effic: 1 have introduc *ould not be co: '•»lue is also deter, '•"stribution of cas , ^'hat the statis '" actual results is This does not m 'wugh onethird of t A ratio close to f ' The statement in 
Tags
Toelichtingen
Geef een Toelichting voor NIVRAHA151_00848