Test Scoring and Analysis Using SAS by Ron Cody & Jeffrey K. Smith

Author:Ron Cody & Jeffrey K. Smith [Cody, Ron]
Language: eng
Format: epub
Publisher: SAS Institute
Published: 2014-12-01T08:00:00+00:00

where is the estimated correlation for a test that is N times longer (or shorter) than the given test and is the correlation of the original test.

Because a split-half correlation is assessed on a test half the length of the original, you can apply the Spearman-Brown formula to make the adjustment, like this:

where is the estimated correlation and is the split-half correlation. Although this method is no longer used (now that we have computers), it is still interesting to compute this statistic. If nothing else, it makes for some interesting SAS programming.

The first step in computing a split-half correlation is to divide the test in half and score the two halves. Here is the program:

Program 7.1: Computing the Score for the Odd- and Even-Numbered Items

*Split-Half Reliability;

data split_half;

set score;

array ans[56] $ 1;

array key[56] $ 1;

array score[56];

*Score odd items;

Raw_Odd = 0;

Raw_Even = 0;

do Item = 1 to 55 by 2;

Score[Item] = ans[Item] eq key[Item];

Raw_odd + Score[Item];

end;

*Score even items;

do Item = 2 to 56 by 2;

Score[Item] = ans[Item] eq key[Item];

Raw_Even + Score[Item];

end;

keep ID Raw:;

run;

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.