reader

ITEM-TOTAL CORRELATIONS ON EXCEL Department of Psychology, University of Cape Town

1.) Copy your data from the reliability calculator excel sheet and put it in a new, blank excel workbook. This is because the reliability calculator has formulae built into it, and doing additional calculations there might well interfere with the formulae that are already in there.

2.) Your data in the new, blank workbook should be set out in the same way as the reliability calculator - i.e., each question gets a column, and each respondent/participant gets a row. See the screencap below under point (3), which shows each question (Q1-12) in a column, and each participant (P1-12) in a row.

3.) You then need to get the sum total scores of each person on your questionnaire - put these in a new column in excel. To get sum total scores, you simply add up a person's scores on each item. This can be done easily using the =SUM excel formula to get the sum of the first person's scores. See the screencap below, which shows me getting the sum scores for participant 1 (P1). Note that I have summed the scores from left to right in Excel, i.e., I am summing cell B2 to cell M2.

Page 1 of 3

4.) After you have the sum of the first person's scores, you can autofill down to get the sum scores for every person. To do this, move the cursor over bottom right hand corner of the cell with the sum score in it, and it will become a small black cross. Drag that down until the last person's row, and all the sum scores will be automatically calculated. In the screencap below, you can see I now have all the sum total scores for the participants.

5.) You now have the total scores of everyone on your questionnaire. To do an item-total correlation for, e.g., Item 1, go to a blank cell and type in =CORREL or =PEARSON, and then a bracket (a round bracket). Excel will ask you for array 1 and array 2. To input the data for array 1, simply highlight the column that has all the persons' scores for Item 1, and to input the data for array 2, simply highlight the column which has the sum total scores in. In the screencap below, you can see that I have entered as Array 1 the column of scores for Q1, and Array 2 is the sum total scores. This gives me an item-total correlation for Item 1 of .92 – very good.

Page 2 of 3

6.) Repeat Step 5 above for all your test items, but each time Array 1 will change to the column containing the scores on the individual test questions, e.g., the column with scores on Item 2, Item 3, etc. Array 2 always stays as the sum total scores. In the screencap below I have completed the item-total correlations for all the items. It is clear that Q2 and Q8 have very low item-total correlations. When I go back to the reliability calculator and delete these two items, overall alpha goes up from .92 to .96 – so it is clear that these two test items are bringing down the reliability of this test quite a bit.

Page 3 of 3