Hi Mei,

Before you use any statistical test in SPSS, you need to recode your negative items as previously suggested by Ravi and Jane. Simply  go to TRANSFORM Menu in SPSS and select Recode into the same variable. Then select the negative item from the list of variables and click on select old and new value. In the next dialog box select old (negative) value and in the new value put the corresponding opposite value. For example (assuming you are using 5 points Likert scale) if negative value is 5, use one in new value, for 4 use 2. 3 will remain unchanged as it is neutral point.

For Pre and Post you will need to use paired sample t-test in SPSS. For this, click on Analyze, Compare Means and the select paired sample t-test.

The process is not that simple but hope it will help you start in doing your analysis.

Thanks

First, you need to reverse score the negative items. If the responses for instance were on a 5-point Likert scale, the positive items are scored 0, 1, 2, 3, 4. Then the negative items should be rescored as 4, 3, 2, 1, 0.

In principle, then you could calculate mean, median etc from all you items.

However, it is often the case that the respondents react differently to positive and negative formulated items.

It is recommended that you check your questionnaire with Confirmatory Factor Analysis (CFA), in order to secure that all items in the scale are well choosen for the scale. It can be done with different programs, the best of which is Mplus (se www.statmodel.com). 

Such analyses will also give you indications as to the behavior of the negative formulated items. There are different methods for rectify the situation if there in fact is a difference in the behavior of positive and negative items.

A general introduction to CFA is:

Brown, T.A. (2006). Confirmatory factor analysis for applied research. New York: Guilford.

As suggested by Jan First, you need to reverse score the negative items. If the responses for instance were on a 5-point Likert scale, the positive items are scored 0, 1, 2, 3, 4. Then the negative items should be rescored as 4, 3, 2, 1, 0.

After that what you do use mean value or median according to the data. And use the new scale in which your have changed the values.

After that you can apply paired t test for seeing the before after mean difference is statistically significant or not in case the data is normal.

If normality condition is not followed then go for wilcoxon signed rank sum test , if going for non-parametric test.

Thanks

Hi Mei,

Before you use any statistical test in SPSS, you need to recode your negative items as previously suggested by Ravi and Jane. Simply  go to TRANSFORM Menu in SPSS and select Recode into the same variable. Then select the negative item from the list of variables and click on select old and new value. In the next dialog box select old (negative) value and in the new value put the corresponding opposite value. For example (assuming you are using 5 points Likert scale) if negative value is 5, use one in new value, for 4 use 2. 3 will remain unchanged as it is neutral point.

For Pre and Post you will need to use paired sample t-test in SPSS. For this, click on Analyze, Compare Means and the select paired sample t-test.

The process is not that simple but hope it will help you start in doing your analysis.

Thanks

Thanks all. I have got idea.

You can ask Şenol ÇİFTÇİ about this problem

Now, he is now working in the faculty of agriculture

I think some traits or phenomena or subjects are already negatives , for example, if the research tittle was " Obstacles of using Twitter in scientific research" , if we consider that all items of the tool of this negative subject were negatives , then after collecting data , no need to re code the scores of each item ( 1 to 5 or 5 to 1) , and the interpretation of items ( or dimensions) will go smoothly with items themselves.

You can ask statistician Dr. Şenol Çelik

His mail: [email protected]

Greetings

Hasan

Very interesting topic..

I am completely agree with Mr @Arif

In his explanation..

You can also very simple way using *COMPUTE in SPSS*

for example if you are using Five Likert Scale 1 strongly disagree and 5 strongly agree, then go to compute and firstly change the name of variable (for instance the previous one was "variable A" you have to change it to for example ( variable A_1) and put in dialog box numeric expression (6- Variable A_1), but Why (6 minus ), because it is five likert scale, if seven likert scale you have to put (8- variable A_1). Then the value will be changed directly, the 5 will be 1 and 4 will be 2 but 3 will remain the same value

.