1. ## Need help to correct minimum and maximum error Please

I am using scales to test my constructs for example product category involvement and the scale is made up of 11 items ( measure using 5 point likert scale1=totally agree, 5 totally disagree) and I want to compute them into a new variable by adding all of the scores from each of the 11 items together then dividing by 11 to get the numbers back to normal (in order to have one variable to represent the construct) the issue is when I do this my minimum and maximum values are wrong. When I look a the 11 items individual and run frequencies the min and max values are correct and I really don't know where this error is coming from or how to solve it. The min values should be 1 and max 5 but my min is 2 and my max is 4.3. But this is wrong because i have respondents who have 1 and 5 as answers so I don't understand what is going on.

Help is GREATLYYYY APPRECIATED!

2. ## Re: Need help to correct minimum and maximum error Please

So you are creating a scale score by taking the average of 11 items right? In this case for a min scale score to be 1 then you would need at least on participant that gave a 1 response toall 11 items (11*1/11=1). Likewise for a max of 5 you would need at least one participant you gave a 5 on all 11 items (11*5/11 = 5).

Is it possible that none of your participatns fitted into either of these categories and hence the reason you have mins and maxs that are not 1 and 5?

3. ## Re: Need help to correct minimum and maximum error Please

I actually fixed the issue and my numbers are coming out to 1 min and 5 max i had to reverse my scale measure cause some were negative then I recalculated and it came out right. However now I have another issue lol. So once i make the new variable with 12 items (diff scale) add all of them together then divide by 12 I get this under frequencies (pic attached) however I was wondering is there a way where Spss can make the numbers fit under a whole number like if the score of a respondent is 1.7 then they are group under 2 or something like this. round up the numbers so to speak to say a respondent strongly disagrees rather then they are between disagree and strongly disagree?

Can't attach the pic dunno how its a screen shot?