Help Regarding Data Entry in SPSS


New Member
Could anyone plz guide me through these queries.I have a Survey of life experiences, whose responses i need to enter in SPSS. A list of experiences is provided and each is rated on a -3 to +3 likert scale.And the experiences that have not been experienced by respondents are left blank.It provides 3 scores positive experiences score, negative experiences score and total experiences score.

* How do i enter the experiences that have not been experienced by respondents...if i leave them unmarked obviously they get treated as missing data and effect the reliability of LES.As a result of which I haven't yet been able to calculate the reliability of LES for my population. Also how do i calculate the reliability of the survey for my population other than test-retest reliability.

* I also want to find out about the experiences that have not been experienced by population
and the frequency of each experienced experience, that is, how many respondents experienced each experience.
* For getting separate scores for negative experiences and positive experiences I'll have to recode into separate variables by defining the value ranges as -3 to -1 and +3 to +1 respectively.. Right?
Well, I realize it's a bit late to bring this up now, but there's no way for you to tell whether someone didn't have an experience (and therefore left the question blank) or whether they did have the experience (but simply chose not to answer your question). That's a problem that isn't fixable.

If you are simply going to assume that blank answers always equal a lack of experience, then I would use one of the the recode variable options to designate "system missing" as some other option (like 15 or something else outside the span of your likert scale options).

Sure, you can recode negative numbers as a "negative experience" response and positive numbers as a "positive experience." But what about neutral experiences? And what do you gain by doing the recode? You lose the element of the question when you collapse respondents' answers into big categories like that.

And what is LES? I'm not familiar with that term/test.


New Member
thanx for the response...
n LES is Life Experiences Survey...and these are the instructions during administration that respondents leave the unexperienced event blank...and the impact of the experienced event ranges from -3 to +3 .0 indicates no impact...
I was concerned that by assigning missing value will i be able to assess which experiences are not experienced.
and recoding into negative and positive experiences would provide basis for analysis, that is these indicate the level and nature of stress experienced shown via literature...

Also could u help me regarding calculating this survey's reliability??? Normally v enter all items as seperate variables and then run reliability for scale by entering all items but here since it is a survey therefore all items are not responded to, so will i have to find the reliability seperately for negatively experienced and positively experienced events (which would be recoded as seperate variables)
I'm not sure I really understand your questions, but I'll try to point you in the right direction.

Your respondents could provide 1 of 8 responses: -3, -2, -1, 0, 1, 2, 3, or not experienced. Yes? If so, I would recode that variable into a new variable. I would set old value range -3 to -1 as new value -1 (negative experience), old value range 1 to 3 as new value 1 (positive experience), 0 stays 0 (neutal experience), and all other values or system missing as 9 (not applicable). You would do this for each experience you asked about.

I don't understand what you mean regarding the survey's reliability. The only way you can tell if your survey is reliable is through the normal methods of pre-testing, replication, etc.

Do you instead mean how do you know whether your results are significant? I think maybe that's what you're asking. Remember, this is not scale data. It is ordinal data (at least, the original -3 to 3 responses are ordinal). If you collapse your responses into larger categories (negative/positive/neutral/na), it is nominal data. You need to use tests that are appropriate for the type of data you are trying to analyze, such as Chi-Square.