It makes no difference whether you score from disagree (1) to agree (5) or the reverse. Just be careful to interpret results in the correct direction, as correlations will be opposite in sign between the two.

thanks a lot sir.

Dear Mah Bhutta ,

As Dr. Spector explained, it makes no difference whether you score from disagree (1) to agree (5) or the other way around. However, when scoring your items you should make sure that all of the items from the same scale go in the same direction (i.e. lower/higher scores for each item reflect lower/higher scores for the latent variable you are attempting to measure). For example, imagine that you are trying to measure a latent variable called job satisfaction through a scale consisting of four items (these four items are your observed variables), with responses anchored in a 5-point Likert scale:

• I find real enjoyment in my job
• I like my job better than the average person
• Most days I am enthusiastic about my job
• I feel quite dissatisfied with my job

If you score your items from 1 (= strongly disagree) to 5 (=strongly agree), that means that higher levels of agreement (i.e. higher scores) for items 1, 2, and 3 reflect higher levels of job satisfaction whereas higher levels of agreement for item 4 reflect lower levels of job satisfaction. Consequently, you must bear in mind that item 4 should be reverse scored (strongly disagree = 5, strongly agree = 1) so that all of the items go in the same direction.

I hope my explanation helped. Good luck with your research!

Thanks for your detailed answer. I got your point.

Respected Prof @ Paul E. Spector,

Can you please explain a little more on this issue that why, at the first place, one would label all positive statements of a measurement scale inversely (Strongly agree=1 , Strongly disagree=5)?

Can you please explain with some example and provide a reference to study in this regard?

Moreover, does it make no difference at al in Correlation and Regression analysis in any type of statistical analysis package e.g., SPSS, SmartPLS

Best Regards

I will agree with all that it makes no difference either you assigned a value of 5 or 1 to strongly agree. But please be aware of the culture of your sample.

As in some cultures they generally consider higher value means the more better like in Pakistan and Malysia. So in these countries if you use the lower number for high value (like for for strongly agree) you will get problem. As we know that respondents generally do not give much time or care in filling the Questionnaire so there are high chances that their responses will be wrong in case you assign values to the scale in a manner that is against the norm

Agree that it doesn't matter, and answers can easily be recoded. Still it makes sense if you have more than one questionnaire measuring different variables, that there should be a common concept e.g. higher scores mean better health, or greater pathology.

In ranking livable city, we follow the classroom style, the student getting maximum number is ranked 1. We can start reversed ranking, statistically no problem, psychologically huge.

I highly appreciate all these valuable viewpoints and perspectives from all of you. Thanks a lot in in clarifying this confusing issue to me :)) They will aid me in my thesis defense. Jazakallah hul Khair !

Dear Colleagues... It has been valuable discussion on the matter ... can anyone just make me clear that if a researcher uses 1=Strongly agree to 7=Strongly disagree for the all positive items of a scale ...............

DOES HE NEED TO REVERSE CODE ALL ITEMS OR NOT???

Best Regards

Muhammed, it really is your choice but you need to be aware of the meaning either way. If the 'strongly agree' gives you a high score for whatever the person is agreeing with (e.g. I have lots of problems), then the higher the score, the more problems there are. If it's the opposite (as in your example), then the higher the score the less problems the person is experiencing.

And you reverse code if you want the opposite concept to have a high score.

Thanks Rachel for the explanation .... as for I understood from your response ... If a researcher code let say a statement like this "I work with passion" with 1=Strongly agree to 7=Strongly disagree ; and all variables of the study have same positive statements as well .... Researcher sure understand that 1 means what and what 7 means what BUT to tell the machine these scores would be reverse coded .... am i right?

Hi Muhammad, not quite sure I understand your question but...

If all the questions that are positive statments about a person's attitude towards work receive a LOW score as in your example, then the lower the total score, the more positive the person's attitude towards work is. A high score would mean the person is not passional about work.

But if you wanted a high score to mean that the respondent is positive about his/her work, you would reverse the coding ie 1=strongly disagree 7=strongly agree.

And if this questionnaire (about work attitudes) also contains NEGATIVE statements e.g. "I really hate my work", these questions would need to be scored in the opposite direction from the positive statements. Hope this helps.

Thanks again Rachel, now it is more clear. Regards

@ Man Bhutta, yes technically, we can do it,but I prefer to give high value (5) for best score(highest mean) and worst score (lowest mean) on low scale (1), it makes easy to remember and calculate.