The following two points are from the website linked below:

• A Likert scale is a multi-item scale, not a single item. A single item, regardless of its format, should not be called a Likert scale. 
• A Likert item, which is a single item or question, should adhere to certain format requirements. An item that is merely ordered-categorical, even if it is combined with similar items in a composite scale, should not be called a Likert item or a Likert-type item. The term, discrete visual analog scale (DVAS) is more generic and an alternative in many cases.

Which do you have, scales or items?  People will me much more willing to accept use of a paired t-test if you have multi-item scales.  HTH.

http://john-uebersax.com/stat/likert.htm

Thanks Bruce, 

I looked at the website you referred to. I think i have items as they are single item questions. 

Question example "i have good knowledge of grey nurse sharks"

Strongly disagree =1 to Strongly agree= 5

"i am confident in delivering grey nurse shark information to customers"

Wilcoxon Matched- Pairs Signed Ranks Test, W would be a better option.

Assumption for applying t test is that data normalywnormaly  distributed, if data is not  normally distributed and if there is outliers in data thenpowerfull sign rank test is more powerfull... u can check normallity by applying boxplot graph...

Assumption for applying t test is that data must be normally  distributed, if data is not  normally distributed and if there is outliers in data then wilcoxon sign rank test is more powerfull... u can check normallity by applying boxplot graph...

With 5-point Likert items, you must have a large number of tied scores.  The common rank-based tests have ways of correcting for ties, but I think it important to bear in mind that they were designed for use with ranks, not with variables that have a small number of ordered categories.  I.e., scores on Likert items are not the same thing as ranks. 

Bear in mind too that parametric tests are generally very robust to violations of their assumptions.  See the recent article by Geoff Norman, for example (link below). 

HTH.

http://link.springer.com/article/10.1007/s10459-010-9222-y#page-1

https://en.wikipedia.org/wiki/Likert_scale

Whether individual Likert items can be considered as interval-level data, or whether they should be treated as ordered-categorical data is the subject of considerable disagreement in the literature,[10][11] with strong convictions on what are the most applicable methods. This disagreement can be traced back, in many respects, to the extent to which Likert items are interpreted as being ordinal data.

The t-test makes no sense logically, because the tested hypothesis is about a mean difference, and "means" do not make sense on Likert data.

The sign-rank test mentioned above may be compromised by the low number of different catagories and the large expected number of ties. I would instead prefer a binomial test to test the "stochastic equality" of the pre- and post data.