Hello
Would be grateful for a wee bit of help with my research.

I have two different groups (group 1 n=72, group 2 n=97) and I have a likert scale consist of seven items and each participant had to indicate degree of agree on a scale of 1 - 7 ie, 1 strongly disagree 7 strongly agree.

I would like to find out if there is a difference between the degree of agreement for each group on each on of the items

i.e. hypothesis = group 1 has a more positive view than group 2

I have used an independant samples t test but the more i look at it prehaps I should have used a mann-whitney u or wilcoxon as some books are indicating that you cannot compare means for a likert scale.

Basically i'm looking for some reassurance that I have done the right the or a point in the right direction if i'm not.

Thanks you

Whether you can treat a Likert scale as comprising interval-level data is one of those tough questions in applied research - there isn't really an easy answer. That said, the more options on the response scale, generally the more confident you can feel about assuming that the scale comprises interval-level data, so it helps that you've got 7 options, meaning that you can probably feel reasonably justified in comparing means.

It's also worth considering the distribution of the item responses (the the t-test requires normally distributed responses). If a Shapiro-Wilks test and skewness/kurtosis values indicate that responses are not normally distributed, you have another reason to switch to the non-parametric Mann-Whitney test.