CHI SQUARE ANALYSIS OF SURVERY RESULTS

Hi,

I have done a chi square test on some survey data and someone had objected that my methodology pseudoreplicates.

The situation is that each year I survey all staff and get a 60% response rate.
I wanted to look if there was a statistically significant difference between last year and this. All the questions require a yes no response.

Lets say there are 100 people who participate and they each answer 4 questions. The questions are all discrete. Now I have looked for statistically sign differences in the number of staff answering yes to a question in the survey - so my population is 400. Is this a problem? I can always do the chi square test for every question, but is it wrong to do what I have done, or invalid, if I have clearly stated it ( no cells have a count of less than 5)

I'd be very grateful for any comments.

Best wishes,

Andrew