If I understand your question correctly, and are interested in assessing the average value of the response, then I do not believe the sample size calculator would be valid.
Consider – if you are actually investigating the mean response, then you would wish to have an estimate of the variation (std dev), consider the practical difference you would like to detect, the power of your test (e.g. 0.8) and your type 1 error rate (e.g. 0.05), as well as the form of the alternative hypothesis (e.g. does not equal). Armed with this information then you could use the power and sample size functionality from a statistical software package. This is not present in the tool at https://www.isixsigma.com/offsite.asp?A=Fr&Url=http://www.surveyguy.com/SGcalc.htm if that is indeed what you’re alluding to. That device is interested in proportions, not means (e.g. %of people who would vote for candidate X). For more info on 1-sample t-tests you may be interested in the paper at http://www.minitab.com/resources/KeepingTAB/kt33/index.htm and investigating power and sample size determination in a book on introductory statistics.
Hope this helps.
.
.