One sample chi-square I have analysed very data set using an online calculator. I used a chi-square comparing the differences between 


*

*'strongly agree/agree' 
      versus

*'strongly disagree/disagree' to specific questions. 


I only had 2 groups (or two sets of data)instead of the usual 4.
I now need to calculate the Cramers V in order to report the effect size in my results write up, but online calculators don't provide this tool. I now need to analyse the data using SPSS ticking the option to calculate the Cramer's V.
I am looking for advice as to how to I should lay my data out in SPSS. There is a lot of information online for when using 4 groups for observed frequencies, but nothing for 2 groups.
 A: I'm not sure why you're collapsing your 5 point Likert scale to two groups and throwing away the neutral response but with that aside, it seems you want to assess whether the proportion who selected SA/A is significantly different compared to the proportion that selected SD/D.  Since you only have two categories a Chi-Squared test doesn't make sense.  You can do a one sample proportion test for the SA/A (or vice versa) category and see if it is significantly different than 0.5.
For example, if out of 200 users (ignoring those who selected neutral) 120 selected SA/A then your estimated proportion is 0.6 and you can run a one sample proportion test.  There are effect sizes for proportions if you want to calculate those.  
You seem pretty set on Cramer's V, probably because someone asked for it, but I don't see how it would apply here (which is probably why you're not finding a reference online). 
I also have concerns analyzing your data this way but hope that helps.
Sample R code:
    library(vcd)
    library(powerAnalysis)
    set.seed(0807)
    # x vector represents your data, 0 could be SD/D and 1 A/SA
    # Note, these tests should all come back non significant
    x=sample(0:1,100,replace=TRUE)
    tab=table(x)
    print(tab)
    prop.table(tab)
    chisq.test(tab) 
    prop.test(tab)
    binom.test(tab)
    # effect size proportion 
    ES.proportions(p1 = prop.table(tab)[2], p2 = 0.5, alternative = c("two.sided",
                                                         "one.sided"))
    # does not work
    assocstats(tab)

A: I agree with the answer by @Glen .
I think there's a little confusion in the subsequent discussion with the term "chi-square test".
Cramér's V is usually used in cases of a chi-square test of association (or test of independence).  In this case, each side of the contingency test needs to have a least two categories.  That is, the table is at least 2 x 2 in size.
The question asks about a situation for which a chi-square goodness-of-fit test is applicable.  Here, the contingency table is 1 x k.  As indicated by @Glen, an effective way to express the effect size in this case is simply to present the proportions.  In the case of a 1 x 2 table, this makes complete sense, and is easily interpreted by the audience.
There is a variant of Cramér's V that could be used in the case of the chi-square goodness-of-fit test.  Cramér's V in the usual case is sqrt(Chi.sq/N/min(R - 1, C - 1)), where Chi.sq is the chi-square statistic, N is the total number of observations, R is the number of categories in rows, and C is the number of categories in columns.  By analogy, in the case of chi-square goodness of fit, a statistic can be calculated as sqrt(Chi.sq/N/(K - 1)), where K is the number of categories.
Unfortunately, I don't have any good reference for this modified statistic except for the R function I wrote.  There is also a function in the lsr package that works in this case.  I note here that if the theoretical proportions for the categories are equal, that Cramér's V for goodness-of-fit tests will range from 0 to 1.  If theoretical proportions are not equal, Cramér's V can exceed 1.
### Code for R:

observed = c(10, 20, 30, 40)

library(rcompanion)
cramerVFit(observed)

   ### Cramer V 
   ###  0.2582

library(lsr)
cramersV(observed)

   ### [1] 0.2581989

