Hello all,
Firstly, thank you to all the Statalist members who give up their time to make this such a useful resource - it has been a great help to me on many occasions but this is the first time I've asked my own question.
I am using Stata 12.1 for Windows.
I have a dataset containing about 93,000 cases of cancer from a particular region. I am hoping to compare the mean age at incidence (age_inc) of different types of cancer (identified by the variable icdgroup) by deciles of an index of socioeconomic deprivation (Scottish Index of Multiple Deprivation; variable name simd_dec, where 1 – most affluent and 10 is most deprived).
I have therefore run one-way ANOVA on age_inc by simd_dec for each of four values of icdgroup (i.e. the four different cancers), which has worked fine. However, I would now like to do a pairwise comparison with a Bonferroni correction. When I try this, I get an error message of “simd_dec variable not found”, even though it is the same variable on which the ANOVA has just been successfully run.
I have pasted the output below for the first type of cancer (icdgroup==1) as an example:
I have looked for solutions from the Stata manuals and from previous Statalist queries and found very little (except the following instance of the same problem, but it is unclear from this how it was resolved - http://www.statalist.org/forums/foru...-post-hoc-test).
Any suggestions would be very much appreciated.
Many thanks,
Emily
Firstly, thank you to all the Statalist members who give up their time to make this such a useful resource - it has been a great help to me on many occasions but this is the first time I've asked my own question.
I am using Stata 12.1 for Windows.
I have a dataset containing about 93,000 cases of cancer from a particular region. I am hoping to compare the mean age at incidence (age_inc) of different types of cancer (identified by the variable icdgroup) by deciles of an index of socioeconomic deprivation (Scottish Index of Multiple Deprivation; variable name simd_dec, where 1 – most affluent and 10 is most deprived).
I have therefore run one-way ANOVA on age_inc by simd_dec for each of four values of icdgroup (i.e. the four different cancers), which has worked fine. However, I would now like to do a pairwise comparison with a Bonferroni correction. When I try this, I get an error message of “simd_dec variable not found”, even though it is the same variable on which the ANOVA has just been successfully run.
I have pasted the output below for the first type of cancer (icdgroup==1) as an example:
Code:
. oneway age_inc simd_dec if icdgroup==1, tab SIMD decile Summary of Age at incidence 2006 Mean Std. Dev. Freq. 1 70.258799 12.291913 1449 2 70.086885 11.79507 1830 3 70.51696 11.687249 1592 4 70.742821 11.878411 1567 5 69.860345 11.568012 1740 6 71.111269 11.334224 2112 7 70.940016 11.536052 2534 8 70.743123 11.82404 2690 9 70.662172 11.379479 2892 10 70.13876 11.896448 3243 Total 70.525244 11.7027 21649 Analysis of Variance Source SS df MS F Prob > F Between groups 3125.62482 9 347.291647 2.54 0.0066 Within groups 2961636.83 21639 136.865698 Total 2964762.45 21648 136.953181 Bartlett's test for equal variances: chi2(9) = 20.6710 Prob>chi2 = 0.014 . pwcompare simd_dec, effects mcompare(bonferroni) variable simd_dec not found r(111);
Any suggestions would be very much appreciated.
Many thanks,
Emily
