Inputting data with lots of variables/ which test to use?


Really hope someone can help. For my final year project i am looking at language development in twins v singletons and comparing their results on different tests them by age and gender. I have no idea how to input this data to SPSS, I thought all along I would use an independent t-test but now I'm not sure should i be using ANOVA? So I need to input twins versus singletons, male/female, age i.e 1 female 8 months v 1 male 8 months and their actual test scores :?


TS Contributor
In most programs, including SPSS, the rows are your units and the columns your variables. So each child needs its own row. Each charateristic (e.g. is the child male or female, his or her age, etc.) needs its own column. This is the raw material with which your program works, and it uses that to compute the comparisons for you.

Think of your dataset as a record of what you have seen (I have seen a child, male singleton 8 months old that scored x on test y, I have seen a second child, male twin ... )


TS Contributor
Once you have created your dataset you can do whatever analysis you think is appropriate. The way you input the data says nothing about which test you can or should or cannot or should not perform.


Fortran must die
If you are comparing the impact of multiple variables on a dependent variable you would want to use something like ANOVA or regression rather than independent t tests.
Oh that's great to know thanks!! So would twin or singleton be my dependent variable and gender, age and test score independent variables? Thanks so much for the help!


TS Contributor
That is very unlikely, you don't become a twin because you are female or 8 months old... Think about what is affecting what.

I guess, but that depends on the details of your research project, that the test score is influenced by twin-status, gender, and age, so test score would be your dependent and the other varialbes your independent variables. But in the end, it is your decision.
oh great! i think i understand it now! so test score is my dependent variable and its going to be affected by being male/twin/age etc! :)