Hello,
My friend created a hypothesis "Is x (variable) is significant?". I found this hypothesis weird:) Anyway, he insists on testing this. I calculated the mean of the scale (x) on SPSS. Is testing signficance of mean scale is a remedy for testing the hypothesis?
Thnks on advance,
Dear Ceren
If you have Only one variable you can Only test the hypithesis the mean of the variable different from zero or any hypothesized value.
The hypothesis makes no sense. end of story. Good luck, David booth
Reinforcing David Eugene Booth's response, Ceren, you should explain to your colleague what "statistical significance" means. Most of us still use the traditional Null Hypothesis Statistical Test (NHST) method. That means we ask whether a statistic is different from zero. For example, a one-sample t-test asks whether the mean of a variable is different from zero: a two-sample t-test asks if the difference between the means of two groups is different from zero. If you leave everything on default in SPSS and just give it one variable then it will calculate the mean of that variable and then test whether that mean is different from zero. Now think about that particular measurement and its properties: What is the lowest possible value? What is the highest possible value? Does it make sense if the measure could be zero? You haven't told us what has been measured, but you've pressed a button for me because I see this often among naive undergraduates and occasionally among my own colleagues who should know better. If your colleague is using, say, a 1-5 rating scale, then it's not possible for the measure to have an average of zero. So it's guaranteed to be "statistically significant". And it will make no sense.
Dear Ceren
If you have Only one variable you can Only test the hypithesis the mean of the variable different from zero or any hypothesized value.
"Significant" is a concept related to disparities between sample observations and null-hypothesized distribution properties. The most common properties are location, dispersion, trends, symmetry, or specific distributions like normal, gamma, etc. If you have more than one sample, you might also consider additional properties like association.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS