# How do I analyse this limited data? It is probably obvious but I have gone blank

##### New Member
I have a job interview where I need to present the main findings of some data but I am struggling to perform a data analysis on it. They have only given us the average but not the raw data so I cannot perform an ANOVA to see if there is a significant difference between the groups.

How can I tell if there is a significant difference in this data?
What data analysis can I perform?

Am I just being silly here and completely missing it? I was thinking of doing a histogram of the two tables and that it shows differences in life expectancy in regions and differences between time periods but are they significant differences?

I was also going to do an ANOVA between the two tables but I don't think that shows anything particularly useful as they are completely different data sets and unrelated so don't have much real-life relevance right?

I have attached the data and the presentation task.

Thank you so much!

#### rogojel

##### TS Contributor
hi,
it seems to me that simpler solutions are required - and obviously the two datasets can not be compared. Do you see any trend in the female data, any regional differences in the male data? Which regions differ significantly from one another, how would you show that with a graph - that sort of thing.

regards

##### New Member
Hi!

But how do I find if the differences are significant? And how do I find a trend? I can show one with a line graph for the female data and a bar chart for the male data but that seems way too basic to just point out that some numbers are bigger than others.

Thanks!

#### rogojel

##### TS Contributor
Well, for the male sata you can look at the confidence intervals and those that do not overlap point to significant differences. It is a rough measure though, slightly overlapping intervals could also be significantly different.
l
As for the trend, I would calculate the differences from one period to the other. They are all positive, whereas if there were no trend you would expect a scatter of plus and minus values. Maybe even look at the probability of randomly chosing 12 signs from (plus, minus) and getting only plusses.

regards