0
$\begingroup$

I have a quick question regarding which statistical test to use to test for a significant difference.

My set of data shows the monthly output of a product for the years between 2013-2017. A new system was put in place in mid-2017 to make the output process more efficient which has led to increased output.

I have calculated the mean output level (units) for each month 2013-2016 and want to compare this to the actual output in 2017.

Which statistical test can I use to show a significant difference for output in the months of June, July, Aug and Sept 2017 which have a higher output level compared to the 2013-2016 average?

enter image description here

enter image description here

Kind regards,

Harvey

$\endgroup$
  • $\begingroup$ Are you really interested in just testing for a statistically significant difference or a meaningful difference? Would it really be helpful if you found a 0.00001 difference and that was statistically significant? $\endgroup$ – StatsStudent Jan 6 at 20:15
  • $\begingroup$ Hi there, Thanks for the reply. I was hoping to find a significant difference (p<0.05). I was thinking perhaps a T-Test or an ANOVA test ? $\endgroup$ – Harvey Jan 6 at 20:32
  • $\begingroup$ Was the supposedly superior new process in place only during May-Aug? Or did it continue from May through the end of the year. // It seems you should make the comparison only for months in which the process was different. $\endgroup$ – BruceET Jan 6 at 20:37
  • $\begingroup$ Hi there, the process was introduced in May and continued after that. The process aimed to alleviate a backlog of output orders hence the spike. Which statistical test could I use to easily show a sig. difference? $\endgroup$ – Harvey Jan 6 at 20:47
  • $\begingroup$ Comment continued: With data ` old = c(20.25, 20.5, 24.25, 33.75); new = c(87, 48, 52, 44)` a one-sided paired t test in R using ` t.test(old, new, pair=T, alte="less")` returns P-value 3.5%. A one-sample Wilcoxon test on the four differences cannot show a P-value smaller than 5% because there aren't enough pairs. $\endgroup$ – BruceET Jan 6 at 20:47

Your Answer

By clicking "Post Your Answer", you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.