I obtained 4 values from triplicate determination when subjected to statistics at p≤0.05 of four samples and obtained their mean±SD values as 30.02±0.01, 30.04±0.04, 35.10±0.03 and 50.00±0.10. Please can anyone interpret the result based on significance and correctly assign the right superscript letter using letter a,b,c or d on them.
Please interest is particularly on how the calculation of comparison was done i.e. on how you arrive at a conclusion that one is significantly different or not. Thanks.
Hi Egbuna,
You must use one-way ANOVA test for the calculation. And if you obtain a p value of this test as
If I understand what you mean, none of the 4 data is not statistically significant. For instance, the observation 30.02+/-0.01 is significant because 30.02 is more than twice its standard error of 0.01. The same argument is applicable to the rest of the data. The data are already in their ascending order a, b, c and d even though a=30.02 and b=30.04 do not seem to be statistically different.
My Dear Egbuna
First you can use ANOVA to test main factor by SPSS (or any Software),if this factor is signinficant (sig.
Hello Egbuna, previous answers are correct. I add that from the definition, the p-value is the minimum significance level at which the null hypothesis is rejected. Once the p-value is less than alpha=0.05, (or whatever critical significance level), you reject the null hypothesis( In your case that means across the four samples are the same). Using ANOVA(since you can't use the t-test as the groups are more than 2), now that you have sufficient evidence to claim that the means are different between groups, you would want to know which groups' means are different and how much. Then you proceed with posthoc analyses such as Tukey HSD, Beniferon tests etc. In SAS, those letters a, b, c etc are assigned in your output. Hope this helps.
Why is your standard deviation so tiny? What the heck are you measuring?
And – very important – why are you asking strangers on RG to do your analysis for you? You cannot defend this in a publication or a thesis, and there's a good reason why not.
Tell us some more about the data and the problem. And tell us what you have been doing to analyse it. We can help, but anyone who answers a question posed in the abstract – as yours is – is making so many assumptions about your study that their advice is valueless, and potentially seriously misleading.
I quite agree with Prof. Ronan Michael Conroy. In getting a well expatiated answer, you must present your question in a detailed manner.
Thank you Dear John Ng'ombe, Tahir Kemal Şahin, Ette Etuk, Fadhil Abdulabbas Abidy for your suggestions. They are very helpful. The clear point there is do Post hoc! Well not actually a novice like Ronan Michael Convoy might think. Sometimes you need to reflect to see if you can find better ways of doing things. I think the best way to be smart in amassing RG scores is to keep shut when you don't understand a question or have nothing in your head to contribute. True scientists are not strangers. I don't need to sip a coffee of tea with you before you contribute.
I am quite curious as to why are the standard errors so small. I think (without looking at your data) that you have 3 technical replicates, which basically is measuring the same thing 3 times. e.g. I measure your weight 3 times, to be sure and take the average. However if you treat these as 3 individual measurements - which I think what is happening here, you will end up with such small standard errors (almost close to 0). This suggests that your measurement has no error or zero errors.
A measurement on a scale around 30, with a standard error of 0.04 suggests that you are trying to treat technical replicates as replicated samples.
Ah – Umar makes a very good point. Can you tell us, please, why the standard errors are so small? I think that all the good advice above relies on the data being independent observations, and the standard errors suggest that they are not and that, therefore, the advice isn't just wrong, but dangerous.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics