Dr Smith is trying to detect whether there is a correlation between soil acidity and the number of flowers on plants She grows a treatment group of plants on acidified soil and a control group on normal soil She measures 40 plants in each group and finds that there is a weak correlation but it is not quite significant p 0 054 so she measures another 10 plants in each group and recalculates the p value a 0 046 She rejects the null hypothesis In her paper she writes only that the sample size was 50 and the correlation was significant What is the problem with this statistical analysis O Cheating O Undisclosed flexibility O Confounding variables O Non independent replicates O Multiple comparisons