标准差和标准误的选择-(SD)-和-(SEM)

标准差和标准误的选择(SD)和(SEM)Whicherrorbarshouldyouchoose?Itiseasytobeconfusedaboutthedifferencebetweenthestandarddeviation(SD)andstandarderrorofthemean(SEM).TheSDquantifiesscatter-howmuchthevaluesvaryfromoneanother.TheSEMquantifieshowaccuratelyyouknowthetruemeanofthepopulation.TheSEMgetssmallerasyoursamplesgetlarger.Thismakessense,becausethemeanofalargesampleislikelytobeclosertothetruepopulationmeanthanisthemeanofasmallsample.TheSDdoesnotchangepredictablyasyouacquiremoredata.TheSDquantifiesthescatterofthedata,andincreasingthesizeofthesampledoesnotincreasethescatter.TheSDmightgouporitmightgodown.Youcan'tpredict.Onaverage,theSDwillstaythesameassamplesizegetslarger.Ifthescatteriscausedbybiologicalvariability,yourprobablywillwanttoshowthevariation.Inthiscase,graphtheSDratherthantheSEM.YoucouldalsoinstructPrismtographtherange,witherrorbarsextendingfromthesmallesttolargestvalue.Alsoconsidergraphingeveryvalue,ratherthanusingerrorbars.Ifyouareusinganinvitrosystemwithnobiologicalvariability,thescattercanonlyresultfromexperimentalimprecision.Inthiscase,youmaynotwanttoshowthescatter,butinsteadshowhowwellyouhaveassessedthemean.GraphthemeanandSEMorthemeanwith95%confidenceintervals.Ideally,thechoiceofwhicherrorbartoshowdependsonthesourceofthevariabilityandthepointoftheexperiment.Infact,manyscientistsalwaysshowthemeanandSEM,tomaketheerrorbarsassmallaspossible.