D ata A nalysis

D ata A nalysis

M akin g gen eralized com m en ts about data an alysis is difficult because such an alyses w ill vary greatly depen din g on th e particular question (s) asked an d w h at param eters are m easured. D ifferen t levels of an alysis can be appropriate for m ost param eters.

G raph ical T ech n iques

G raph in g data is very useful an d im portan t for un derstan din g th e ch aracteristics of th e "data set" (i.e. th e total am oun t of data collected for a particular m on itorin g site or project) an d iden tifyin g an y poten tial relation sh ips. E xam ples in clude bar ch arts,

X Y graph s, frequen cy distribution s, or pie ch arts. For exam ple, by graph in g stream tem perature versus

distan ce from a divide, an un derstan din g of basin tren ds can develop. B y graph in g stream tem perature versus tim e, an un derstan din g of w h en th e h igh est tem peratures occurred can be gain ed. T h is also provides a m ean s to ch eck th e data for accuracy.

D escriptive Statistics T h ese are th e very basic statistics th at describe a data set (for m ore in form ation on statistical an alysis, refer to th e m on itorin g m en tors listed on page 7 in

C h apter 2). C om m on ly reported statistics are: m edian , average, m axim um , m in im um , an d stan dard deviation . B y graph in g th e average plus an d m in us th e stan dard deviation , data collectors begin to un derstan d th e distribution of th eir data.

Statistical M eth ods T h e presen tation of data in a valid scien tific m an n er requires th at a statem en t of th e in vestigator’ s con fiden ce in th at data be in cluded. Statistical m eth ods are th e tools used to sh ow w h at levels of con fiden ce, or th e am oun t of error, in vestigators h ave in th e data. A n um ber of statistical m eth ods or m odels are available for an alyzin g data.

H ow ever, it is critical to un derstan d th e assum ption s of th ese m odels prior to usin g th em . For exam ple, m an y n atural resource data sets m ay n ot be n orm ally H ow ever, it is critical to un derstan d th e assum ption s of th ese m odels prior to usin g th em . For exam ple, m an y n atural resource data sets m ay n ot be n orm ally

W ater Q uality C riteria

sh aped” curve on a graph ) an d th erefore stan dard O regon w ater quality criteria are provided on th e w eb an alytical m eth ods m ay result in an alyses th at are at < h ttp://w aterquality.deq.state.or.us/w q/ flaw ed. T h ese problem s can often be addressed by w qrules/w qrules.h tm l> . T h ese criteria m ay be in

logarith m ic or pow er tran sform ation s of th e data. term s of a seven -day m ovin g average of th e daily N on -param etric m eth ods are also available (H irsch et m axim um or m in im um tem peratures. Special

al. 1992). Som e statistical an alyses in clude: con dition s m ay also be recogn ized w h ich n aturally

A N O V A , m ultiple an d lin ear regression , cause w ater quality to exceed th e stan dards. For

m ultivariate an alyses, an d correlation an alyses. exam ple extrem e low stream flow s or prolon ged Som e user-frien dly softw are packages are available w arm periods can cause stream s to exceed state to aid statistical an alyses. W ith out fam iliarity or tem perature stan dards. It is useful to an alyze th e

train in g in statistical an alysis, h ow ever, h elp in data collected an d com pare th e results to th e w ater developin g statistical m odels w ill be n eeded.

quality criteria.

C on tact on e of th e region al m on itorin g coordin ators listed on page 7 in C h apter 2 for furth er assistan ce.

D epositing D ata

T h e O PSW M on itorin g T eam is curren tly explorin g option s for storage of th e m on itorin g data collected

T h ese protocols w ill con form to th e for th e O PSW . Som e of th e attach ed protocols

recom m en dation s for data storage th at are bein g con tain exam ple data sh eets. T h ese sh eets provide a

developed an d w ill, in th e future, provide guidelin es tem plate for organ izin g th e data collected by

for tran sportin g an d deliverin g th e data to th e O PSW volun teers in to a form at com patible w ith th e O PSW

database. A t a m in im um , guidelin es for th e data database. In gen eral, som e im portan t com pon en ts