• Ei tuloksia

LIITE 8. VERTAILUARVOJEN MITTAUSEPÄVARMUUDET Appendix 8. Uncertainties of the assigned values

35 LIITE 10/1 LIITE 10. TULOKSISSA ESIINTYVIÄ KÄSITTEITÄ

Appendix 10. Explanations for the result sheets Laboratoriokohtaiset tulostaulukot (liite 11):

Analyte Määritys

Unit Yksikkö

Sample Näyte

z-Graphics z-arvo graafisesti esitettynä z-value z-arvo, joka lasketaan seuraavasti:

z = (x1 - X)/s

x. = yksittäisen laboratorion tulos X = vertailuarvo (the assigned value)

s = kokonaiskeskihajonnalle asetettu tavoitearvo (starget(%~)

Outl test OK yes -jos tulos on läpäissyt käytetyn outlier-testin C - jos tulos on hylätty Cochran -testissä H -jos tulos on hylätty Hampel —testissä G-2 -jos tulos on hylätty Grubbs-2-testissä

Cochranin outlier -testiä käytetään rinnakkaistulosten testaukseen. Cochranin outlier -testi testaa laboratorioiden sisäistä hajontaa.

Hampel-testi perustuu mediaanin j a yksittäisen laboratorion tuloksen erotuksen itseisarvoon (tulos x,, mediaavi Xmedj a erotukset d. = I Xmed — Xi I . Laboratorion tulos on harha-arvo, jos d. > 5,06MAD (erotusten d, mediaan).

Outlier -testit suoritettiin 95%:n merkitsevyystasolla.

Tulosaineiston normaalisuus tarlcistettiin aineiston Kolmogorov-Smimov -testillä.

Assigned value Vertailuarvo (tässä vertailukokeessa laskennallinen pitoisuus, tulosaineiston mediaavi tai valittuj en laboratorioiden tulosten keskiarvo)

2* Targ SD % Käytettykokonaiskeskihajonta(95 %merkitsevyystaso) Lab's result Laboratorionilmoittamatulos

Md. Mediaani

Mean Keskiarvo

SI) Keskihajonta

SI) % Keskihajontaprosenttina

SI) Robust-keskihaj onta: s* =1,5 x mediaani erotuksesta x. - x , joka on saatu korvaamalla yksittäinen tulos x. arvolla s* kunnes arvoissa x*jas* ei tapahdu merkitsevää muutosta

Passed Tilastolliseen käsittelyyn hyväksytyt tulokset Missing Puuttuva tulos, esim. <määritysraja

Num of labs Osallistuvien laboratorioiden kokonaismäärä

LIITE 10/2 36 Yhteenvetotaulukko (liite 12):

A - accepted ( -2 < z < 2), hyväksytyt tulokset

p - questionable (2< z < 3), arveluttava tulos (positiivinen virhe, tulos > vertailuarvo)

n - questionable (-3 < z< -2), arveluttava tulos (negatiivinen virhe, tulos pienempi kuin vertailuarvo) P- non- accepted (z > 3), ei hyväksytty tulos (positiivinen virhe, tulos > vertailuarvo)

]v- non- accepted (z <-3), ei hyväksytty tulos (negatiivinen virhe, tulos < vertailuarvo)

37 LI TE

APPENDIX 11 /

LI TE 11. LABORATORIOKOHTAISET TULOKSET

Appendix 11. Results of the interlaboratory comparison test

Analyte i Unit Sample z-Graphics Z- value Outi Assig- 2 Lab's Md. Mean SD SD o Pas- Outl. Mis- Num

Outlier test failed: C - Cohcran, G1 - Grubbs(1-outlier algorithm), G2 - Grubbs(2-outliers algorithm), H - Hampel, M - manual SYKE - Interlaboratory comparison test 3/2002

LIITE APPENDIX 11 /

Outlier test failed: C - Cohcran, G1 - Grubbs(1-outlier algorithm), G2 - Grubbs(2-outliers algorithm), H - Hampel, M - manual SYKE - Interlaboratory comparison lest 3/2002

39 LITE

Outlier test failed: C - Cohcran, G1 - Grubbs(1-outlier algorithm), G2 - Grubbs(2-outliers algorithm), H - Hampel, M - manual SYKE - Interlaboratory comparison lest 3/2002

LITE 11 / 4 - 40

Outlier test failed: C - Cohcran, G1 - Grubbs(1-outlier algorithm), G2 - Grubbs(2-outliers algorithm), H - Hampel, M - manual SYKE - Interlaboratory comparison test 3/2002

41 LIITE

APPENDIX

Analyte Unit Sample z-Graphics Z- value Outi Assig- 2* Lab's Md. Mean SD SD% Pas- Outi. Mis- Num

3 -2 -1 0 +1 +2 +3 test ned Targ result sed tai- sing of

OK value SD% led labs

Laboratory 34

a-chlorophyll pgll Cl -02108 yes 21,35 20 2O. 21,35 21,39 3,075 iS 2T 1 0 28

pg/l C2 -0,3836 yes 5,605 20 5.39 5,605 5,518 0,6945 12,5 26 1 0 27

pg11 C3 0 yes 53,5 15 53.5 53,5 52,73 2,922 5,5 25 4 0 29

Si 02 -1 mg/ 1 H 0,2991 yes 10,7 10 10,86 10,57 10,45 0,4241 4,1 12 0 0 12

mgA P2 1,321 yes 0,737 15 0,81 0,737 0,7563 0,07018 9,3 9 0 0 9

mg/I P3 . 0,7679 yes 5,73 10 5,95 5,73 5,715 0,1689 3 12 0 0 12

Si 02 -2 mgll P1 0823 yes 10,45 10 10,88 10,45 10,38 0,4202 4 12 0 0 12

mg/I P2 1,481 yes 0,72 15 0,8 0,72 0,7325 0,05016 6,8 8 1 0 9

mg/1 P3 1,024 yes 5,665 10 5,955 5,665 5,679 0,1983 3,5 12 0 0 12

Outlier test failed: C - Cohcran, G1 - Grubbs(1-outlier algorithm), G2 - Grubbs(2-outliers algorithm), H - Hampel, M - manual SYKE - Interlaboratory comparison test 3/2002

LUTE

APPENDIX 42

LIITE 12. YHTEENVETO LABORATORIOIDEN TULOKSISTA, z -ARVOT Appendix 12. Summary of the z scores

Antilyte., ] $arnplelt_ab 1 2 3 4 5 , 6 7 8 . 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

%* - percentage of accepted results

Totally accepted, % In all: 89 In accredited: 96 In non-accredited: 84

SYKE - Interlaboyatory comparison test 3/2002

43

KUVAILULEHTI

Julkaisija Julkaisun päivämäärä

Suomen ympäristökeskus lokakuu 2002

Tekijä(t) (toimielimestä: nimi, puheenjohtaja, sihteeri)

Irma Mäkinen, Sami Huhtala, Kaija Korhonen, Teemu Näykki, 011i Järvinen, Pasi Salonen ja Miia Aalto Julkaisun nimi (myös ruotsinkielinen)

Laboratorioiden välinen vertailukoe 3/2002

Julkaisun laji Toimeksiantaja Toimielimen asettamispvm

Raportti Julkaisun osat

Tiivistelmä

Suomen ympäristökeskuksen laboratorio j är] esti toukokuussa 2002 vertailukokeen a-klorofyllin, silikaatin (Si02), saliniteetin ja orgaanisen hiilen (TOC) määrittämisestä luonnonvesistä.

Vertailukokeeseen osallistui yhteensä 34 laboratoriota.

Tulosten arvioimiseksi laskettiin z-arvoja sitä varten asetettiin kokonaiskeskihaj orman tavoitearvot.

Kokonaiskeskihajonnan tavoitearvot olivat a-klorofyllille, Si02:lle ja TOC:lle 10 - 20 % sekä saliniteetille 0,02 %o (95 % merkitsevyystaso). Vertailuarvona (the assigned value) käytettiin synteettisille näytteille (TOC/T1 ja Si02-1/S1) laskennallista arvoa. Muille näytteille vertailuarvoksi asetettiin tulosaineiston mediaani.

Vertailukokeessa seurattiin erityisesti silikaattinäytteiden säilyvyyttä (säilytyslämpötila ja -aika).

Näytteissä ei tapahtunut selviä muutoksia eri säilytyslämpötiloissa seuratun aj anj akson aikana (noin 2,5 kuukautta).

Erot eri analyysimenetelmillä saatujen tulosten välillä olivat pieniä. Tulosten keskihajonta oli yleensä pienempi kuin 10 %.

Vertailukokeen tuloksista hyväksyttiin 89 %. Laboratorloista 38 % käytti akkreditoituja analyysimenetelmiä. Näiden laboratorioiden tuloksista hyväksyttiin 96 %, kun kokonaan akkreditoimattomia menetelmiä käyttäneiden laboratonoiden tuloksista hyväksyttiin 84 %.

Asiasanat (avainsanat)

vesianalyysi, vesi- ja ympäristölaboratoriot, vertailukoe, interkalibrointi, pätevyystesti Muut tiedot

English summary

Sarjan nimija numero ISBN ISSN

Suomen ympäristökeskuksen moniste 263 952-11-1251-4 1455-0792

Kokonaissivumäärä Kieli Hinta Luottamuksellisuus

44 suomi julkinen

Jakaja Kustantaja

Suomen ympäristökeskus Suomen ympäristökeskus

Asiakaspalvelu PL 140

Puh: (09) 4030 0119 00251 Helsinki

Fax: (09) 4030 0190

44

DOCUMENTATION PAGE

Published by Date of publication

Finnish Environment Institute October 2002

Author(s)

Irma Mäkinen, Sami Huhtala, Kaija Korhonen, Teemu Näykki, 011i Järvinen, Pasi Salonen and Miia Aalto Title of publication

Interlaboratory Comparison 3/2002

Type of publication Commissioned by

Report

Parts of publication

Abstract

The Finnish Environment Institute (SYKE) carried out the interlaboratory comparison for analyses of a-chlorophyll, silicate (SiO2), salinity and total organic carbon (TOC) in May 2002. One artificial sample and two natural water samples were distributed. In total 34 laboratories participated in the comparison.

The mean value, the median value, the standard deviation and the relative standard deviation were calculated after rejection of the outliers according to the Cochran test, Grubb-2 test or Hampel test.

Either the calculated concentration (artificial samples: TOC/T1 and Si02/S1) or the median value of the results (other samples) was chosen to be the assigned value. The performance of the laboratories were evaluated by using z scores (Appendices 11 and 12).

The analytical methods are presented in Appendix 7.1. The differences of the results obtained by different analytical methods were rather small (Appendices 7.2 and 7.3).

In particular, the stability of silicate samples were tested (temperature, time). The samples were stable and no significant changes occured between different storage temperatures during the 2,5 month observation time.

In this comparison test 89 % of the data was regarded to be acceptable when deviation of 10-20 % for a-chlorophyll, Si02 TOC and 0,02 %o for salinity (95 % confidence level) from the assigned value was accepted. Mostly the standard deviation of the data was smaller than 10 %. 38 % of the participating laboratories used accredited methods. 96 % of their results were accepted. Only 84 % of the results was accepted in the cases, when the laboratories used non-accredited methods.

Keywords

water analyses, water and environmental laboratories, interlaboratory comparison, performance testing, intercalibration Other information

Series (key title and no.) Mimeograph series of

the Finnish Environment Institute no 263

Pages Language

Distributed by Publisher

Finnish Environment Institute Finnish Environment Institute

Tel: (09) 4030 0119 P.O. Box 140

Fax: (09) 4030 0190 FIN-00251 Helsinki

ISBN 952-I 1-1251-4 ISSN 1455-0792

0

■ ■ ■ • • ■ •

r a

• r • • • • • • ■ • ■ ■ • • • • • • • • • •