 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Math Forum / Math Software / SPSS / March 2010HELP ME, please... How to introduce a set of data and perform Kruskal-Wallis
Hi! I have never used SPSS before because I didn´t have to make statisics. Now thinks changed and I am fighting with a sample of data and cannot find the solucio... Could somebody tell me how to perform the Kruskal–Wallis test with the following set of data. I even cannot introduce them into SPSS, e.g. instead of S10, S11, etc. appear S1... I hope somebody can help me... I would like to find out if there is any significant difference witin th variables an between them. The answers are on a Likert scale: 0 - incorrect answer, 1 - "less incorrect" answer, 0 - I don´t know, 1- "less correct" answer, 2 - corret answer ID Group Item Condition Answer S1 G1 P1 C1 4 S1 G1 P2 C1 4 S1 G1 P3 C2 4 S1 G1 P4 C2 3 S1 G1 P5 C2 0 S1 G1 P6 C3 0 S1 G1 P7 C3 1 S1 G1 P8 C3 0 S1 G1 P9 C4 0 S1 G1 P10 C4 0 S1 G1 P11 C5 0 S1 G1 P12 C5 0 S2 G1 P1 C1 4 S2 G1 P2 C1 4 S2 G1 P3 C2 4 S2 G1 P4 C2 4 S2 G1 P5 C2 1 S2 G1 P6 C3 0 S2 G1 P7 C3 1 S2 G1 P8 C3 0 S2 G1 P9 C4 0 S2 G1 P10 C4 1 S2 G1 P11 C5 0 S2 G1 P12 C5 1 S3 G1 P1 C1 4 S3 G1 P2 C1 3 S3 G1 P3 C2 4 S3 G1 P4 C2 4 S3 G1 P5 C2 4 S3 G1 P6 C3 1 S3 G1 P7 C3 0 S3 G1 P8 C3 0 S3 G1 P9 C4 1 S3 G1 P10 C4 1 S3 G1 P11 C5 4 S3 G1 P12 C5 3 S4 G1 P1 C1 4 S4 G1 P2 C1 3 S4 G1 P3 C2 4 S4 G1 P4 C2 3 S4 G1 P5 C2 4 S4 G1 P6 C3 0 S4 G1 P7 C3 1 S4 G1 P8 C3 1 S4 G1 P9 C4 0 S4 G1 P10 C4 1 S4 G1 P11 C5 0 S4 G1 P12 C5 1 S5 G1 P1 C1 3 S5 G1 P2 C1 4 S5 G1 P3 C2 1 S5 G1 P4 C2 3 S5 G1 P5 C2 3 S5 G1 P6 C3 0 S5 G1 P7 C3 0 S5 G1 P8 C3 0 S5 G1 P9 C4 0 S5 G1 P10 C4 0 S5 G1 P11 C5 3 S5 G1 P12 C5 1 S6 G1 P1 C1 3 S6 G1 P2 C1 4 S6 G1 P3 C2 4 S6 G1 P4 C2 4 S6 G1 P5 C2 3 S6 G1 P6 C3 0 S6 G1 P7 C3 0 S6 G1 P8 C3 1 S6 G1 P9 C4 1 S6 G1 P10 C4 1 S6 G1 P11 C5 0 S6 G1 P12 C5 1 S7 G1 P1 C1 4 S7 G1 P2 C1 4 S7 G1 P3 C2 3 S7 G1 P4 C2 1 S7 G1 P5 C2 2 S7 G1 P6 C3 0 S7 G1 P7 C3 0 S7 G1 P8 C3 0 S7 G1 P9 C4 0 S7 G1 P10 C4 1 S7 G1 P11 C5 1 S7 G1 P12 C5 0 S8 G1 P1 C1 4 S8 G1 P2 C1 3 S8 G1 P3 C2 4 S8 G1 P4 C2 4 S8 G1 P5 C2 4 S8 G1 P6 C3 1 S8 G1 P7 C3 0 S8 G1 P8 C3 0 S8 G1 P9 C4 3 S8 G1 P10 C4 4 S8 G1 P11 C5 0 S8 G1 P12 C5 4 S9 G1 P1 C1 4 S9 G1 P2 C1 4 S9 G1 P3 C2 4 S9 G1 P4 C2 3 S9 G1 P5 C2 3 S9 G1 P6 C3 3 S9 G1 P7 C3 0 S9 G1 P8 C3 1 S9 G1 P9 C4 1 S9 G1 P10 C4 0 S9 G1 P11 C5 1 S9 G1 P12 C5 0 S10 G1 P1 C1 4 S10 G1 P2 C1 4 S10 G1 P3 C2 3 S10 G1 P4 C2 4 S10 G1 P5 C2 3 S10 G1 P6 C3 0 S10 G1 P7 C3 0 S10 G1 P8 C3 0 S10 G1 P9 C4 0 S10 G1 P10 C4 1 S10 G1 P11 C5 4 S10 G1 P12 C5 4 S11 G2 P1 C1 4 S11 G2 P2 C1 4 S11 G2 P3 C2 4 S11 G2 P4 C2 3 S11 G2 P5 C2 1 S11 G2 P6 C3 2 S11 G2 P7 C3 0 S11 G2 P8 C3 0 S11 G2 P9 C4 0 S11 G2 P10 C4 1 S11 G2 P11 C5 3 S11 G2 P12 C5 0 S12 G2 P1 C1 4 S12 G2 P2 C1 4 S12 G2 P3 C2 4 S12 G2 P4 C2 4 S12 G2 P5 C2 3 S12 G2 P6 C3 3 S12 G2 P7 C3 4 S12 G2 P8 C3 3 S12 G2 P9 C4 1 S12 G2 P10 C4 0 S12 G2 P11 C5 4 S12 G2 P12 C5 3 S13 G2 P1 C1 4 S13 G2 P2 C1 4 S13 G2 P3 C2 4 S13 G2 P4 C2 4 S13 G2 P5 C2 3 S13 G2 P6 C3 0 S13 G2 P7 C3 4 S13 G2 P8 C3 3 S13 G2 P9 C4 0 S13 G2 P10 C4 3 S13 G2 P11 C5 0 S13 G2 P12 C5 0 S14 G2 P1 C1 4 S14 G2 P2 C1 3 S14 G2 P3 C2 2 S14 G2 P4 C2 3 S14 G2 P5 C2 3 S14 G2 P6 C3 0 S14 G2 P7 C3 1 S14 G2 P8 C3 0 S14 G2 P9 C4 3 S14 G2 P10 C4 4 S14 G2 P11 C5 0 S14 G2 P12 C5 1 S15 G2 P1 C1 4 S15 G2 P2 C1 4 S15 G2 P3 C2 3 S15 G2 P4 C2 4 S15 G2 P5 C2 3 S15 G2 P6 C3 0 S15 G2 P7 C3 0 S15 G2 P8 C3 0 S15 G2 P9 C4 0 S15 G2 P10 C4 3 S15 G2 P11 C5 1 S15 G2 P12 C5 0 S16 G2 P1 C1 4 S16 G2 P2 C1 4 S16 G2 P3 C2 3 S16 G2 P4 C2 2 S16 G2 P5 C2 3 S16 G2 P6 C3 0 S16 G2 P7 C3 0 S16 G2 P8 C3 1 S16 G2 P9 C4 4 S16 G2 P10 C4 3 S16 G2 P11 C5 4 S16 G2 P12 C5 4 S17 G2 P1 C1 4 S17 G2 P2 C1 4 S17 G2 P3 C2 4 S17 G2 P4 C2 3 S17 G2 P5 C2 4 S17 G2 P6 C3 0 S17 G2 P7 C3 1 S17 G2 P8 C3 1 S17 G2 P9 C4 3 S17 G2 P10 C4 1 S17 G2 P11 C5 3 S17 G2 P12 C5 1 S18 G2 P1 C1 4 S18 G2 P2 C1 3 S18 G2 P3 C2 3 S18 G2 P4 C2 3 S18 G2 P5 C2 1 S18 G2 P6 C3 0 S18 G2 P7 C3 0 S18 G2 P8 C3 0 S18 G2 P9 C4 1 S18 G2 P10 C4 3 S18 G2 P11 C5 4 S18 G2 P12 C5 4 S19 G2 P1 C1 4 S19 G2 P2 C1 4 S19 G2 P3 C2 4 S19 G2 P4 C2 4 S19 G2 P5 C2 4 S19 G2 P6 C3 0 S19 G2 P7 C3 0 S19 G2 P8 C3 0 S19 G2 P9 C4 0 S19 G2 P10 C4 1 S19 G2 P11 C5 4 S19 G2 P12 C5 4 S20 G2 P1 C1 4 S20 G2 P2 C1 0 S20 G2 P3 C2 4 S20 G2 P4 C2 4 S20 G2 P5 C2 4 S20 G2 P6 C3 1 S20 G2 P7 C3 0 S20 G2 P8 C3 1 S20 G2 P9 C4 0 S20 G2 P10 C4 3 S20 G2 P11 C5 2 S20 G2 P12 C5 4 S21 G3 P1 C1 4 S21 G3 P2 C1 4 S21 G3 P3 C2 4 S21 G3 P4 C2 4 S21 G3 P5 C2 3 S21 G3 P6 C3 1 S21 G3 P7 C3 0 S21 G3 P8 C3 0 S21 G3 P9 C4 4 S21 G3 P10 C4 4 S21 G3 P11 C5 0 S21 G3 P12 C5 0 S22 G3 P1 C1 4 S22 G3 P2 C1 4 S22 G3 P3 C2 4 S22 G3 P4 C2 1 S22 G3 P5 C2 0 S22 G3 P6 C3 1 S22 G3 P7 C3 0 S22 G3 P8 C3 0 S22 G3 P9 C4 1 S22 G3 P10 C4 2 S22 G3 P11 C5 4 S22 G3 P12 C5 4 S23 G3 P1 C1 4 S23 G3 P2 C1 4 S23 G3 P3 C2 4 S23 G3 P4 C2 4 S23 G3 P5 C2 4 S23 G3 P6 C3 0 S23 G3 P7 C3 0 S23 G3 P8 C3 1 S23 G3 P9 C4 1 S23 G3 P10 C4 0 S23 G3 P11 C5 1 S23 G3 P12 C5 1 S24 G3 P1 C1 4 S24 G3 P2 C1 4 S24 G3 P3 C2 3 S24 G3 P4 C2 4 S24 G3 P5 C2 1 S24 G3 P6 C3 1 S24 G3 P7 C3 0 S24 G3 P8 C3 0 S24 G3 P9 C4 3 S24 G3 P10 C4 4 S24 G3 P11 C5 3 S24 G3 P12 C5 3 S25 G3 P1 C1 4 S25 G3 P2 C1 4 S25 G3 P3 C2 4 S25 G3 P4 C2 4 S25 G3 P5 C2 4 S25 G3 P6 C3 0 S25 G3 P7 C3 0 S25 G3 P8 C3 0 S25 G3 P9 C4 4 S25 G3 P10 C4 3 S25 G3 P11 C5 4 S25 G3 P12 C5 0 S26 G3 P1 C1 4 S26 G3 P2 C1 4 S26 G3 P3 C2 3 S26 G3 P4 C2 4 S26 G3 P5 C2 4 S26 G3 P6 C3 1 S26 G3 P7 C3 0 S26 G3 P8 C3 1 S26 G3 P9 C4 0 S26 G3 P10 C4 2 S26 G3 P11 C5 4 S26 G3 P12 C5 4 S27 G3 P1 C1 4 S27 G3 P2 C1 4 S27 G3 P3 C2 4 S27 G3 P4 C2 3 S27 G3 P5 C2 4 S27 G3 P6 C3 0 S27 G3 P7 C3 0 S27 G3 P8 C3 1 S27 G3 P9 C4 3 S27 G3 P10 C4 3 S27 G3 P11 C5 4 S27 G3 P12 C5 3 S28 G3 P1 C1 4 S28 G3 P2 C1 4 S28 G3 P3 C2 3 S28 G3 P4 C2 4 S28 G3 P5 C2 4 S28 G3 P6 C3 0 S28 G3 P7 C3 0 S28 G3 P8 C3 0 S28 G3 P9 C4 0 S28 G3 P10 C4 0 S28 G3 P11 C5 3 S28 G3 P12 C5 4 S29 G3 P1 C1 4 S29 G3 P2 C1 4 S29 G3 P3 C2 4 S29 G3 P4 C2 3 S29 G3 P5 C2 4 S29 G3 P6 C3 0 S29 G3 P7 C3 4 S29 G3 P8 C3 3 S29 G3 P9 C4 0 S29 G3 P10 C4 4 S29 G3 P11 C5 4 S29 G3 P12 C5 4 S30 G3 P1 C1 4 S30 G3 P2 C1 4 S30 G3 P3 C2 4 S30 G3 P4 C2 3 S30 G3 P5 C2 3 S30 G3 P6 C3 0 S30 G3 P7 C3 0 S30 G3 P8 C3 1 S30 G3 P9 C4 0 S30 G3 P10 C4 3 S30 G3 P11 C5 4 S30 G3 P12 C5 4 S31 G4 P1 C1 4 S31 G4 P2 C1 4 S31 G4 P3 C2 4 S31 G4 P4 C2 4 S31 G4 P5 C2 4 S31 G4 P6 C3 4 S31 G4 P7 C3 4 S31 G4 P8 C3 3 S31 G4 P9 C4 4 S31 G4 P10 C4 4 S31 G4 P11 C5 4 S31 G4 P12 C5 4 S32 G4 P1 C1 4 S32 G4 P2 C1 4 S32 G4 P3 C2 3 S32 G4 P4 C2 4 S32 G4 P5 C2 4 S32 G4 P6 C3 4 S32 G4 P7 C3 4 S32 G4 P8 C3 3 S32 G4 P9 C4 3 S32 G4 P10 C4 4 S32 G4 P11 C5 4 S32 G4 P12 C5 4 S33 G4 P1 C1 4 S33 G4 P2 C1 4 S33 G4 P3 C2 4 S33 G4 P4 C2 4 S33 G4 P5 C2 4 S33 G4 P6 C3 4 S33 G4 P7 C3 4 S33 G4 P8 C3 4 S33 G4 P9 C4 3 S33 G4 P10 C4 1 S33 G4 P11 C5 4 S33 G4 P12 C5 4 S34 G4 P1 C1 4 S34 G4 P2 C1 4 S34 G4 P3 C2 3 S34 G4 P4 C2 4 S34 G4 P5 C2 4 S34 G4 P6 C3 4 S34 G4 P7 C3 4 S34 G4 P8 C3 4 S34 G4 P9 C4 3 S34 G4 P10 C4 4 S34 G4 P11 C5 4 S34 G4 P12 C5 4 S35 G4 P1 C1 4 S35 G4 P2 C1 4 S35 G4 P3 C2 4 S35 G4 P4 C2 4 S35 G4 P5 C2 4 S35 G4 P6 C3 3 S35 G4 P7 C3 3 S35 G4 P8 C3 4 S35 G4 P9 C4 3 S35 G4 P10 C4 4 S35 G4 P11 C5 4 S35 G4 P12 C5 4 S36 G4 P1 C1 4 S36 G4 P2 C1 4 S36 G4 P3 C2 4 S36 G4 P4 C2 4 S36 G4 P5 C2 4 S36 G4 P6 C3 4 S36 G4 P7 C3 4 S36 G4 P8 C3 4 S36 G4 P9 C4 3 S36 G4 P10 C4 4 S36 G4 P11 C5 4 S36 G4 P12 C5 4 S37 G4 P1 C1 4 S37 G4 P2 C1 4 S37 G4 P3 C2 4 S37 G4 P4 C2 4 S37 G4 P5 C2 4 S37 G4 P6 C3 4 S37 G4 P7 C3 4 S37 G4 P8 C3 3 S37 G4 P9 C4 3 S37 G4 P10 C4 4 S37 G4 P11 C5 4 S37 G4 P12 C5 4 S38 G4 P1 C1 4 S38 G4 P2 C1 4 S38 G4 P3 C2 4 S38 G4 P4 C2 4 S38 G4 P5 C2 4 S38 G4 P6 C3 3 S38 G4 P7 C3 3 S38 G4 P8 C3 4 S38 G4 P9 C4 4 S38 G4 P10 C4 3 S38 G4 P11 C5 4 S38 G4 P12 C5 4 S39 G4 P1 C1 4 S39 G4 P2 C1 4 S39 G4 P3 C2 4 S39 G4 P4 C2 4 S39 G4 P5 C2 4 3S9 G4 P6 C3 4 S39 G4 P7 C3 4 S39 G4 P8 C3 4 S39 G4 P9 C4 4 S39 G4 P10 C4 2 S39 G4 P11 C5 4 S39 G4 P12 C5 4 S40 G4 P1 C1 4 S40 G4 P2 C1 4 S40 G4 P3 C2 4 S40 G4 P4 C2 4 S40 G4 P5 C2 3 S40 G4 P6 C3 4 S40 G4 P7 C3 4 S40 G4 P8 C3 4 S40 G4 P9 C4 4 S40 G4 P10 C4 3 S40 G4 P11 C5 4 S40 G4 P12 C5 4 > Hi! I have never used SPSS before because I didn´t have to make > statisics. Now thinks changed and I am fighting with a sample of [quoted text clipped - 6 lines] > 0 - incorrect answer, 1 - "less incorrect" answer, 0 - I don't know, > 1- "less correct" answer, 2 - corret answer Do you mean: 0 - incorrect answer, 1 - "less incorrect" answer, 2 - I don't know, 3 - "less correct" answer, 4 - correct answer ? The Kruskal-Wallis test is for data collected in a simple one-way design, but your design is more complicated. You have two crossed nests: subjects within groups, crossed with items within conditions; a further complication is that different conditions have different numbers of items (just as different groups could, but do not, have different numbers of subjects). An analysis of variance is possible, but I do not know how to tell SPSS what the design is. [... data deleted; see the original post ...] > > Hi! I have never used SPSS before because I didn´t have to make > > statisics. Now thinks changed and I am fighting with a sample of [quoted text clipped - 19 lines] > > [... data deleted; see the original post ...] Nesting can be indicated in two ways in SPSS: 1. ID within group item within condition 2. ID(group) item(condition) So I think you could do something like the following: UNIANOVA Answer BY Group Item Condition ID /RANDOM=ID /EMMEANS=TABLES(Group) /EMMEANS=TABLES(Condition) /EMMEANS=TABLES(Group*Condition) /DESIGN=Group ID within group condition item within condition group*condition group*item within condition . Or equivalently: UNIANOVA Answer BY Group Item Condition ID /RANDOM=ID /EMMEANS=TABLES(Group) /EMMEANS=TABLES(Condition) /EMMEANS=TABLES(Group*Condition) /DESIGN=Group ID(group) condition item(condition) group*condition group*item(condition) . Am I missing any terms, Ray? Bear in mind that one has to omit the last error term, because SPSS will try to compute a residual, so if all terms (including the last error term) are included, the residual will equal 0, and the model won't run. There's probably a way to run this model via the MIXED procedure too (i.e., as a multilevel model). GLM Repeated Measures might work too, but the data would have to be restructured to have one row per person. However, I think the different number of items per condition might make that approach problematic. -- Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/Home "When all else fails, RTFM." >>> Hi! I have never used SPSS before because I didn´t have to make >>> statisics. Now thinks changed and I am fighting with a sample of [quoted text clipped - 67 lines] > bwea...@lakeheadu.cahttp://sites.google.com/a/lakeheadu.ca/bweaver/Home > "When all else fails, RTFM." Not having used SPSS since the maroon-manual days, I'm in no position to say anything about the code. I organize the sources of variance in crossed-nest designs by making a table with one nest pair as row labels and the other as column labels. The interactions are at the intersections. In the present case the nest pairs are G, S|G and C, I|C, where G = group, S = subject, C = condition, I = item. It's arbitrary which pair goes down the side and which goes across the top. (For completeness, I put the grand mean µ in the corner.) µ C I|C G GC GI|C S|G SC|G SI|GC I assume G and C are fixed, and S is random. If I is fixed then each term involving S is the error terms for those above it in the table. If I is random then things are messier. For instance, one form of the test of G would be F' = MS(G)/[MS(S|G) + MS(GI|C) - MS(SI|GC)]. Another would be F" = [MS(G) + MS(SI|GC)]/[MS(S|G) + MS(GI|C)]. In both cases the degrees of freedom would be found using the Welch- Satterthwaite approximation. See your favorite text that covers combined mean squares and quasi-Fs. > >>> Hi! I have never used SPSS before because I didn´t have to make > >>> statisics. Now thinks changed and I am fighting with a sample of [quoted text clipped - 93 lines] > Satterthwaite approximation. See your favorite text that covers > combined mean squares and quasi-Fs. OK, so I omitted SC|G term, or in SPSS syntax, condition*ID(group). Here's the syntax with that term included: UNIANOVA Answer BY Group Item Condition ID /RANDOM=ID /EMMEANS=TABLES(Group) /EMMEANS=TABLES(Condition) /EMMEANS=TABLES(Group*Condition) /DESIGN=Group ID(group) condition item(condition) condition*ID(group) group*condition group*item(condition) . The SI|GC term is omitted, and shows up as the residual. -- Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/Home "When all else fails, RTFM." |
Reply to this Message |
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2010 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.