topic Re: Parametric / non-parametric ? in Statistical Procedures

Parametric / non-parametric ?

koksiang100 — Sat, 08 Nov 2014 02:42:09 GMT

Actual 1100 1300 1400 1500 1600 1100 1200 1600 2100 1300 1600 1300 1600 2200 2300 1700 1800 800 1400 900 2100 1400 1800 1900 1000 1800 1700 2100 800 1100 900 1600 1700 1400 1100 1200 1700 900 700 900 1300 700 1500 700 1300 1100 1700 1600 1800 2000 1500 2100

Demand 1500 2100 1600 1500 2000 1600 1200 2000 2200 2000 2200 2000 2000 2500 2500 2000 2000 1000 2000 1500 2500 1500 2500 2500 2000 2000 2500 2500 1500 1500 1400 2000 2000 2000 1500 1500 2500 1500 1500 1500 2500 1500 2000 1500 1500 2000 2000 2500 2500 2500 2500 2500

These are the data collected. Both are sequential from 1st - 52nd weeks. How do determine whether it is parametric or non-parametric?

Re: Parametric / non-parametric ?

stat_sas — Sat, 08 Nov 2014 03:27:12 GMT

Try Kolmogorov Smirnov test for this.

Re: Parametric / non-parametric ?

koksiang100 — Sat, 08 Nov 2014 03:36:05 GMT

The maximum difference between the cumulative distributions, D, is: 0.5000 with a corresponding P of: 0.000

Data Set 1:

52 data points were entered

Mean = 1440.

95% confidence interval for actual Mean: 1322. thru 1559.

Standard Deviation = 425.

High = 2.300E+03 Low = 700.

Third Quartile = 1.700E+03 First Quartile = 1.100E+03

Median = 1450.

Average Absolute Deviation from Median = 352.

KS finds the data is consistent with a normal distribution: P= 0.84 where the normal distribution has mean= 1449. and sdev= 422.6

KS is not particularly happy calling this data log normally distributed: P= 0.17 where the log normal distribution has geometric mean= 1332. and multiplicative sdev= 1.394

Items in Data Set 1:

700. 700. 700. 800. 800. 900. 900. 900. 900. 1.000E+03 1.100E+03 1.100E+03 1.100E+03 1.100E+03 1.100E+03 1.200E+03 1.200E+03 1.300E+03 1.300E+03 1.300E+03 1.300E+03 1.300E+03 1.400E+03 1.400E+03 1.400E+03 1.400E+03 1.500E+03 1.500E+03 1.500E+03 1.600E+03 1.600E+03 1.600E+03 1.600E+03 1.600E+03 1.600E+03 1.700E+03 1.700E+03 1.700E+03 1.700E+03 1.700E+03 1.800E+03 1.800E+03 1.800E+03 1.800E+03 1.900E+03 2.000E+03 2.100E+03 2.100E+03 2.100E+03 2.100E+03 2.200E+03 2.300E+03

Data Set 2:

52 data points were entered

Mean = 1948.

95% confidence interval for actual Mean: 1829. thru 2067.

Standard Deviation = 426.

High = 2.500E+03 Low = 1.000E+03

Third Quartile = 2.500E+03 First Quartile = 1.500E+03

Median = 2000.

Average Absolute Deviation from Median = 340.

KS says it's unlikely this data is normally distributed: P= 0.02 where the normal distribution has mean= 1923. and sdev= 419.7

KS says it's unlikely this data is log normally distributed: P= 0.00 where the log normal distribution has geometric mean= 1851. and multiplicative sdev= 1.288

Items in Data Set 2:

1.000E+03 1.200E+03 1.400E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.500E+03 1.600E+03 1.600E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.000E+03 2.100E+03 2.200E+03 2.200E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03 2.500E+03

Data Reference: 2786

KS finds the data is consistent with a normal distribution: P= 0.84 where the normal distribution has mean= 1449. and sdev= 422.6

KS is not particularly happy calling this data log normally distributed: P= 0.17 where the log normal distribution has geometric mean= 1332. <---- what do these 2 sentences mean?

Re: Parametric / non-parametric ?

stat_sas — Sat, 08 Nov 2014 04:29:23 GMT

Seems like data2 variable does not follow normal distribution.

Re: Parametric / non-parametric ?

koksiang100 — Sat, 08 Nov 2014 04:38:28 GMT

so my data1 is following normal distribution and data2 is not.... therefore which test should i use and what assumption should i make ?

Re: Parametric / non-parametric ?

stat_sas — Sat, 08 Nov 2014 04:49:21 GMT

Why do you want to compare two variables which are not following same distribution?

Re: Parametric / non-parametric ?

koksiang100 — Sat, 08 Nov 2014 04:53:53 GMT

I have to prove that in order to justify the problem in my project. If these 2 variables are not following same distribution,is that any way to compare them?

Re: Parametric / non-parametric ?

stat_sas — Sat, 08 Nov 2014 04:59:31 GMT

Your data2 variable does not seem a random variable, it contains values which are very similar to each other. It is evident that two variables are different.

Re: Parametric / non-parametric ?

koksiang100 — Sat, 08 Nov 2014 05:01:18 GMT

in case of that, what test should I use in order to prove the significant difference in these 2 data?

Re: Parametric / non-parametric ?

stat_sas — Sat, 08 Nov 2014 05:17:34 GMT

Okay, I would say just divide data points of two variables into 4 to 5 classes and run chi-square test to see if there is an association between two variables.

Re: Parametric / non-parametric ?

koksiang100 — Sat, 08 Nov 2014 05:30:27 GMT

what will be my hypothesis in chi square test?

Re: Parametric / non-parametric ?

stat_sas — Sat, 08 Nov 2014 05:34:07 GMT

Is there any association between two classifications?

Re: Parametric / non-parametric ?

koksiang100 — Sat, 08 Nov 2014 05:42:57 GMT

Ok. Thanks a lot... I will proceed with that later...

If I proved that the 2 variables are independent , what test should I use then?

how about I prove they are dependent ?

Re: Parametric / non-parametric ?

stat_sas — Sat, 08 Nov 2014 05:51:02 GMT

Please try then we can go from there.

Re: Parametric / non-parametric ?

koksiang100 — Sat, 08 Nov 2014 10:19:46 GMT

Chi-Square Test for Association: Worksheet rows, Worksheet columns

Rows: Worksheet rows Columns: Worksheet columns

600-1000 1100-1500 1600-2000 2100-2600 All

1 10 19 17 6 52

5.50 17.50 17.50 11.50

2 1 16 18 17 52

5.50 17.50 17.50 11.50

All 11 35 35 23 104

Cell Contents: Count

Expected count

Pearson Chi-Square = 12.910, DF = 3, P-Value = 0.005

Likelihood Ratio Chi-Square = 14.316, DF = 3, P-Value = 0.003

here is the result that i obtained from minitab

Re: Parametric / non-parametric ?

koksiang100 — Sat, 08 Nov 2014 10:50:30 GMT

The chi-square statistic is 12.9102. The P-Value is 0.004835. The result is significant at p < 0.05.